WO2013190092A1 - Gene signatures for copd diagnosis - Google Patents

Gene signatures for copd diagnosis Download PDF

Info

Publication number
WO2013190092A1
WO2013190092A1 PCT/EP2013/062996 EP2013062996W WO2013190092A1 WO 2013190092 A1 WO2013190092 A1 WO 2013190092A1 EP 2013062996 W EP2013062996 W EP 2013062996W WO 2013190092 A1 WO2013190092 A1 WO 2013190092A1
Authority
WO
WIPO (PCT)
Prior art keywords
genes
copd
test sample
biomarker
sample
Prior art date
Application number
PCT/EP2013/062996
Other languages
French (fr)
Inventor
Stéphanie BOUE
Florian Martin
Marja TALIKKA
Yang Xiang
Original Assignee
Philip Morris Products S.A.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Philip Morris Products S.A. filed Critical Philip Morris Products S.A.
Publication of WO2013190092A1 publication Critical patent/WO2013190092A1/en

Links

Classifications

    • C—CHEMISTRY; METALLURGY
    • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
    • C12Q—MEASURING OR TESTING PROCESSES INVOLVING ENZYMES, NUCLEIC ACIDS OR MICROORGANISMS; COMPOSITIONS OR TEST PAPERS THEREFOR; PROCESSES OF PREPARING SUCH COMPOSITIONS; CONDITION-RESPONSIVE CONTROL IN MICROBIOLOGICAL OR ENZYMOLOGICAL PROCESSES
    • C12Q1/00—Measuring or testing processes involving enzymes, nucleic acids or microorganisms; Compositions therefor; Processes of preparing such compositions
    • C12Q1/68—Measuring or testing processes involving enzymes, nucleic acids or microorganisms; Compositions therefor; Processes of preparing such compositions involving nucleic acids
    • C12Q1/6876—Nucleic acid products used in the analysis of nucleic acids, e.g. primers or probes
    • C12Q1/6883—Nucleic acid products used in the analysis of nucleic acids, e.g. primers or probes for diseases caused by alterations of genetic material
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B25/00—ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression
    • G16B25/10—Gene or protein expression profiling; Expression-ratio estimation or normalisation
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
    • G16B20/10—Ploidy or copy number detection
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
    • G16B20/20—Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G16B40/20—Supervised data analysis
    • C—CHEMISTRY; METALLURGY
    • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
    • C12Q—MEASURING OR TESTING PROCESSES INVOLVING ENZYMES, NUCLEIC ACIDS OR MICROORGANISMS; COMPOSITIONS OR TEST PAPERS THEREFOR; PROCESSES OF PREPARING SUCH COMPOSITIONS; CONDITION-RESPONSIVE CONTROL IN MICROBIOLOGICAL OR ENZYMOLOGICAL PROCESSES
    • C12Q2600/00—Oligonucleotides characterized by their use
    • C12Q2600/158—Expression markers
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B25/00—ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression

Definitions

  • the present invention relates to gene signatures that are useful for the diagnosis of Chronic Obstructive Pulmonary Disease (COPD).
  • COPD Chronic Obstructive Pulmonary Disease
  • the present invention also relates to methods of diagnosing COPD.
  • the invention further relates to arrays and computer readable media comprising gene signatures for diagnosis of COPD.
  • COPD Chronic Obstructive Pulmonary Disease
  • COPD ulcerative colitis
  • emphysema chronic bronchitis
  • the airways in the lung chronically narrow, limiting the amount of airflow through the airways.
  • Individuals affected by COPD experience decreased elasticity in their airways and air sacs, the destruction of the walls between air sacs, the inflammation of the walls of the airways, and the clogging of airways by increased production of mucus.
  • COPD is caused by long-term exposure to lung irritants in the environment, such as air pollution, chemical fumes, dust and tobacco smoke. Because long-term exposure to lung irritants is typically required, most individuals who suffer from COPD are more than 40 years old when symptoms first present. Such symptoms include an ongoing cough or a cough that produces an excess of mucus; shortness of breath, especially with physical activity; wheezing and chest tightness. COPD symptoms slowly worsen over time, and most affected individuals do not notice symptoms at first because they are so mild or easy to correct by lifestyle adjustment.
  • COPD is typically diagnosed by signs and symptoms, including medical history, family history and test results. Diagnostic tests for COPD include lung function tests, such as spirometry, lung volume measurement, and lung diffusion capacity; chest X-rays, chest CT scans and arteriole blood gas tests. Accordingly, the diagnostic tests for COPD require the disease to have progressed to the point that lung function is moderately affected. Thus, there is a need for a diagnostic test that can identify COPD in patients at early stages. Recently, there have been attempts to identify genes associated with COPD or the treatment thereof. See, e.g. , US 201 1/0160070; US 2010/01 19474; and US 2009/0186951 . A need still persists to understand the molecular mechanisms of COPD, which may allow for the design or optimization of therapies to treat the disease, instead of just the symptoms.
  • the present invention is directed to gene signatures for classifying, diagnosing or grading Chronic Obstructive Pulmonary Disease (COPD).
  • COPD Chronic Obstructive Pulmonary Disease
  • a first aspect of the invention provides a method of diagnosing, classifying or grading COPD in an individual at risk for or having COPD.
  • the method comprises classifying a test sample as COPD or non-COPD.
  • the method comprises measuring the expression levels of at least 2 genes listed in Table 1 in a test sample: and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as COPD or non-COPD.
  • a differential pattern of expression levels of said at least 2 genes in the test sample diagnoses, classifies or grades the COPD.
  • the differential pattern of expression levels is identified by a classifier based on a plurality of genes listed in Table 1 , including said at least two genes, said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms.
  • the differential pattern of expression levels is identified by a classifier based on a pl ural ity of genes listed in Table 1 , including said at least two genes, said classifier having been trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T-filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude.
  • the classifier may be trained with at least the data in the Gene Expression Omnibus datasets GSE10106, GSE10135, GSE1 1906, GSE1 1952, GSE13933, GSE19407, GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545.
  • the method comprises detecting the expression level of at least 2 genes listed in Table 1 in a test sample obtained from the individual; and comparing the expression level of the genes listed in Table 1 in the test sample to the expression level of the genes listed in Table 1 in a control sample. In some embodiments, if the expression level of the genes listed in Table 1 is different in the test sample than in the control sample, then the individual suffers from COPD. In some embodiments, the expression level of the genes listed in Table 2 is higher in the test sample than in the control sample. Optionally, the expression level of the genes listed in Table 3 is lower in the test sample than in the control sample. In some embodiments, the method further comprises detecting the expression level of at least 2 genes listed in Table 1 in the control sample.
  • the at least 2 genes are selected from the group consisting of: PROS l , IRAKI , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR IA, PAPD4, VCL, PPP2R5C, DGKA, and CYP51 A1 .
  • the test sample or control sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine.
  • the test sample is obtained from a large airway of the individual, such as from a bronchial brush inserted into the large airway of the individual.
  • the control sample is obtained from a large airway of an individual not affected with COPD, such as from a bronchial brush inserted into the large airway of the individual not affected with COPD.
  • the control sample is obtained from the individual at risk for or having COPD prior to onset of COPD. In other embodiments, the control sample is obtained from an individual that does not suffer from COPD.
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring mRNA levels.
  • the expression level of the genes listed in Table 1 the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1 and said at least 2 genes.
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring the level of proteins encoded by the genes.
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring both mRNA levels and the level of proteins encoded by the genes.
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control biological sample are compared by in silica analysis (e.g., network-based analysis or machine-learning methods).
  • a second aspect of the invention provides an array for use in diagnosing or prognosing COPD.
  • the array comprises polynucleotides immobilized on a solid surface that can hybridize to at least 2 signature genes, wherein the COPD signature genes are selected from the group consisting of the genes listed in Table 1.
  • the array comprises polynucleotides hybridizing to at least 2 lung cancer signature genes immobilized on a solid surface, wherein the lung cancer signature genes are selected from the genes listed in Table 1 .
  • the array is not a human genome-wide array.
  • a third aspect of the invention provides a panel for use in diagnosing or prognosing COPD.
  • the panel comprises antibodies immobilized on a solid surface that bind to proteins encoded by at least 2 COPD signature genes, wherein the COPD signature genes are selected from the group consisting of the genes listed in Table 1.
  • a fourth aspect of the invention provides a computer readable medium or a computer program for use in diagnosing or prognosing COPD.
  • the computer readable medium or computer program comprises a COPD gene signature, wherein the gene signature comprises at least 2 genes selected from the genes listed in Table 1.
  • the computer readable medium or computer program product comprises a classifier based on at least two genes listed in Table 1 , said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms.
  • the classifier is trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T-fi!ter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude.
  • the classifier may be trained with at least the data in the Gene Expression Omnibus datasets GSE10106, GSE10135, GSE1 1906, GSE1 1952, GSE13933, GSE19407, GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545. In some
  • the at least two genes are selected from the group consisting of PROS 1 , IRAKI , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR1A, PAPD4, VCL, PPP2R5C, DGKA, and CYP51A1.
  • a fifth aspect of the invention provides a device for diagnosing or prognosing COPD.
  • the device comprises means for detecting the expression level of the genes listed in Table 1 in a test sample; means for correlating the expression level with a diagnosis or prognosis of the COPD; and means for outputting the COPD diagnosis or prognosis.
  • the device further comprises means for detecting the expression level of the genes listed in Table 1 in a control sample.
  • a sixth aspect of the invention provides a kit for diagnosing or prognosing COPD.
  • the kit comprises a set of reagents that detects expression levels of the genes listed in Table 1 in a test sample and instructions for using said kit for diagnosing the COPD.
  • the kit is for assessing the prognosis of COPD in an individual.
  • the kit comprises a set of reagents that detects expression levels of the genes listed in Table 1 in a test sample from the individual and instructions for using said kit
  • the set of reagents that detects expression levels of the genes listed in Table 1 in the test sample may also be used to detect expression levels of the genes listed in Table 1 in a control sample.
  • An seventh aspect of the invention provides a method of diagnosing COPD n an individual or of assessing the prognosis of an individual with COPD.
  • the method comprises a) measuring the expression level of at least 2 genes/biomarkers selected from the group consisting of the genes listed in Table 1 in a biological sample obtained from the individual; b) calculating a numerical biomarker score for the individual based on the expression levels of the biomarkers measured in step a); wherein the numerical biomarker score is predictive of the diagnosis of COPD in the individual.
  • the method comprises measuring the expression level of at least 2 genes/biomarkers selected from the group consisting of the genes listed in Table 1 in a biological sample obtained from the individual; calculating a numerical biomarker score for the individual based on the expression levels of the biomarkers measured in step a); wherein the numerical biomarker score is predictive of the prognosis of COPD in the individual.
  • the biological sample is selected from blood,
  • the tissue is lung tissue, such as tissue obtained by biopsy from a tumor.
  • the expression level of the genes listed in Table 1 in the biological sample is detected by measuring mRNA levels.
  • the expression level of the genes listed in Table 1 the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1 and said at least 2 genes.
  • the expression level of the genes listed in Table 1 in the biological sample is detected by measuring the level of proteins encoded by the genes.
  • the expression level of the genes listed in Table 1 in the biological sample is detected by measuring both mRNA levels and the level of proteins encoded by the genes.
  • the numerical biomarker score is calculated by in siiico analysis.
  • the in silico analysis may be network based analysis or machine-learning methods.
  • the biomarkers are proteins encoded by the genes selected from the group consisting of the genes listed in Table 1 .
  • a method of diagnosing an individual as being at risk for or having Chronic Obstructive Pulmonary Disease (COPD) comprising
  • test sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine. 6. The method according to paragraph 5, wherein the test sample is obtained from a large airway of the individual.
  • test sample is obtained from a bronchial brush inserted into the large airway of the individual.
  • control sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine.
  • control sample is obtained from a large airway of an individual not affected with COPD.
  • control sample is obtained a bronchial brush inserted into the large airway of the individual not affected with COPD.
  • control sample is obtained from the individual at risk for or having the COPD prior to onset of the COPD.
  • control sample is obtained from an individual that does not suffer from COPD.
  • hybridization is a dot blot, a slot blot, an RNase protection assay, microarray hybridization, or in situ hybridization.
  • a panel comprising antibodies immobilized on a solid surface that bind to proteins encoded by at least 2 COPD signature genes, wherein the COPD signature genes are selected from the genes listed in Table 1 .
  • a computer readable medium comprising a gene signature, wherein the gene signature comprises at least 2 genes selected from the genes listed in Table 1 .
  • a device for diagnosing COPD the device comprising: means for detecting the expression level of at least 2 genes listed in Table 1 in a test sample; means for correlating the expression level with a diagnosis of the COPD; and means for outputting the COPD diagnosis.
  • a device for prognosing COPD comprising: means for detecting the expression level of at least 2 genes listed in Table 1 in a test sample; means for correlating the expression level with a prognosis of the COPD; and means for outputting the COPD prognosis.
  • a kit for classifying and grading COPD comprising a set of reagents that detects expression levels of at least 2 genes listed in Table 1 in a test sample and instructions for using said kit for classifying and grading COPD in said individual.
  • a kit for assessing the prognosis of COPD in an individual comprising a set of reagents that detects expression levels of at least 2 genes listed in Table 1 in a test sample from the individual and instructions for using said kit for determining the prognosis of the COPD in said individual.
  • a method of diagnosing, prognosing, classifying or grading COPD in a biological sample or an individual comprising measuring the expression levels of at least 2 genes listed in Table 1 in said biological sample or a test sample obtained from said individual; and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as COPD or non-COPD.
  • a method of diagnosing an individual as being at risk for or having Chronic Obstructive Pulmonary Disease comprising detecting the expression level of at least 2 of the genes listed in Table 1 in a test sample obtained from the individual; wherein a differential pattern of expression levels of said at least 2 genes in the test sample diagnoses the individual as suffering from COPD.
  • COPD Chronic Obstructive Pulmonary Disease
  • differential pattern of expression levels is identified by a classifier based on a plurality of genes listed in Table 1 , including said at least two genes, said classifier having been trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T- filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning,
  • An array comprising polynucleotides hybridizing to at least 2 COPD signature genes immobilized on a solid surface, wherein the COPD signature genes are selected from the genes listed in Table 1 and said array is not a human genome-wide array.
  • a device comprising antibodies immobilized on a solid surface that bind to proteins encoded by at least 2 COPD signature genes, wherein the COPD signature genes are selected from the genes listed in Table 1 .
  • a computer readable medium or computer program product comprising a classifier based on at least two genes listed in Table 1 , said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms.
  • Figure 1 provides a features selection and classification algorithm(s) used for prediction of a gene signature.
  • antibody refers to an immunoglobulin molecule capable of specific binding to a target, such as a carbohydrate, polynucleotide, lipid, polypeptide, etc., through at least one antigen recognition site, located in the variable region of the immunoglobulin molecule.
  • the term i intended to encomnass not onlv intact nolvclonal or monoclonal antibodies, but also engineered antibodies (e.g., chimeric, humanized and/or derivatized to alter effector functions, stability and other biological activities) and fragments thereof (such as Fab, Fab', F(ab')2, Fv), single chain (ScFv) and domain antibodies, including shark and camelid antibodies), and fusion proteins comprising an antibody portion, multivalent antibodies, multispecific antibodies (e.g., bispecific antibodies so long as they exhibit the desired biological activity) and antibody fragments as described herein, and any other modified configuration of the immunoglobulin molecule that comprises an antigen recognition site.
  • engineered antibodies e.g., chimeric, humanized and/or derivatized to alter effector functions, stability and other biological activities
  • fragments thereof such as Fab, Fab', F(ab')2, Fv), single chain (ScFv) and domain antibodies, including shark and camelid antibodies
  • An antibody includes an antibody of any class, such as IgG, IgA, or IgM (or sub class thereof), and the antibody need not be of any particular class.
  • immunoglobulins can be assigned to different classes. There are five major classes of immunoglobulins: IgA, IgD, IgE, IgG, and IgM, and several of these may be further divided into subclasses (isotypes), e.g., IgG l , IgG2, IgG3, IgG4, IgAl and IgA2 in humans.
  • the heavy chain constant domains that correspond to the different classes of immunoglobulins are called alpha, delta, epsilon, gamma, and mu, respectively.
  • the subunit structures and three dimensional configurations of different classes of immunoglobulins are well known.
  • the term "array” refers to the arrangement of biomarker detection molecules, such as nucleic acid probes or antibodies, on a solid support that allows for high-throughput screening of a sample to detect the presence and/or quantity of a biomarker. Such arrays may be used, e.g. , to evaluate the expression levels of several genes of interest in a single high-throughput reaction.
  • the array may be a nucleic acid array, such as a nucleic acid microarray; a protein array, such as a protein microarray; a peptide array, such as a peptide microarray; a tissue microarrav. such as a tissue microarrav or an antibodv microarrav. such as an antibody microarray.
  • the solid substrate may be a microscopic bead, a glass slide, a plastic chip or a silicon chip.
  • biomarker refers to a characteristic whose presence, absence or level indicates a biological state. Typically, the properties of biomarkers indicate a normal process, a pathogenic process or a response to a pharmaceutical or therapeutic intervention.
  • a biomarker can be a cell, a gene, a gene product, an enzyme, a hormone, a protein, a peptide, an antibody, a nucleic acid molecule, a metabolite, a lipid, a free fatty acid, cholesterol or some other chemical compound.
  • a biomarker can be a morphologic biomarker (for example, a histological change, DNA ploidy, malignancy-associated changes in the ceil nucleus and premalignant lesions) or a genetic biomarker (for example, DNA mutations, DNA adducts and apoptotic index).
  • a morphologic biomarker for example, a histological change, DNA ploidy, malignancy-associated changes in the ceil nucleus and premalignant lesions
  • a genetic biomarker for example, DNA mutations, DNA adducts and apoptotic index
  • COPD Chironic Obstructive Pulmonary Disease
  • COPD refers to a complex disease that results in progressive loss of lung function.
  • COPD is typically characterized by the occurrence of chronic bronchitis or emphysema, both of which result in airway narrowing.
  • Clinically, COPD is typically detected by low airflow in lung function tests.
  • COPD is typically irreversible and gets progressively worse over time. Symptoms of COPD include chronic cough, chronic sputum production, dyspnea, rhonchi, wheezing, chest tightness, tiredness and decreased airflow in lung function tests.
  • COPD ulcerative colitis
  • BODE index is a scoring system that measures FEV 1 , body-mass index, 6-minute walk distance, and a modified MRC (Medical Research Council) dyspnea scale to estimate outcomes in COPD.
  • COPD refers to a method for determining the type of COPD from which a subject suffers.
  • COPD can be classified as primarily bronchial or primarily emphysematous. Such classifications are made by simply analyzing clinical, functional, and radiological findings or by detecting biomarkers.
  • bronchial COPD lung damage and inflammation occurs in the large airways resulting in chronic bronchitis, which is characterized by hyperplasia and hypertrophy of goblet cells and mucous glands in the airway.
  • COPD emphysematous COPD
  • lung damage and inflammation occurs in the alveoli and is characterized by enlargement of the air spaces distal to the terminal bronchioles, with destruction of their walls and a reduction in lung elasticity.
  • COPD may be classified based on the presence, absence, alteration or levels of biomarkers.
  • COPD may be classified based on the COPD gene signature. Classifying COPD may also refer to distinguishing between bronchial or emphysematous COPD.
  • a computer program refers to a sequence of instructions, written to perform a specified task within a computer.
  • a computer program product is described, the product comprising computer-readable instructions that, when executed in a computerized system comprising at least one processor, cause the processor to carry out one or more steps of any of the methods described above.
  • a computerized system is described, the system comprising a processor configured with non-transitory computer-readable instructions that, when executed, cause the processor to carry out any of the methods described herein.
  • the computer program product and the computerized methods described herein may be implemented in a computerized system having one or more computing devices, each including one or more processors.
  • the computerized systems described herein may comprise one or more engines, which include a processor or devices, such as a computer, microprocessor, logic device or other device or processor that is configured with hardware, firmware, and software to carry out one or more of the computerized methods described herein. Any one or more of these engines may be physically separable from any one or more other engines, or may include multiple physically separable components, such as separate processors on common or different circuit boards.
  • the computer systems of the present invention comprises means for implementing the methods and its various embodiments as described herein.
  • the computerized system described herein may include a distributed computerized system having one or more processors and engines that communicate through a network interface. Such an implementation may be appropriate for distributed computing over multiple communication systems.
  • Non- volatile media include, for example, optical, magnetic, or opto- magnetic disks, or integrated circuit memory, such as flash memory.
  • Volatile media include, for example, dynamic random access memory (DRAM), which typically constitutes the main memory.
  • DRAM dynamic random access memory
  • Computer-readable media include, for example, a floppy disk, a flexible disk, hard disk, magnetic tape, magnetic cards, magnetic ink characters, magnetic drums, any other magnetic medium, a CD-ROM, DVD, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, a RAM, a PROM, an EPROM or EEPROM (electronically erasable programmable read-only memory), a
  • FLASH-EEPROM FLASH-EEPROM, barcodes, semiconductors, microchips and any other memory chip or cartridge.
  • control sample refers to a sample against which a test sample is compared in order to diagnose, prognose, classify or grade the test sample.
  • a control sample may be healthy tissue or may be a well-characterized sample from an individual suffering from COPD, including but not limited to, GOLD stage 1 , GOLD stage 2, GOLD stage 3, or GOLD stage 4 COPD.
  • a control sample can be analyzed concurrently with or separately from the test sample, including before or after analyzing the test sample.
  • the data from the analysis of a control sample may be stored, e.g., in a computer readable medium or in a manual, for comparison against test samples analyzed in the future, or as data for training network-based or machine-learning methods.
  • a control sample may be developed as a medical standard for comparison.
  • control sample includes samples that provided a medical standard. Accordingly, a test sample may be compared against a medical standard generated from control samples. For example, expression of a variant or mutated form of a gene may be indicative of a change medical condition. Alternatively, a change in expression level of a gene may be indicative of a change in medical condition.
  • a control sample may be lung tissue, such as tissue obtained by biopsy from a healthy individual, or some other sample.
  • a control sample may be blood, blood cells, serum, plasma, sputum, saliva, tissue, bronchial washing, bronchial aspirates, bronchia brushings, exhaled breath, lymph fluid, and urine.
  • Tissue specimens such as those obtained by biopsy, may be fixed (e.g., formaldehyde-fixed paraffin-embedded (FFPE)).
  • FFPE formaldehyde-fixed paraffin-embedded
  • the control sample may be obtained from a tissue bank.
  • the control sample may also be obtained from a cadaver or an organ donor.
  • the terms "differential pattern of expression” and “differential expression” are used interchangeably herein and refer to a difference in an activity measurement (e.g., the variability or difference of genetic expression) of a biological entity under different conditions.
  • one condition may refer to an experimental treatment (such as exposure to a potentially carcinogenic agent), and another condition may refer to a control treatment (such as a null treatment).
  • a fold-change is a number describing how much a measurement at a node (or biological entity) changes from an initial value to a final value between control data and treatment data, or between two sets of data representing different treatment conditions.
  • the fold-change number may represent the logarithm of the fold-change of the activity of the biological entity between the two conditions.
  • FEV l force expiratory volume in one second
  • FEV l refers to the volume of air that can forcibly be blown out in one second, after full inspiration. Average values for FEV l in healthy individuals depend on sex and age and have been well-characterized in the art. FEVl and the FEV l to FVC ration (FEVl /FVC) are used clinically to grade COPD. In healthy adults
  • FEVl /FVC should be approximately 75-80%. In obstructive diseases, such as COPD, FEV l is diminished because of increased airway resistance to expiratory flow. While the FVC may be decreased as well, due to the premature closure of airway in expiration, FEV l is typically more affected because of the increased airway resistance, so the FEV1/FVC ratio reflects the degree of airway closure compared to lung volume.
  • F VC force vital capacity
  • a gene signature may be characterized by which genes are expressed or at what level each gene is expressed. Gene signatures are particularly useful in diagnosing, prognosing, classifying or grading complex diseases states, which result from the combination of several genetic and environmental factors.
  • the gene signatures disclosed herein may be used, e.g., for the diagnosis, prognosis, classification and/or grading of COPD in an individual. The gene signature may be unique to the class and grade of COPD.
  • grade COPD refers to a method for determining the grade of COPD from which a subject suffers.
  • grade COPD There are several different grades of COPD, which reflect the severity of the disease.
  • GOLD Global Initiative for Chronic Obstructive Lung Disease
  • GOLD stage 0 refers to a high risk population who did not present the symptoms used to describe stage 1 ;
  • stage 1 refers to mild COPD and is characterized by a FEV l /FVC ratio less than 70% and an FEVl greater than 80%;
  • stage 2 refers to moderate COPD and is characterized by a FEV l/FVC ratio less than 70% and an FEV l between 50% and 80%;
  • stage 3 refers to sever COPD and is characterized by a FEV l /FVC ratio less than 70%) and an FEVl between 30% and 50%o and
  • stage 4 refers to very severe COPD and is characterized by a FEVl/FVC ratio less than 70% and an FEV l less than 30% or the presence of chronic renal failure or right heart failure.
  • the term "in silico analysis” refers to analysis performed on a computer or via computer simulation. Gene signature analysis involves detection of gene expression based on identity and expression level for a multitude of genes. In silico analysis may apply one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample, e.g., as COPD or non-COPD. Comparisons between expression levels from test samples and control samples may require computer analysis to determine the degree and significance of any changes observed. See, e.g. , U.S.
  • the term "individual” refers to a vertebrate, preferably a mammal.
  • the mammal can be, without limitation, a mouse, a rat, a cat, a dog, a horse, a pig, a cow, a non-human primate or a human.
  • the term "individual at risk for COPD” refers to an individual who is predisposed to COPD. Predisposition to COPD may be due to one or more genetic or environmental factors. For example, an individual related to a COPD patient is more likely to get COPD than an individual who is not related to a COPD patient. Further, exposure to environmental factors such as radon gas, asbestos, tobacco smoke, and air pollution can increase the risk for COPD and predispose an individual to COPD.
  • COPD can be bronchial or emphysematous and may be detected by analyzing clinical, functional, and radiological findings or detecting relevant biomarkers.
  • MALDI-TOF matrix-assisted laser
  • Time-of- flight (TOF) mass spectrometry refers to a method in which an ion's mass-to-charge ratio is determined via the time that it takes an ionized particle to reach a detector at a known distance.
  • machine learning methods refers to methods that allow a machine, such as a programmable computer, to improve its performance at a certain predictive task that is based on the known properties of examples or training data.
  • Machine learning methods include, without limitation, support vector machines (SVMs), network-based SVMs, ensemble classifiers, neural network-based classifiers, logistic regression classifiers, decision tree-based classifiers, classifiers employing a linear discriminant analysis technique, a random-forest analysis technique, or both.
  • SVMs support vector machines
  • network-based SVMs ensemble classifiers
  • neural network-based classifiers logistic regression classifiers
  • decision tree-based classifiers classifiers employing a linear discriminant analysis technique, a random-forest analysis technique, or both.
  • network-based methods refers to methods for identifying biomarkers that is based on the properties of groups of functionally interrelated genes that form a network in a biological system, instead of treating individual genes in the biological system a priori as completely independent and identical.
  • numbererical biomarker score refers to a number that is representative of the result(s) of one or more of the network-based analysis or machine learning methods.
  • polynucleotide hybridizing to refers to a polynucleotide molecule that binds to a target nucleic acid molecule through complementary base pair sequencing. Hybridization typically requires two nucleic acids that contain complementary sequences, although depending on the stringency of the
  • hybridization mismatches between bases are possible.
  • the appropriate stringency for hybridizing nucleic acids depends on the length of the nucleic acids and the degree of complementation, variables well known in the art. Exemplary high stringent hybridization conditions are equivalent to about 20-27 °C below the melting temperature (T m ) of the DNA duplex formed in about 1 M salt.
  • T m melting temperature
  • Many equivalent procedures exist and several popular molecular cloning manuals describe suitable conditions for highly stringent hybridization and, furthermore. provide formulas for calculating the length of hybrids expected to be stable under these conditions (see, e.g., Current Protocols in Molecular Biology, John Wiley & Sons, N.Y. (1989), 6.3.1 6 or 13.3.6; or pages 9.47-9.57 of Sambrook, et al.
  • High stringency refers to hybridization and/or washing conditions at 68 °C in 0.2 x SSC, at 42 °C in 50 % formamide, 4 x SSC, or under conditions that afford levels of hybridization equivalent to those observed under either of these two conditions.
  • SELDI-TOF surface-enhanced laser
  • Time-of-flight (TOF) mass spectrometry refers to a method in which an ion's mass-to-charge ratio is determined by measuring the time that it takes an ionized particle to reach a detector at a known distance.
  • test sample refers to a sample obtained from an individual at risk for, having or suffering from COPD.
  • a test sample may be any sample suspected of containing or exhibiting a biomarker.
  • the test sample is analyzed and compared to a control sample, including medical standards developed from control samples, to diagnose, prognose, classify or grade COPD in the individual.
  • a test sample may be obtained from lung tissue, such as tissue obtained by biopsy from a tumor, or other biological tissue.
  • a test sample may be blood, blood cells, serum, plasma, sputum, saliva, tissue, bronchial washing, bronchial aspirates, bronchia brushings, exhaled breath, lymph fluids, and urine.
  • Tissue specimens, such as those obtained by biopsy may be fixed (e.g., formaldehyde-fixed paraffin- embedded (FFPE)).
  • FFPE formaldehyde-fixed paraffin- embedded
  • a training data set includes a set of samples, and each sample may correspond to a measurement from a different patient.
  • a machine learning technique is applied to the training data set to generate a "classifier," which corresponds to a way of assigning each sample in the training data set to a category (such as "disease positive" or "disease free”).
  • a training class set is known.
  • the training class set includes a known category assigned to each sample (or person).
  • the categories predicted by the classifier are compared to the known categories. If the predicted categories mostly match the known categories, the classifier has performed well.
  • the parameters of the machine learning technique may be updated, and the updated machine learning technique is applied. These steps are repeated until the performance of a classifier exceeds a threshold, and the final classifier is provided.
  • the final classifier may then be applied to a test data set.
  • the test data set may correspond to measured samples from different patients, but the patients in the test data set may have unknown categories (disease states). Thus, applying the final classifier to the test data set thus allows for prediction of the disease states of the patients.
  • the gene signature comprises at least 2 genes selected from the genes listed in Table 1. In some embodiments, the gene signature comprises at least 2, at least 3, at least 4, at 1 of ⁇ it 1 at Ipoct 1 ⁇ at lpacf at Ip cf at l ⁇ acf ⁇ n af l a ⁇ ⁇ at
  • the gene signature comprises each of the genes listed in Table 1.
  • said at least 2, at least 3, at least 4, at least 5, or at least 10 genes are selected from the group consisting of: PROS1, IRAKI, VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR1A, PAPD4, VCL, PPP2R5C, DGKA, and CYP51A1, which are the genes that appear in 4 of the 5 lists generated in Example 1.
  • the gene signature includes a degree of up- regulation of a subset of genes in the gene signature compared to the control sample.
  • each up-regulated gene in the gene signature may, independently, be up-regulated at least 1.5-fold, at least 2-fold, at least 2.5-fold, at least 3-fold, at least 3.5-fold, at least 4-fold, at least 4.5-fold, at least 5-fold, at least
  • the gene signature includes a degree of down-regulation of a subset of genes in the gene signature compared to the control sample.
  • each down-regulated gene in the gene signature may, independently, be down-regulated at least 1 .5-fold, at least 2-fold, at least 2.5-fold, at least 3 -fold, at least 3.5-fold, at least 4-fold, at least 4.5-fold, at least 5-fold, at least 6-fold, at least
  • the present invention encompasses the following gene signatures
  • xvii A, B, C, D, E, F, G, II, I, J, K, L, M, N, 0, P, Q, and R; xviii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, and S;
  • xlvii. A, B, C, D, E. F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU and AV;
  • AQ AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF and BG; lix.
  • Ixi A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
  • AA AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR.
  • AS AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE. BF, BG, BH, BE BJ, BK, BE, BM, BN, BO, BP and BQ;
  • BG BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR and BS;
  • BG BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW and BX;
  • BG BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX and BY;
  • BZ, CA, CB, CC, CD, CE, CF and CG are independently selected from the genes listed in Table 1 and each of A, B, C, D, E, F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN.
  • BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE, CF and CG are different.
  • each of A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, ⁇ , BY, BZ, CA, CB, CC, CD, CE, CF and CG are independently selected from the genes listed in Table 2 and each of A, B, C, D, E, F, G, H, I, J, J, K, L, M
  • each of A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE, CF and CG are independently selected from the genes listed in Table 3 and each of A, B, C, D, E, F, G, I I, I,
  • biomarkers and gene signatures of the invention may be used in methods of diagnosing, prognosing, classifying or grading COPD in biological sample or an individual.
  • One aspect of the invention provides a method of diagnosing, classi fying or grading COPD in an individual at risk for or having
  • the method comprises classifying a test sample as COPD or non-COPD. In some embodiments, the method comprises measuring the expression levels of at least 2 genes listed in Table 1 in a test sample; and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as COPD or non-COPD.
  • the expression levels of at least 2, at least 3, at least 4, at least 5, at - J J - least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1 are measured.
  • a differential pattern of expression levels of said at least 2 genes in the test sample diagnoses, classifies or grades the COPD.
  • the methods of the invention can be used to identify a gene signature and a classifier (e.g., a gene-signature-based classifier) that can distinguish datasets obtained a COPD sample from those datasets obtained from a non-COPD or healthy sample.
  • a classifier e.g., a gene-signature-based classifier
  • control data is not collected or used.
  • a classifier or a previously established standard may be used to determine whether a test sample is a COPD sample.
  • a classifier that is obtained by training with network-based or machine-learning based methods using datasets obtained from subjects with COPD and datasets from subjects without COPD can be used.
  • one or more numerical scores (e.g., average fold change or rank abs tvai as shown in Table 1 ) generated by the algorithms described herein may be used as a previously established standard.
  • the levels of expression of one or more of the genes listed in Table 1 in a test sample may be compared to the previously established standard, and the comparison may be used to classify the test sample as a COPD sample or a normal sample.
  • the invention provides a method of diagnosing COPD in a biological sample, wherein the method comprises determining the properties (for example, absence, presence or expression level) of one or more genes listed in Table 1 in the biological sample; and applying in silico analysis with a classifier obtained from a network-based method, a machine-learning based method, or a combination of the foregoing methods.
  • the classifier can be obtained from the network-based methods, the machine-learning based methods, or a combination of the foregoing methods by training with datasets obtained from subjects with COPD, subjects with COPD of a certain determined GOLD stage, or healthy subjects.
  • a classifier may be obtained, given appropriate class(es) of training datasets, to identify a specific prognosis of the
  • COPD to indicate transition from a more severe or less severe GOLD stage or to indicate that a particular treatment regimen should be used to treat the individual who provided the biological sample.
  • the methods of the invention comprise obtaining a test sample (such as bronchial brushings) from an individual, determining the absence, presence or expression level of one or more of the genes listed in Table 1 in the test sample, comparing said absence, presence or expression level to the absence, presence or expression level of the same gene(s) in a control sample, and selecting a COPD treatment regimen based on the comparison.
  • a test sample such as bronchial brushings
  • the invention provides a method for monitoring the progress of a COPD treatment in an individual, said method comprising determining at suitable time intervals before, during, or after therapy (for example, at different time points during the treatment) in a sample taken from said individual differential expression of a panel of at least 2 genes selected from the genes listed in ' fable 1 .
  • the invention encompasses a method that comprises collecting data on the properties of one or more genes in the gene signature without generating a gene signature.
  • the method of the invention comprises obtaining a test sample from an individual, and detecting the absence, presence or the expression level of one or more of the genes listed in Table 1 in the sample.
  • the invention encompasses a method that comprises using data on the properties of one or more genes in a gene signature that are already collected as training data to generate an improved gene signature using one or more network-based methods, one or more machine learning methods, or a combination of the foregoing methods.
  • the invention encompasses a method that comprises collecting data on the properties of one or more genes in a biological system which is included in a gene signature, and using the data to predict a classification of the state of the biological system associated with the collected data.
  • the method comprises detecting the expression level of at least 2 genes listed in Table 1 in a test sample obtained from the individual; and comparing the expression level of the genes listed in Table 1 in the test sample to the expression level of the genes listed in Table 1 in a control sample. In some embodiments, if the expression level of the genes listed in Table 1 is different in the test sample than in the control sample, then the individual suffers from COPD. In some embodiments, the expression level of the genes listed in Table 2 is higher in the test sample than in the control sample. Optionally, the expression level of the genes listed in Table 3 is lower in the test sample than in the control sample. In some embodiments, the method further comprises detecting the expression level of the genes listed in Table 1 in the control sample.
  • the expression levels of at least 2, at least 3 , at least 4, at least 5, at least 10, at least 1 5, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1 are detected.
  • the test sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine.
  • the test sample is obtained from a large airway of the individual, such as from a bronchial brush inserted into the large airway of the individual.
  • control sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine.
  • control sample is obtained from a large airway of an individual not affected with COPD, such as from a bronchial brush inserted into the large airway of the individual not affected with COPD.
  • control sample is obtained from the individual at risk for or having COPD prior to onset of COPD.
  • control sample is obtained from an individual that does not suffer from COPD.
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring mRNA levels.
  • mRNA level is measured by amplification, hybridization, mass spectroscopy, serial analysis of gene expression, or massive parallel signature sequencing.
  • the amplification is reverse transcription PGR, real time quantitative PGR, differential display or TaqMan PGR.
  • the hybridization is a dot blot, a slot blot, an RNase protection assay, microarray hybridization, or in situ hybridization.
  • the mass spectroscopy may be MALDI-TOF mass spectroscopy.
  • the expression level of the genes listed in Table 1 in the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1 .
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring the level of proteins encoded by the genes.
  • the protein level is measured using an antibody assay or by mass spectroscopy.
  • the antibody assay is selected from Western analysis, immunofluorescence, ELISA, and immunohistochemistry.
  • the mass spectroscopy may be MALDI-TOF mass spectroscopy or SELDI-TOF mass spectroscopy.
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring both mRNA levels and the level of proteins encoded by the genes. In some embodiments, expression levels are measured using the ampliiication, hybridization, mass spectroscopy, serial analysis of gene expression, massive parallel signature sequencing, and antibody assays discussed above.
  • the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control non-tumor biological sample are compared by in silico analysis.
  • the in silico analysis may be network based analysis or a machine-learning method. Methods of Biomarker Detection, Arrays and Panels
  • Detection of the nucleic acid and/or protein biomarkers described herein in a test sample or a control sample may be performed in a variety of ways.
  • the methods of the invention rely on the detection of the presence or absence of biomarker genes and/or biomarker gene expression, or the qualitative or quantitative assessment of either over- or under-expression of a biomarker gene in a population of cells in a test sample relative to a standard (for example, a control sample).
  • a standard for example, a control sample.
  • Such methods utilize reagents such as biomarker polynucleotides and biomarker antibodies.
  • the presence, absence or level of expression of a biomarker gene may be determined by measuring the amount of biomarker messenger RNA (mRNA), for example, by DNA-DNA hybridization, RNA-DNA hybridization, reverse transcription-polymerase chain reaction (PGR), real time quantitative PCR, differential display or TaqMan PCR; followed by comparing the results to a reference based on a control sample (for example, samples from clinically- characterized patients and/or cell lines of a known genotype/phenotype).
  • mRNA biomarker messenger RNA
  • a control sample for example, samples from clinically- characterized patients and/or cell lines of a known genotype/phenotype.
  • microRNA expression or turnover may be measured.
  • Hybridization, mass spectroscopy e.g., MALDI-TOF or SELDI-TOF mass spectroscopy
  • serial analysis of gene expression or massive parallel signature sequencing assays can also be performed.
  • hybridization assays include a singleplex or a multiplexed aptamer assay, a dot blot, a slot blot, an RNase protection assay, microarray hybridization, Southern or Northern hybridization analysis and in situ hybridization (e.g. , Fluorescent in situ hybridization).
  • these techniques find application in mi croarray-based assays that can be used to detect and quantify the amount of biomarker gene transcript using cDNA- or oligonucleotide-based arrays.
  • Microarray technology allows multiple biomarker gene transcripts and/or samples from different subjects to be analyzed in one reaction.
  • mRNA isolated from a sample is converted into labeled nucleic acids by reverse transcription and optionally in vitro transcription (cDNAs or cRNAs labelled with, for example, Cy3 or Cy5 dyes) and hybridized in parallel to probes present on an array. See, for example, Schulze et al, Nature Cell Biol.
  • the expression level of the genes listed in Table 1 in the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1.
  • biomarkers are detected using reagents that specifically detect the biomarker.
  • reagents may bind to a target gene or a target gene product (e.g., mRNA or protein), such that levels of the gene product may be quantified.
  • reagents may be nucleic acid molecules that hybridize to the mRNA or cDNA of target gene products.
  • the reagents may be molecules that label mRNA or cDNA for later detection, e.g., by binding to an array.
  • the reagents may bind to proteins encoded by the genes of interest.
  • the reagent may be an antibody or a binding protein that specifically binds to a protein encoded by a target gene of interest.
  • the reagent may label proteins for later detection, e.g., by binding to an antibody on a panel.
  • reagents are used in histology to detect histological and/or genetic changes in a sample.
  • the present invention provides isolated biomarker polynucleotides or variants thereof, which can be used, for example, as hybridization probes or primers ("biomarker probes” or “biomarker primers”) to detect or amplify nucleic acids encoding a biomarker polypeptide, particularly a biomarker polypeptide encoded by a biomarker gene or polynucleotide selected from the group depicted in Table 1 , Table 2, or Table 3.
  • biomarker probes or “biomarker primers”
  • Nucleic acid molecules comprising nucleic acid sequences encoding the biomarker polypeptides or proteins of the invention, or genomic nucleic acid sequences from the biomarker genes (e.g., intron sequences, 5 ' and 3 ' untranslated sequences), or complements thereof (i.e. , antisense polynucleotides), are collectively referred to as "biomarker genes,” “biomarker polynucleotides” or “biomarker nucleic acid sequences” of the invention.
  • biomarker polynucleotides or variants thereof which can be used, for example, as hybridization probes or primers ("biomarker probes” or “biomarker primers”) to detect or amplify nucleic acids encoding a biomarker polypeptide of the invention.
  • biomarker gene product thus
  • RNA encompasses both mRNA as well as translated polypeptide as a gene product of a biomarker.
  • the isolated biomarker polynucleotide according to the invention may comprise flanking sequences (i. e. , sequences located at the 5' or 3' ends of the nucleic acid), which naturally flank the nucleic acid sequence in the genomic DNA of the organism from which the nucleic acid is derived.
  • flanking sequences i. e. , sequences located at the 5' or 3' ends of the nucleic acid
  • the isolated biomarker polynucleotide can comprise less than about 5 kb, 4 kb, 3 kb, 2 kb, 1 kb, 0.5 kb or 0.1 kb of nucleotide sequences which naturally flank the coding sequence in genomic DNA of the cell from which the nucleic acid is derived.
  • polynucleotide is about 10-20, 21 -50, 51 -100, 101 -200, 201 -400, 401 -750, 75 1 - 1000, or 1001 - 1500 bases in length.
  • the biomarker polynucleotides of the invention are used as molecular probes in hybridization reactions or as molecular primers in nucleic acid extension reactions as described herein.
  • the biomarker polynucleotides may be referred to as biomarker probes and biomarker primers, respectively, and the biomarker polynucleotides present in a sample which are to be detected and/or quantified are referred to as target biomarker
  • Two biomarker primers arc commonly used in DNA
  • biomarker forward primer and biomarker reverse primer depending on their 5 ' to 3 ' orientation relative to the direction of transcription.
  • the invention encompasses methods of detecting genetic change in a biomarker gene (e.g., a mutation or a change in copy number). In another embodiment, the invention encompasses methods of detecting a change in the rnethvlatiosi of a biomarker gene.
  • a biomarker probe or a biomarker primer is typically an oligonucleotide which binds through complementary base pairing to a subsequence of a target biomarker polynucleotide.
  • the biomarker probe may be, for example, a DNA fragment prepared by amplification methods such as by PGR or it may be chemically synthesized. A double-stranded fragment may then be obtained, if desired, by annealing the chemically synthesized single strands together under appropriate conditions or by synthesizing the complementary strand using DNA polymerase with an appropriate primer.
  • a specific nucleic acid sequence is given, it is understood that the complementary strand is also identified and included as the complementary strand will work equally well in situations where the target is a double stranded nucleic acid.
  • a nucleic acid probe is
  • biomarkers may be detected in the test sample or the control sample by gene expression profiling.
  • mRNA is prepared from a sample and mRNA expression levels are measured by reverse transcription quantitative polymerase chain reaction (RT-PCR followed with qPCR).
  • RT-PCR is used to create a cDNA from the corresponding mRNA.
  • the cDNA may be used in a qPCR assay to produce fluorescence as the D A amplification process progresses.
  • qPCR can produce an absolute measurement such as number of copies of mRNA per cel l.
  • Northern blots, microarrays, Invader assays, and RT-PCR combined with capillary electrophoresis may be used to measure expression levels of mRNA in a sample. Further details are provided, for example, in ''Gene Expression Profiling: Methods and Protocols," Richard A. Shimkets, editor, Humana Press, 2004 and US patent application 2010/0070191 .
  • the invention encompasses an array comprising polynucleotides that hybridize to genes listed in Table 1 , Table 2, or Table 3.
  • the array may comprise polynucleotides that hybridize to at least 2, at least 3, at ' least 4, at least 5, at ' least 10, at least 1 5, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, at least 84 genes or all genes listed in Table 1 , Table 2, or Table 3.
  • the array may comprise polynucleotides that hybridize to at least 2, at least 3, at ' least 4, at least 5, at ' least 10, at least 1 5, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, at least 84 genes or all genes listed in Table 1 , Table 2, or Table 3.
  • the polynucleotides are immobilized on a solid surface.
  • solid surfaces include paper, filter, nylon or other type of membrane, slide including glass slide, and chip (e.g., silicon, microarray chip).
  • the polynucleotides may be single-stranded nucleic acid molecules (e.g., antisense oligonucleotides or iragments of cDNA).
  • the array is not a human genome- wide array. Examples of human genome-wide array include, but are not limited to, Exon 1.0 ST, Gene 1.0 ST, U 95, U 133, U 133A 2.0, and U 133 Plus 2.
  • detection of the biomarkers described herein may be accomplished by an immunoassay procedure.
  • the immunoassay typically includes contacting a test sample with an antibody that specifically binds to or otherwise recognizes a biomarker, and detecting the presence of the antibody/biomarker complex in the sample.
  • the immunoassay procedure may be selected from a wide variety of immunoassay procedures known to those skilled in the art such as, for example, competitive or non-competitive enzyme-based immunoassays, immunoprecipitation, enzyme-linked immunosorbent assays (ELiSA), radioimmunoassay (RIA), immunofluorescence, immunohistochemistry (IHC), cytological assays and Western blots.
  • multiplex assays may be used, including antibody panels or arrays, wherein several desired antibodies are placed on a support, such as a glass bead or plate, and reacted or otherwise contacted with the test sample or the control sample.
  • Antibodies used in these assays may be monoclonal or polyclonal, and may be of any type such as IgG, IgM, IgA, IgD and IgE. Monoclonal antibodies may be used to bind to a specific epitope offered by the biomarker molecule, and therefore mav provide a more specific and accurate result. Antibodies may be produced by immunizing animals such as rats, mice, rabbits and goats. The antigen used for immunization may be isolated from the samples or synthesized by recombinant protein technology. Methods of producing antibodies and of performing antibody-based assays are well-known to the skilled artisan and are described, for example, more thoroughly in Antibodies: A Laboratory Manual (1988) by Harlow & Lane; Immunoassays: A Practical Approach, Oxford
  • the present invention also provides "biomarker antibodies” including polyclonal, monoclonal, or recombinant antibodies, and fragments and variants thereof, that immunospecifically bind the respective biomarker proteins or polypeptides encoded by the genes or cDNAs (including polypeptides encoded by mRNA splice variants) as listed in Tables 1 , 2, and 3.
  • biomarker antibodies including polyclonal, monoclonal, or recombinant antibodies, and fragments and variants thereof, that immunospecifically bind the respective biomarker proteins or polypeptides encoded by the genes or cDNAs (including polypeptides encoded by mRNA splice variants) as listed in Tables 1 , 2, and 3.
  • Various chemical or biochemical derivatives of the antibodies or antibody fragments of the present invention can be produced using known methods.
  • One type of derivative which is diagnostically useful as an immunoconjugate comprising an antibody molecule, or an antigen-binding fragment thereof, to which is conjugated a detectable label.
  • the biomarker antibody is not labeled but in the course of an assay, it becomes indirectly labeled by binding to or being bound by another molecule that is labeled.
  • the invention encompasses molecular complexes comprising a biomarker antibody and a label , as well as immunocomplexes comprising a biomarker polypeptide, a biomarker antibody, and immunocomplexes comprising a biomarker polypeptide, a biomarker antibody, and a label.
  • Examples of detectable substances or detectable labels include various enzymes, prosthetic groups, fluorescent materials, luminescent materials, bioluminescent materials, and radioactive materials.
  • suitable enzymes include horseradish peroxidase, alkaline phosphatase, beta-galactosidase and acetylcholinesterase.
  • suitable prosthetic group complexes include streptavidin/biotin and avidin/biotin.
  • fluorescent materials include umbelliferones, fluoresceins, fluorescein isothiocyanate, rhodamines, dichlorotriazinylarnine fluorescein, dansyl chloride, phycoerythrins, Alexa Fluor 647, Alexa Fluor 680, DilCi 9 (3), Rhodamine Red-X, Alexa Fluor 660, Alexa Fluor 546, Texas Red, YOYO- 1 + DNA, tetramethylrhodamine, Alexa Fluor 594, BODIPY FL, Alexa Fluor 488, Fluorescein, BODIPY TR, BODiPY TMR, carboxy SNARF- 1 , FM 1 -43 , Fura-2, Indo- 1 , Cascade Blue, NBD, DAPI, Alexa Fluor 350, aminomethylcoumarin, Lucifer yellow, Propidium iodide, or dansylamide.
  • An example of a luminescent material is lumino
  • radioactive material include I, I, S or 1 1.
  • Immunoassays for biomarker polypeptides will typically comprise incubating a sample, such as a biological fluid, a tissue extract, freshly harvested cells, or lysates of cells, in the presence of a detectably labeled antibody capable of identifying biomarker gene products or conserved variants or peptide fragments thereof, and detecting the bound antibody by any of a number of techniques well- known in the art.
  • a sample such as a biological fluid, a tissue extract, freshly harvested cells, or lysates of cells
  • a detectably labeled antibody capable of identifying biomarker gene products or conserved variants or peptide fragments thereof
  • ELISA enzyme-linked immunosorbent assay
  • the enzyme either conjugated to the antibody or to a binding partner for the antibody, when later exposed to an appropriate substrate, will react with the substrate in such a manner as to produce a chemical moiety which can be detected, for example, by spectrophotometric, or fluorimetric means.
  • the biological sample may be brought in contact with and immobilized onto a solid phase support or carrier such as nitrocellulose, or other solid support which is capable of immobilizing cells, cell particles or soluble proteins.
  • a solid phase support or carrier such as nitrocellulose, or other solid support which is capable of immobilizing cells, cell particles or soluble proteins.
  • the support may then be washed with suitable buffers followed by treatment with the detectably labeled biomarker antibody.
  • the solid phase support may then be washed with the buffer a second time to remove unbound antibody.
  • the amount of bound label on solid support may then be detected by conventional means.
  • a well known example of such a technique is Western blotting.
  • the present invention provides compositions comprising labelled biomarker polynucleotides, or labelled biomarker antibodies to the biomarker proteins or polypeptides, or labeled biomarker polynucleotides and labeled biomarker antibodies to the biomarker proteins or polypeptides according to the invention as described herein.
  • Antibodies and other reagents may also be used to detect post- translational modifications (e.g., methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycosylation, sugar modification, lipidation, lipid modification, ubiquitination, sumolation, disulphide bonding, cysteinylation, oxidation, glutathionylation, carboxylation,
  • post- translational modifications e.g., methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycosylation, sugar modification, lipidation, lipid modification, ubiquitination, sumolation, disulphide bonding, cyst
  • biomarker proteins or biomarker
  • the invention encompasses a panel comprising antibodies that bind to proteins encoded by genes listed in Table 1 , Table 2 or Table 3.
  • the panel may comprise antibodies that bind to proteins encoded by at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, at least 84 genes or all genes listed in Table 1 , Table 2 and Table 3.
  • the panel of antibodies is immobilized on a solid surface. Examples of solid surfaces include microspheres, plates, wells, slides, and beads (e.g., protein A or protein G agarose).
  • the biomarkers described herein may also be detected and quantified by mass spectrometry.
  • Mass spectrometry is a method that employs a mass spectrometer to detect ionized protein markers or ionized peptides as digested from the protein markers by measuring the mass-to- charge ratio (m/z). Labelling of biomarkers (along with other proteins) with stable heavy isotopes (deuterium, carbon- 13, nitrogen- 1 5, and oxygen- 18) can be used in quantitative proteomics. These are either incorporated metabolically in sample cells cultured briefly in vitro, or directly in samples by chemical or enzymatic reactions.
  • Biomarker peptide ions segregate and their intensity values are used for quantification.
  • analytes may be introduced into an inlet system of the mass spectrometer and ionized in am ionization source, such as a laser, fast atom bombardment, plasma or other suitable ionization sources known to the art.
  • am ionization source such as a laser, fast atom bombardment, plasma or other suitable ionization sources known to the art.
  • the generated ions are typically collected by an ion optic assembly and introduced into mass analyzers for mass separation before their masses are measured by a detector. The detector then translates information obtained from the detected ions into mass-to-charge ratios.
  • the invention also encompasses methods that involve measuring the activity of a biomarker (e.g., enzymatic activity).
  • a biomarker e.g., enzymatic activity
  • enzymatic activity include, without limitation, kinase, phosphatase, protease, ubiquitination, oxidase and reductase activity.
  • the invention also provides compositions comprising biomarker polynucleotides, biomarker polypeptides, or biomarker antibodies according to the invention as described herein in the various embodiments.
  • the invention further provides diagnostic or detection reagents for use in the methods of the invention, for example, reagents for flow cytometry and/or immunoassays that comprise fluorochrome-labeled antibodies that bind to one of the biomarker polypeptides of the invention.
  • the invention provides diagnostic or detection reagents that comprise one or more biomarker probes, or one or more biomarker primers.
  • a diagnostic reagent may comprise biomarker probes and/or biomarker primers from the same biomarker gene or from multiple biomarker genes.
  • the invention also provides diagnostic compositions that comprise one or more biomarker probes and target biomarker polynucleotides, or one or more biomarker primers and target polynucleotides, or biomarker primers, biomarker probes and biomarker target polynucleotides.
  • the diagnostic compositions comprise biomarker probes and/or biomarker primers and a sample suspected to comprise biomarker target polynucleotides.
  • Such diagnostic compositions comprise biomarker probes and/or biomarker primers and the nucleic acid molecules (including NA, mRNA, cRNA, cDNA, and/or genomic DNA) of a subject in need of a diagnosis/prognosis of COPD.
  • biomarker probes and/or biomarker primers and the nucleic acid molecules (including NA, mRNA, cRNA, cDNA, and/or genomic DNA) of a subject in need of a diagnosis/prognosis of COPD.
  • Biomarkers and gene signatures of the invention may be predicted based on gene expression patterns in COPD.
  • biomarker and gene signature prediction comprises gene expression patterns in control (e.g., non- COPD) biological samples.
  • a heterogeneous ensemble learning approach may be used to classify samples based on their gene expression profiles. Such an approach may combine predictions from different approaches that use genes, gene set- derived features and/or causal network-derived features in order to get a classification and a prediction confidence for each classified sample.
  • Methods that may be used to generate biomarkers and gene signatures of the invention include shrunken centroids, factor rotation, logistic regression models, network-based approaches, disease module-based approaches, linkage methods, modularity or pathway-based methods and diffusion-based methods.
  • the biological data (such as training data and test data) used in these methods may be drawn from the literature, databases (including data from preclinical, clinical and post-clinical trials of pharmaceutical products or medical devices), genome databases (genomic sequences and expression data, e.g. , Gene Expression Omnibus by National Center for Biotechnology Information or ArrayExpress by European Bioinformatics Institute (Parkinson et al . 2010, Nucl. Acids Res., doi : 10.1093/nar/gkql 040. Pubmed ID 21071405)), commercially available databases (e.g., Gene Logic, Gaithersburg, MD, USA) or experimental work.
  • databases including data from preclinical, clinical and post-clinical trials of pharmaceutical products or medical devices
  • genome databases genome databases (genomic sequences and expression data, e.g. , Gene Expression Omnibus by National Center for Biotechnology Information or ArrayExpress by European Bioinformatics Institute (Parkinson et al . 2010, Nucl. Acids Res., doi : 10.109
  • the REACTOME, EGG or BIOCARTA pathway gene set collections from the Broad Institute may be used.
  • the data may be related to nucleic acid (e.g. , absolute or relative quantities of specific DNA or RNA species, changes in DNA sequence, RNA sequence, changes in tertiary structure, or methyl ation pattern as determined by sequencing,
  • hybridization particularly to nucleic acids on microarray, quantitative polymerase chain reaction, or other techniques known in the art
  • protein/peptide e.g. , absolute or relative quantities of protein, specific fragments of a protein, peptides, changes in secondary or tertiary structure, or posttranslational modifications as determined by methods known in the art
  • functional activities e.g., enzymatic activities, proteolytic activities, transcriptional regulatory activities, transport activities, binding affinities to certain binding partners
  • Modifications including posttranslational modifications of protein or peptide, can include, but are not limited to, methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycosylation, sugar modification, lipidation, lipid modification, ubiquitination, sumolation, disuiphide bonding, cysteinylation, oxidation, glutathionylation, carboxylation,
  • a protein can be modified posttranslationally by a series of reactions such as Amadori reactions, Schiff base reactions, and Maillard reactions resulting in glycated protein products.
  • test data sets may be processed and have their quality controlled together if they are obtained from the same technology platform (e.g., an
  • raw data files may be read by the ReadAffy function of the affy package (Gautier et al., Bioinformatics, 20:307-3 1 5 (2004)) belonging to Bioconductor (Gentleman et al., Genome Biol, 5(1 0):R80 (2004)) in R (R Development Core Team, R: A Language and Environment for Statistical Computing. 2007).
  • the quality may be controlled by: 1.
  • RNA degradation plots using the AffyRNAdeg function of the affy package (Gautier, 2004)
  • NUSE and RLE plots using the function affyPLM (Brettschneider et al., Technometrics, 50(3):241 -264 (2008)), calculating the MA(RLE) values;
  • Arrays passing quality control checks may be normalized using the gcrma algorithm (Wu et al., Journal of the American Statistical Association, 99:909
  • the samples classification may be obtained from the series matrix file of the same database for each dataset.
  • the output of this part of the method may consist of: a gene expression matrix on training samples and test samples, probesets, and the class information for the training samples.
  • Non-limiting examples of methods that may be used to generate predictions are: transformation invariant (Tranlnv) (U.S. Provisional Patent Application entitled “Systems and Methods for Generating Biomarker Signatures with Integrated Bias Correction and Class Prediction,” filed concurrently with the instant application and having the attorney docket no.
  • Generalized simulated annealing may be modified for binary functions.
  • a dual binary generalized simulated annealing based method may be used (DualGensemble) (U.S. Provisional Patent Application entitled “Systems and Methods for Generating Biomarker Signatures with Integrated Dual Ensemble and Simulated Annealing Techniques," filed concurrently with the instant application, incorporated herein by reference in its entirety and having the attorney docket no. 106500-0031 -001 ).
  • T-filter is a method of filtering genes based on the t-test by setting P-value and fold-change thresholds.
  • CORG may be modified by calculating activity scores by leveraging the F-test instead of the T- test. CORG may also be combined with SVM.
  • Dual bagging is a combination of bagging (Breiman, The Elements of Statistical Learning: Data Mining, Inference, and Prediction. 2nd ed., ed. T. Hastie, R. Tibshirani, and J. Friedman, (2009)) and the random subspace method (Bryll, Pattern Recognition, 20(6): 1291 - 1302 (2003) Ho, IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(8):832- 844 ( 1998); Skurichina, Pattern Analysis and Applications, 5(2): 121 - 135 (2002)).
  • the single and pairs method may include the following steps:
  • the forward learning method may include the following steps:
  • N for example, 20, 100 or 200.
  • n 1 , . . . N do a.
  • ntree 500
  • TPr true positive rates
  • TNr true negative rates
  • the Laplacian based learning method may include the following steps:
  • KPC main kernel principal components
  • Network-based analysis can be combined with machine learning methods to generate predictions, for example, combining any one of CORG, dual bagging or T-filter with a network-based analysis.
  • methods used to generate predictions are further combined with another classification method (e.g., a method that is used for cross- validation).
  • classification methods include PAMR (Tibshirani et al, Proc Natl Acad Sci USA, 99(10):6567-6572 (2002)),
  • LDA Discrimination Analysis
  • ELDA Eigengene-based Linear Discrimination Analysis
  • PCA Principal Components Analysis
  • RPART Recursive Partitioning Tree
  • SVM Support Vector Machine
  • KNN K-Nearest Neighbors
  • PLS.DA Partial Least Squares Discriminant Analysis
  • NPA Network-based Analysis that uses NPA may be combined with SVM (U.S. Provisional Patent Application entitled “Systems and Methods Relating to Network-Based Biomarker Signatures,” filed concurrently with the instant application, incorporated herein by reference in its entirety and having the attorney docket no. 106500-0022-001.
  • these methods may further include a step of oversampling to balance classes.
  • the methods may include a step of filtering genes based on a simple T-test between the categories to be classified.
  • the filtering step may reduce the number of genes to less than 1 ,500 or less than 2,000.
  • the genes obtained by these methods may be mapped to gene symbols using any suitable platform, for example, the Confero platform (Hermida et al., Confer o: an integrated Contrast and Gene Set Platform for Computational Analysis and Biological Interpretation of Omics Data, submitted, 2012).
  • the Confero platform Hermida et al., Confer o: an integrated Contrast and Gene Set Platform for Computational Analysis and Biological Interpretation of Omics Data, submitted, 2012.
  • the numerical methods for generating the gene signatures of the invention may include a testing step and confidence statistics for the genes.
  • the testing step (or phase) is an exemplary use of the gene signature in carrying out the embodied method.
  • the invention encompasses a method for classifying a test sample as a COPD sample or a non-COPD sample, the method comprising: measuring the expression levels of at least 2 genes listed in Table 1 in a test sample; and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as cither a COPD sample or a non-COPD sample.
  • the classifier has been trained by in silico analysis or one or more feature selection and classification algorithms.
  • the computer program product may include a classifier based on at least two genes listed in Table 1.
  • the classifier may be based on at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50. at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, least about 85, at least 87 or all genes listed in Table 1.
  • the classifier is trained by in silico analysis or one or more feature selection and classification algorithms.
  • the classifier is trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T-filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude.
  • the classifier may be trained with at least the data in Gene Expression Omnibus datasets GSE10106, GSE10135, GSE 1 1906, GSE 1 1 952, GSE13933, GSE19407, GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545. Devices and Kits
  • One aspect of the invention encompasses devices useful for performing methods of the invention.
  • the devices may be used for diagnosing, classifying and/or grading COPD.
  • the devices can comprise means for detecting the expression level of at least 2 genes listed in Table 1 or the level of at least 2 gene products of such genes in a test sample.
  • Such means may include
  • Such components may include one or more of an amplification chamber (for example a thermal cycler), a plate reader, robotic sample handling
  • the devices can comprise means for detecting the expression levels of at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1 .
  • the devices can comprise means for detecting the expression levels of the gene products of at least 2, at least 3, at least 4, at least 5, at least 1 0, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1.
  • the devices optionally comprise a means for identi fying a given test sample, and of linking the results obtained to that sample.
  • Such means can include manual labels, barcodes, and other indicators which can be linked to a sample container or receptacle.
  • Identification means may optionally be included in the sample itself, for example where an encoded particle is added to the sample.
  • the results may be linked to the sample, for example in a computer memory that contains a sample designation and a record of expression levels obtained from the sample. Linkage of the results to the sample can also include a linkage to a particular sample container or receptacle in the device, which is also linked to the sample identity.
  • the devices may comprise an excitation and/or a detection means.
  • Any instrument that provides a wavelength that can activate a label (e.g., fluorophore, fluorochrome and fluorescent dye) used on a detection reagent and is shorter than the emission wavelength(s) to be detected can be used for excitation.
  • excitation sources include a broadband ultraviolet light source such as a deuterium lamp with an appropriate filter, the output of a white light source such as a xenon lamp or a deuterium lamp after passing through a monochromator to extract out the desired wavelcngth(s), a continuous wave (cw) gas laser, a solid state diode laser, or any pulsed lasers.
  • Emitted light can be detected through any suitable component or technique; many suitable approaches are known in the art.
  • a fluorimeter or spectrophotometer may be used to detect whether the test sample emits light of a wavelength characteristic of a label used in a method of the invention.
  • the devices may comprise a means for correlating the expression levels of the genes being analyzed with COPD status, prognosis, grade and/or
  • Such means may comprise one or more of a variety of correlative techniques, including lookup tables, algorithms, multivariate models, and linear or nonlinear combinations of expression models or algorithms, such as any of the in silico and machine learning methods described above.
  • the expression levels may be converted to one or more biomarker scores, indicating that the individual providing the sample is not suffering from COPD or is suffering from stage 1 , stage 2, stage 3 or stage 4 COPD.
  • the models and/or algorithms can be provided in computer readable format.
  • the devices may also comprise output means for outputting the COPD status, prognosis, grade and/or classification.
  • output means can take any form which transmits the results to an individual and/or a healthcare provider, and may include a monitor, a display, and/or a printer.
  • Output means may record the results to a computer readable medium.
  • the device may use a computer system for performing one or more of the steps provided.
  • kits for practicing the methods of the invention may be used for diagnosing, classifying and grading COPD or for assessing the prognosis of COPD in an individual.
  • the kits can be used for ciinical diagnosis and/or laboratory research.
  • a kit comprises in one or more containers one or more reagents that detect expression levels of genes that serve as biomarkers of COPD in a test sample.
  • the kit also comprises instructions in any tangible medium (e.g., written, tape, CD-ROM, DVD) on the use of the detection reagent(s) in one or more methods of the invention.
  • a detection reagent in the kit may comprise at least one polynucleotide, probe, and/or primer specific for the COPD genes listed in Table 1 .
  • the nucleic-acid based detection reagents may comprise sequences complementary to a portion of the signature genes or sequences that are portions of the signature genes.
  • Such a kit may optionally provide in separate containers enzymes and/or buffers for reverse transcription, in vitro transcription, and/or DNA polymerization, nucleotides, and/or labeled nucleotides.
  • a detection reagent in the kit may comprise a biomarker antibody, which may be labeled or labelable.
  • the antibodies may bind to proteins encoded by the COPD genes listed in Table 1 .
  • the detection reagents recognize a post-translational modification (e.g., methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycosylation, sugar modification, lipidation, lipid modification, ubiquitination, sumolation, disulphide bonding, cysteinylation, oxidation, glutathionylation, carboxylation, giucuronidation, and deamidation) of a protein encoded by a gene selected from the genes listed in Table 1 .
  • a post-translational modification e.g., methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycos
  • the kit may include a substrate for the biomarker and a detection reagent that recognizes the products and/or byproducts of the activity being measured.
  • a detection reagent that recognizes the products and/or byproducts of the activity being measured.
  • Such a kit may optionally provide, in separate containers, buffers, secondary antibodies, signal generating accessory molecules, and/or labeled secondary antibodies, including fluorochrome-labeled secondary antibodies.
  • the kit may also include unlabeled or labeled antibodies to various cell surface antigens which can used for identification or sorting of subpopulations of cells.
  • the detection reagents may be labeled or labelable by one or more detectable labels.
  • detectable labels include, without limitation, radiolabels (e.g. radioactive nuclides), dyes, fluorescent proteins or materials (e.g., fluorochromes, fluorophores, fluorescein and rhodamine), luminescent proteins or materials, bioluminescent proteins or materials (e.g., luciferase, aequorin and luciferin), enzymes (e.g., beta-galactosidase, alkaline phosphatase, horseradish peroxidase and acetylcholinesterase) and prosthetic groups (e.g., biotin, streptavidin and avidin).
  • radiolabels e.g. radioactive nuclides
  • dyes e.g., fluorescent proteins or materials (e.g., fluorochromes, fluorophores, fluorescein and rhodamine), luminescent proteins or
  • the detection reagents in the kit may be immobilized on a solid surface or packaged separately with reagents to immobilize them on a solid surface.
  • kits may be positive and negative controls for the methods of the invention.
  • the positive and/or negative controls included in a kit can be nucleic acids, polypeptides, ceil lysate, cell extract, whole cells from patients, or whole cells from cell lines.
  • RNA degradation plots with the AffyRNAdeg function of the a fly package
  • NUSE and RLE plots with the function affyPLM
  • Training set sample classifications are obtained from the series matrix file of the GEO database for each dataset.
  • GSM252824.CEL GSM252826.CEL, GSM298246.CEL, GSN 252819.CEL, GSM252821 .CEL, GSM252842.CEL, GSM252843.CEL, GSM252844.CEL, GSM252845.CEL, GSM252846.CEL, GSM252847.CEL, ,GSM252848.CEL, GSM252849.CEL, GSM252850.CEL, GSM252851 .CEL, GSM252852.CEL, GSM252853.CEL, GSM252854.CEL, GSM1 14096.CEL, GSM 1 14098. CEL, GSM1 14099.
  • INPUT gene expression matrix X nxp on n samples and p genes, training samples and test samples, and the class information for the training samples
  • OUTPUT Class prediction for each test sample and a list of genes involved. [0136] Prior to applying features selection and classification methods, the following steps are performed: (1 ) oversampling is, optionally, used to balance classes in the training dataset; (2) mapping probe sets to gene symbols (Entrcz gene ids) using Confero platform (Hermida, 2012); and (3) optionally filtering the genes in the matrix based on simple T-test between the categories to be classified so that less than 1500 genes (for Dual Ensemble or T-filter methods) or less than 2000 genes (for the other methods) remain.
  • This dual ensemble method builds ensemble of multiple classi fication algorithms applied in randomly perturbed data.
  • the diversity of the ensemble classifier is imposed by using different classification algorithms and is further enhanced by data-level perturbation. See, e.g., Yang, 201 0.
  • a molecular profile of a training dataset, TO. train and its associated phenotype cl. train (control and treatment) are used as input.
  • the molecular profile of the test set TO. test is used to predict the phenotype cl.tcst.
  • Genes are filtered based on t-test to obtain a list of N genes, by setting P- value and fold-change thresholds. Thresholds are decreased (resp. increase) automatically if the list size is over N. Any M is trained on the resulting subspace, N is chosen by cross-validation,
  • This method is modified from CORG method (Chuang, 2007) as activity scores are calculated by leveraging F-test instead of T-test. It uses the c2.cp gene sets collection from the Broad Institute (Cambridge, MA) (Reactome, KEGG and Biocarta pathways). d. Single and Pairs method
  • a threshold (ao) is selected.
  • the obtained list is used to train M on the reduced feature space.
  • ntree 500
  • TPr true positive rates
  • TNr true negative rates
  • the union of the gene signatures extracted from the results of the foregoing methods is considered as the larger gene signature.
  • a weight is given to each of the genes in the union of the gene signature on the basis of the number of times it appears in a generated signature and the length of each generated gene signature.
  • Genes in Table 1 are those that appear in at least 3 of 0 5 lists generated.
  • a further reduction of the number of genes in the signature can be obtained by selecting the genes which has a fold-change value greater than a threshold (see Table 1 , the column "fc>2" wherein a value of 1 is ascribed to those genes that has a fold change value greater than 2).
  • the genes in Table 1 that appear in at least 4 of the 5 lists generated are more predictive of lung cancer status than those appearing in 3 of the 5 lists generated.
  • Table 1 Genes which are specifically regulated in airway brushings of COPD patients compared with control airway brushings
  • Table 2 Genes which are specifically upregulated in airway brushings of COPD patients compared with control airway brushings
  • Table 3 Genes which are specifically downregulated in airway brushings of COPD patients compared with control airway brushings

Abstract

The present invention relates to biomarkers and gene signatures that are useful for diagnosing, classifying and prognosing COPD. The invention also relates to devices, arrays, panels, diagnostic methods and kits using these biomarkers and gene signatures.

Description

Gene Signatures for COPD Diagnosis Field of the invention
[0001] The present invention relates to gene signatures that are useful for the diagnosis of Chronic Obstructive Pulmonary Disease (COPD). The present invention also relates to methods of diagnosing COPD. The invention further relates to arrays and computer readable media comprising gene signatures for diagnosis of COPD.
Background of the Invention
[0002] Chronic Obstructive Pulmonary Disease (COPD) is the third leading cause of death in the U.S. and is projected to be the fourth leading cause of death worldwide by 2030. In Europe, over€20 billion is spent treating COPD, while annmyirnatplv ¾ 1 8 hi l linn i ¾ «npiif in fhp T T *s
[0003] COPD is a progressive disease that affects an individual's ability to breathe over time and includes two main conditions: emphysema and chronic bronchitis. In COPD, the airways in the lung chronically narrow, limiting the amount of airflow through the airways. Individuals affected by COPD experience decreased elasticity in their airways and air sacs, the destruction of the walls between air sacs, the inflammation of the walls of the airways, and the clogging of airways by increased production of mucus.
[0004] COPD is caused by long-term exposure to lung irritants in the environment, such as air pollution, chemical fumes, dust and tobacco smoke. Because long-term exposure to lung irritants is typically required, most individuals who suffer from COPD are more than 40 years old when symptoms first present. Such symptoms include an ongoing cough or a cough that produces an excess of mucus; shortness of breath, especially with physical activity; wheezing and chest tightness. COPD symptoms slowly worsen over time, and most affected individuals do not notice symptoms at first because they are so mild or easy to correct by lifestyle adjustment.
[0005] COPD is typically diagnosed by signs and symptoms, including medical history, family history and test results. Diagnostic tests for COPD include lung function tests, such as spirometry, lung volume measurement, and lung diffusion capacity; chest X-rays, chest CT scans and arteriole blood gas tests. Accordingly, the diagnostic tests for COPD require the disease to have progressed to the point that lung function is moderately affected. Thus, there is a need for a diagnostic test that can identify COPD in patients at early stages. Recently, there have been attempts to identify genes associated with COPD or the treatment thereof. See, e.g. , US 201 1/0160070; US 2010/01 19474; and US 2009/0186951 . A need still persists to understand the molecular mechanisms of COPD, which may allow for the design or optimization of therapies to treat the disease, instead of just the symptoms.
[0006] The present invention is directed to gene signatures for classifying, diagnosing or grading Chronic Obstructive Pulmonary Disease (COPD). A first aspect of the invention provides a method of diagnosing, classifying or grading COPD in an individual at risk for or having COPD. In some embodiments the method comprises classifying a test sample as COPD or non-COPD. in some embodiments, the method comprises measuring the expression levels of at least 2 genes listed in Table 1 in a test sample: and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as COPD or non-COPD. In some embodiments, a differential pattern of expression levels of said at least 2 genes in the test sample diagnoses, classifies or grades the COPD.
[0007] In some embodiments, the differential pattern of expression levels is identified by a classifier based on a plurality of genes listed in Table 1 , including said at least two genes, said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms. Optionally, the differential pattern of expression levels is identified by a classifier based on a pl ural ity of genes listed in Table 1 , including said at least two genes, said classifier having been trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T-filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude. For example, the classifier may be trained with at least the data in the Gene Expression Omnibus datasets GSE10106, GSE10135, GSE1 1906, GSE1 1952, GSE13933, GSE19407, GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545.
[0008] In some embodiments, the method comprises detecting the expression level of at least 2 genes listed in Table 1 in a test sample obtained from the individual; and comparing the expression level of the genes listed in Table 1 in the test sample to the expression level of the genes listed in Table 1 in a control sample. In some embodiments, if the expression level of the genes listed in Table 1 is different in the test sample than in the control sample, then the individual suffers from COPD. In some embodiments, the expression level of the genes listed in Table 2 is higher in the test sample than in the control sample. Optionally, the expression level of the genes listed in Table 3 is lower in the test sample than in the control sample. In some embodiments, the method further comprises detecting the expression level of at least 2 genes listed in Table 1 in the control sample.
[0009] In some embodiments, the at least 2 genes are selected from the group consisting of: PROS l , IRAKI , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR IA, PAPD4, VCL, PPP2R5C, DGKA, and CYP51 A1 .
[0010] In some embodiments, the test sample or control sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine. Optionally, the test sample is obtained from a large airway of the individual, such as from a bronchial brush inserted into the large airway of the individual. In some embodiments, the control sample is obtained from a large airway of an individual not affected with COPD, such as from a bronchial brush inserted into the large airway of the individual not affected with COPD. In some embodiments, the control sample is obtained from the individual at risk for or having COPD prior to onset of COPD. In other embodiments, the control sample is obtained from an individual that does not suffer from COPD.
[0011] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring mRNA levels. Optionally, the expression level of the genes listed in Table 1 the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1 and said at least 2 genes.
[0012] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring the level of proteins encoded by the genes.
[0013] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring both mRNA levels and the level of proteins encoded by the genes.
[0014] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control biological sample are compared by in silica analysis (e.g., network-based analysis or machine-learning methods).
[0015] A second aspect of the invention provides an array for use in diagnosing or prognosing COPD. In some embodiments, the array comprises polynucleotides immobilized on a solid surface that can hybridize to at least 2 signature genes, wherein the COPD signature genes are selected from the group consisting of the genes listed in Table 1. Optionally, the array comprises polynucleotides hybridizing to at least 2 lung cancer signature genes immobilized on a solid surface, wherein the lung cancer signature genes are selected from the genes listed in Table 1 . In some embodiments, the array is not a human genome-wide array.
[0016] A third aspect of the invention provides a panel for use in diagnosing or prognosing COPD. In some embodiments, the panel comprises antibodies immobilized on a solid surface that bind to proteins encoded by at least 2 COPD signature genes, wherein the COPD signature genes are selected from the group consisting of the genes listed in Table 1.
[0017] A fourth aspect of the invention provides a computer readable medium or a computer program for use in diagnosing or prognosing COPD. in some embodiments, the computer readable medium or computer program comprises a COPD gene signature, wherein the gene signature comprises at least 2 genes selected from the genes listed in Table 1.
[0018] In some embodiments, the computer readable medium or computer program product comprises a classifier based on at least two genes listed in Table 1 , said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms. Optionally, the classifier is trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T-fi!ter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude. The classifier may be trained with at least the data in the Gene Expression Omnibus datasets GSE10106, GSE10135, GSE1 1906, GSE1 1952, GSE13933, GSE19407, GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545. In some
embodiments, the at least two genes are selected from the group consisting of PROS 1 , IRAKI , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR1A, PAPD4, VCL, PPP2R5C, DGKA, and CYP51A1.
[0019] A fifth aspect of the invention provides a device for diagnosing or prognosing COPD. In some embodiments, the device comprises means for detecting the expression level of the genes listed in Table 1 in a test sample; means for correlating the expression level with a diagnosis or prognosis of the COPD; and means for outputting the COPD diagnosis or prognosis. Optionally, the device further comprises means for detecting the expression level of the genes listed in Table 1 in a control sample.
[0020] A sixth aspect of the invention provides a kit for diagnosing or prognosing COPD. In some embodiments, the kit comprises a set of reagents that detects expression levels of the genes listed in Table 1 in a test sample and instructions for using said kit for diagnosing the COPD. In other embodiments, the kit is for assessing the prognosis of COPD in an individual. In such embodiments, the kit comprises a set of reagents that detects expression levels of the genes listed in Table 1 in a test sample from the individual and instructions for using said kit
- ιcυ„ι„ ucicmuuuig mc pi uguu„: S„ u„-if -t uic„
Figure imgf000007_0001
: m„ ¾„t„u:u J : m„uJi:»v,;iu Λ—um i . τ i„n aumc
embodiments, the set of reagents that detects expression levels of the genes listed in Table 1 in the test sample may also be used to detect expression levels of the genes listed in Table 1 in a control sample.
[0021] An seventh aspect of the invention provides a method of diagnosing COPD n an individual or of assessing the prognosis of an individual with COPD. In some embodiments, the method comprises a) measuring the expression level of at least 2 genes/biomarkers selected from the group consisting of the genes listed in Table 1 in a biological sample obtained from the individual; b) calculating a numerical biomarker score for the individual based on the expression levels of the biomarkers measured in step a); wherein the numerical biomarker score is predictive of the diagnosis of COPD in the individual. In some embodiments, the method comprises measuring the expression level of at least 2 genes/biomarkers selected from the group consisting of the genes listed in Table 1 in a biological sample obtained from the individual; calculating a numerical biomarker score for the individual based on the expression levels of the biomarkers measured in step a); wherein the numerical biomarker score is predictive of the prognosis of COPD in the individual.
[0022] In some embodiments, the biological sample is selected from blood,
ui
urine. Optionally, the tissue is lung tissue, such as tissue obtained by biopsy from a tumor.
[0023] In some embodiments, the expression level of the genes listed in Table 1 in the biological sample is detected by measuring mRNA levels. Optionally, the expression level of the genes listed in Table 1 the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1 and said at least 2 genes. [0024] In some embodiments, the expression level of the genes listed in Table 1 in the biological sample is detected by measuring the level of proteins encoded by the genes.
[0025] In some embodiments, the expression level of the genes listed in Table 1 in the biological sample is detected by measuring both mRNA levels and the level of proteins encoded by the genes.
[0026] In some embodiments, the numerical biomarker score is calculated by in siiico analysis. The in silico analysis may be network based analysis or machine-learning methods.
[0027] In some embodiments, the biomarkers are proteins encoded by the genes selected from the group consisting of the genes listed in Table 1 .
[0028] Particular embodiments of the invention are set forth in the following numbered paragraphs:
1 . A method of diagnosing an individual as being at risk for or having Chronic Obstructive Pulmonary Disease (COPD) comprising
(1 ) detecting the expression level of at least 2 genes listed in Table 1 in a test sample obtained from the individual; and
(2) comparing the expression level of said at least 2 genes in the test sample to the expression level of said at least 2 genes in a control sample,
wherein, if the expression level of said at least 2 genes is different in the test sample than in the control sample, then the individual suffers from COPD.
2. The method according to paragraph 1 , wherein the expression level of the genes listed in Table 2 is higher in the test sample than in the control sample.
3. The method according to paragraph 1 or 2, wherein the expression level of the genes listed in Table 3 is lower in the test sample than in the control sample.
4. The method according to any one of paragraphs 1 -3, wherein the method further comprises detecting the expression level of said at least 2 genes in the control sample.
5. The method according to any one of paragraphs 1 -4, wherein the test sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine. 6. The method according to paragraph 5, wherein the test sample is obtained from a large airway of the individual.
7. The method according to paragraph 6, wherein the test sample is obtained from a bronchial brush inserted into the large airway of the individual.
8. The method according to any one of paragraphs 1 -7, wherein the control sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine.
9. The method according to paragraph 8, wherein the control sample is obtained from a large airway of an individual not affected with COPD.
1 0. The method according to paragraph 9, wherein the control sample is obtained a bronchial brush inserted into the large airway of the individual not affected with COPD.
1 1. The method according to paragraph 8, wherein the control sample is obtained from the individual at risk for or having the COPD prior to onset of the COPD.
12. The method according to paragraph 8, wherein the control sample is obtained from an individual that does not suffer from COPD.
13. The method according to any one of paragraphs 1 - 1 1 , wherein the expression level of said at least 2 genes in the test sample and the expression level of said at least 2 genes in the control sample are detected by measuring rJ NA levels.
14. The method according to paragraph 13, wherein the mRNA level is measured by amplification, hybridization, mass spectroscopy, serial cinalysis oi gene expression, or massive parallel signature sequencing.
15. The method according to paragraph 14, wherein the amplification is reverse transcription PGR, real time quantitative PGR, differential display or TaqMan PGR.
16. The method according to paragraph 14, wherein the hybridization is a dot blot, a slot blot, an RNase protection assay, microarray hybridization, or in situ hybridization.
17. The method according to paragraph 14, wherein the mass spectroscopy is MALDI-TOF mass spectroscopy. 1 8. The method according to any one of paragraphs 1 - 1 1 , wherein the expression level of said at least 2 genes in the test sample and the expression level of said at least 2 genes in the control sample are detected by measuring the level of proteins encoded by the genes.
19. The method according to paragraph 18, wherein the protein level is measured using an antibody assay or by mass spectroscopy.
20. The method according to paragraph 19, wherein the antibody assay is selected from Western analysis, immunofluorescence, ELISA, and
immunohistochemistry.
21. The method according to any one of paragraphs 1 -20, wherein the expression level of said at least 2 genes in the test sample and, optionally, the expression level of said at least 2 genes in the control sample are compared by in silico analysis.
22. The method according to paragraph 21 , wherein the in silico analysis comprises using a classifier generated by one or more network-based methods or machine-learning based methods.
23. An array comprising polynucleotides hybridizing to at least 2 COPD signature genes immobilized on a solid surface, wherein the COPD signature genes are selected from the genes listed in Table 1.
24. A panel comprising antibodies immobilized on a solid surface that bind to proteins encoded by at least 2 COPD signature genes, wherein the COPD signature genes are selected from the genes listed in Table 1 .
25. A computer readable medium comprising a gene signature, wherein the gene signature comprises at least 2 genes selected from the genes listed in Table 1 . 26. A device for diagnosing COPD, the device comprising: means for detecting the expression level of at least 2 genes listed in Table 1 in a test sample; means for correlating the expression level with a diagnosis of the COPD; and means for outputting the COPD diagnosis.
27. A device for prognosing COPD, the device comprising: means for detecting the expression level of at least 2 genes listed in Table 1 in a test sample; means for correlating the expression level with a prognosis of the COPD; and means for outputting the COPD prognosis. 28. A kit for classifying and grading COPD, comprising a set of reagents that detects expression levels of at least 2 genes listed in Table 1 in a test sample and instructions for using said kit for classifying and grading COPD in said individual.
29. A kit for assessing the prognosis of COPD in an individual, comprising a set of reagents that detects expression levels of at least 2 genes listed in Table 1 in a test sample from the individual and instructions for using said kit for determining the prognosis of the COPD in said individual.
30. A method of diagnosing, prognosing, classifying or grading COPD in a biological sample or an individual comprising measuring the expression levels of at least 2 genes listed in Table 1 in said biological sample or a test sample obtained from said individual; and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as COPD or non-COPD.
3 1. The method according to paragraph 30, wherein a classifier or a previously established standard is used to determine whether a test sample is a COPD or non- COPD.
32. The method according to paragraph 3 1 , wherein the classifier is obtained by training with a network-based method or a machine-learning based method using datasets obtained from subjects with COPD and datasets from subjects without COPD.
33. A method of diagnosing an individual as being at risk for or having Chronic Obstructive Pulmonary Disease (COPD) comprising detecting the expression level of at least 2 of the genes listed in Table 1 in a test sample obtained from the individual; wherein a differential pattern of expression levels of said at least 2 genes in the test sample diagnoses the individual as suffering from COPD.
34. The method according to paragraph 33, wherein the differential pattern of expression levels is identified by a classifier based on a plurality of genes listed in Table 1 , including said at least two genes, said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms.
35. The method according to paragraph 33 or 34, wherein the differential pattern of expression levels is identified by a classifier based on a plurality of genes listed in Table 1 , including said at least two genes, said classifier having been trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T- filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning,
Laplacian based learning and learning method based on network perturbation amplitude.
36. The method according to any one of paragraphs 33-35, wherein said classifier having been trained with at least the data in the Gene Expression
Omnibus datasets GSE10106, GSE10135, GSE 1 1906, GSE l 1952, GSE 13933, GSE 19407, GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545.
37. The method according to any one of paragraphs 33-36, wherein the method further comprises comparing the expression level of said at least 2 genes in the test sample and a control sample; or detecting the expression level of said at least 2 genes in the control sample and comparing the expression level of said at least 2 genes in the test sample and control sample, to identify the differential pattern.
38. The method according to any one of paragraphs 33-37, wherein said at least 2 genes are selected from the group consisting of: PROS 1 , IRAKI , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPRIA, PAPD4, VCL, PPP2R5C, DGKA, and CYP51 A1 .
39. The method according to any one of paragraphs 33-38, wherein the expression level of said at least 2 genes in the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array
comprising polynucleotides of a plurality of genes in Table 1 and said at least 2 genes.
40. The method according to any one of paragraphs 33-38, wherein the expression level of said at least 2 genes in the test sample are detected by measuring the level of proteins encoded by the genes.
41. An array comprising polynucleotides hybridizing to at least 2 COPD signature genes immobilized on a solid surface, wherein the COPD signature genes are selected from the genes listed in Table 1 and said array is not a human genome-wide array. 42. A device comprising antibodies immobilized on a solid surface that bind to proteins encoded by at least 2 COPD signature genes, wherein the COPD signature genes are selected from the genes listed in Table 1 .
43. A computer readable medium or computer program product comprising a classifier based on at least two genes listed in Table 1 , said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms.
44. The computer readable medium or computer program product according to paragraph 43 , wherein said classifier is trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T- filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude.
45. The computer readable medium or computer program product according to paragraph 44, wherein said classifier is trained with at least the data in the Gene
Expression Omnibus datasets GSE 10106, GSE 10135, GSE 1 1 06, GSE1 1952, GSE13933, GSE 1 9407, GSE 1 9667, GSE20257, GSE5058, GSE7832, and GSE8545.
46. The computer readable medium or computer program product according to any one of paragraphs 43-45, wherein said at least two genes are selected from the group consisting of PROS 1 , IRAKI , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR1A, PAPD4, VCL, PPP2R5C, DGKA, and CYP51A1.
Brief Description of the Drawings
[0029] Figure 1 provides a features selection and classification algorithm(s) used for prediction of a gene signature.
Detailed Description of the invention
[0030] In order that the invention described herein may be fully understood, the following detailed description is set forth.
[0031 ] Unless defined otherwise, all technical and scienti fic terms used herein have the same meaning as those commonly understood by one of skill in the art to which this invention belongs. In case of conflict, the present specification, including definitions, will control. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, suitable methods and materials are described below. The materials, methods and examples are illustrative only, and are not intended to be limiting. All publications, patents and other documents mentioned herein are incorporated by reference in their entirety.
[0032] Throughout this specification, the word "comprise" or variations such as "comprises" or "comprising" will be understood to imply the inclusion of a stated integer or groups of integers but not the exclusion of any other integer or group of integers.
[0033] The term "antibody" refers to an immunoglobulin molecule capable of specific binding to a target, such as a carbohydrate, polynucleotide, lipid, polypeptide, etc., through at least one antigen recognition site, located in the variable region of the immunoglobulin molecule. As used herein, unless otherwise indicated bv context, the term i intended to encomnass not onlv intact nolvclonal or monoclonal antibodies, but also engineered antibodies (e.g., chimeric, humanized and/or derivatized to alter effector functions, stability and other biological activities) and fragments thereof (such as Fab, Fab', F(ab')2, Fv), single chain (ScFv) and domain antibodies, including shark and camelid antibodies), and fusion proteins comprising an antibody portion, multivalent antibodies, multispecific antibodies (e.g., bispecific antibodies so long as they exhibit the desired biological activity) and antibody fragments as described herein, and any other modified configuration of the immunoglobulin molecule that comprises an antigen recognition site. An antibody includes an antibody of any class, such as IgG, IgA, or IgM (or sub class thereof), and the antibody need not be of any particular class. Depending on the antibody amino acid sequence of the constant domain of its heavy chains, immunoglobulins can be assigned to different classes. There are five major classes of immunoglobulins: IgA, IgD, IgE, IgG, and IgM, and several of these may be further divided into subclasses (isotypes), e.g., IgG l , IgG2, IgG3, IgG4, IgAl and IgA2 in humans. The heavy chain constant domains that correspond to the different classes of immunoglobulins are called alpha, delta, epsilon, gamma, and mu, respectively. The subunit structures and three dimensional configurations of different classes of immunoglobulins are well known.
[0034] The term "array" refers to the arrangement of biomarker detection molecules, such as nucleic acid probes or antibodies, on a solid support that allows for high-throughput screening of a sample to detect the presence and/or quantity of a biomarker. Such arrays may be used, e.g. , to evaluate the expression levels of several genes of interest in a single high-throughput reaction. The array may be a nucleic acid array, such as a nucleic acid microarray; a protein array, such as a protein microarray; a peptide array, such as a peptide microarray; a tissue microarrav. such as a tissue microarrav or an antibodv microarrav. such as an antibody microarray. The solid substrate may be a microscopic bead, a glass slide, a plastic chip or a silicon chip.
[0035] The term "biomarker" refers to a characteristic whose presence, absence or level indicates a biological state. Typically, the properties of biomarkers indicate a normal process, a pathogenic process or a response to a pharmaceutical or therapeutic intervention. A biomarker can be a cell, a gene, a gene product, an enzyme, a hormone, a protein, a peptide, an antibody, a nucleic acid molecule, a metabolite, a lipid, a free fatty acid, cholesterol or some other chemical compound. A biomarker can be a morphologic biomarker (for example, a histological change, DNA ploidy, malignancy-associated changes in the ceil nucleus and premalignant lesions) or a genetic biomarker (for example, DNA mutations, DNA adducts and apoptotic index).
[00361 The term "Chronic Obstructive Pulmonary Disease" or "COPD" refers to a complex disease that results in progressive loss of lung function. COPD is typically characterized by the occurrence of chronic bronchitis or emphysema, both of which result in airway narrowing. Clinically, COPD is typically detected by low airflow in lung function tests. COPD is typically irreversible and gets progressively worse over time. Symptoms of COPD include chronic cough, chronic sputum production, dyspnea, rhonchi, wheezing, chest tightness, tiredness and decreased airflow in lung function tests. Individuals suffering from very severe COPD can develop respiratory failure and present with cyanosis, headaches, drowsiness, and/or asterixis. COPD is a progressive disease and prognosis of the disease can be predicted by severe airflow obstruction, poor exercise capacity, shortness of breath, being significantly underweight or overweight, respiratory failure, cor pulmonale, and frequent acute exacerbations. COPD prognosis can be evaluated using the BODE index, which is a scoring system that measures FEV 1 , body-mass index, 6-minute walk distance, and a modified MRC (Medical Research Council) dyspnea scale to estimate outcomes in COPD.
[0037] The term "classifying COPD" refers to a method for determining the type of COPD from which a subject suffers. COPD can be classified as primarily bronchial or primarily emphysematous. Such classifications are made by simply analyzing clinical, functional, and radiological findings or by detecting biomarkers. In bronchial COPD, lung damage and inflammation occurs in the large airways resulting in chronic bronchitis, which is characterized by hyperplasia and hypertrophy of goblet cells and mucous glands in the airway. In emphysematous COPD, lung damage and inflammation occurs in the alveoli and is characterized by enlargement of the air spaces distal to the terminal bronchioles, with destruction of their walls and a reduction in lung elasticity. COPD may be classified based on the presence, absence, alteration or levels of biomarkers. COPD may be classified based on the COPD gene signature. Classifying COPD may also refer to distinguishing between bronchial or emphysematous COPD.
[0038] As used herein, the term "computer program" refers to a sequence of instructions, written to perform a specified task within a computer. For example, a computer program product is described, the product comprising computer-readable instructions that, when executed in a computerized system comprising at least one processor, cause the processor to carry out one or more steps of any of the methods described above. In another example, a computerized system is described, the system comprising a processor configured with non-transitory computer-readable instructions that, when executed, cause the processor to carry out any of the methods described herein. The computer program product and the computerized methods described herein may be implemented in a computerized system having one or more computing devices, each including one or more processors.
Generally, the computerized systems described herein may comprise one or more engines, which include a processor or devices, such as a computer, microprocessor, logic device or other device or processor that is configured with hardware, firmware, and software to carry out one or more of the computerized methods described herein. Any one or more of these engines may be physically separable from any one or more other engines, or may include multiple physically separable components, such as separate processors on common or different circuit boards. The computer systems of the present invention comprises means for implementing the methods and its various embodiments as described herein. The computerized system described herein may include a distributed computerized system having one or more processors and engines that communicate through a network interface. Such an implementation may be appropriate for distributed computing over multiple communication systems.
[ 0039 ] The term "computer readable medium" refers to a medium capable of storing data, such that the data may be accessed by a computer. Such a medium may take many forms, including but not limited to, non-volatile media and volatile media. Non- volatile media include, for example, optical, magnetic, or opto- magnetic disks, or integrated circuit memory, such as flash memory. Volatile media include, for example, dynamic random access memory (DRAM), which typically constitutes the main memory. Common forms of computer-readable media include, for example, a floppy disk, a flexible disk, hard disk, magnetic tape, magnetic cards, magnetic ink characters, magnetic drums, any other magnetic medium, a CD-ROM, DVD, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, a RAM, a PROM, an EPROM or EEPROM (electronically erasable programmable read-only memory), a
FLASH-EEPROM, barcodes, semiconductors, microchips and any other memory chip or cartridge.
100401 The term "control sample" refers to a sample against which a test sample is compared in order to diagnose, prognose, classify or grade the test sample. A control sample may be healthy tissue or may be a well-characterized sample from an individual suffering from COPD, including but not limited to, GOLD stage 1 , GOLD stage 2, GOLD stage 3, or GOLD stage 4 COPD. A control sample can be analyzed concurrently with or separately from the test sample, including before or after analyzing the test sample. The data from the analysis of a control sample may be stored, e.g., in a computer readable medium or in a manual, for comparison against test samples analyzed in the future, or as data for training network-based or machine-learning methods. A control sample may be developed as a medical standard for comparison. For example, analysis of control samples has developed medical standards for normal fed and fasted blood glucose levels; normal, at risk, and hypertensive blood pressures, and normal resting heart rates. As used herein, the term "control sample" includes samples that provided a medical standard. Accordingly, a test sample may be compared against a medical standard generated from control samples. For example, expression of a variant or mutated form of a gene may be indicative of a change medical condition. Alternatively, a change in expression level of a gene may be indicative of a change in medical condition. A control sample may be lung tissue, such as tissue obtained by biopsy from a healthy individual, or some other sample. For example, a control sample may be blood, blood cells, serum, plasma, sputum, saliva, tissue, bronchial washing, bronchial aspirates, bronchia brushings, exhaled breath, lymph fluid, and urine. Tissue specimens, such as those obtained by biopsy, may be fixed (e.g., formaldehyde-fixed paraffin-embedded (FFPE)). The control sample may be obtained from a tissue bank. The control sample may also be obtained from a cadaver or an organ donor.
[0041 ] The terms "differential pattern of expression" and "differential expression" are used interchangeably herein and refer to a difference in an activity measurement (e.g., the variability or difference of genetic expression) of a biological entity under different conditions. For example, one condition may refer to an experimental treatment (such as exposure to a potentially carcinogenic agent), and another condition may refer to a control treatment (such as a null treatment). In an example, a fold-change is a number describing how much a measurement at a node (or biological entity) changes from an initial value to a final value between control data and treatment data, or between two sets of data representing different treatment conditions. The fold-change number may represent the logarithm of the fold-change of the activity of the biological entity between the two conditions. A confidence interval for the significance of the fold- change number may also be assessed. [0042] The term "forced expiratory volume in one second" or "FEV l " refers to the volume of air that can forcibly be blown out in one second, after full inspiration. Average values for FEV l in healthy individuals depend on sex and age and have been well-characterized in the art. FEVl and the FEV l to FVC ration (FEVl /FVC) are used clinically to grade COPD. In healthy adults
FEVl /FVC should be approximately 75-80%. In obstructive diseases, such as COPD, FEV l is diminished because of increased airway resistance to expiratory flow. While the FVC may be decreased as well, due to the premature closure of airway in expiration, FEV l is typically more affected because of the increased airway resistance, so the FEV1/FVC ratio reflects the degree of airway closure compared to lung volume.
[0043] The term "forced vital capacity" or "F VC" refers to the volume of air that can forcibly be blown out after full inspiration, measured in liters.
[0044] The terms "gene signature" and "genetic signature" are used
interchangeably herein and refer to a group of genes expressed in a cell, whose combined expression pattern may be indicative of, e.g., a normal state, an at-risk state, a diseased state, a treated state or a recovery state. A gene signature may be characterized by which genes are expressed or at what level each gene is expressed. Gene signatures are particularly useful in diagnosing, prognosing, classifying or grading complex diseases states, which result from the combination of several genetic and environmental factors. The gene signatures disclosed herein may be used, e.g., for the diagnosis, prognosis, classification and/or grading of COPD in an individual. The gene signature may be unique to the class and grade of COPD.
[0045] The term "grading COPD" refers to a method for determining the grade of COPD from which a subject suffers. There are several different grades of COPD, which reflect the severity of the disease. For example, the Global Initiative for Chronic Obstructive Lung Disease (GOLD) characterizes COPD patients into GOLD Stage 1 to 4 depending on the severity of disease. GOLD stage 0 refers to a high risk population who did not present the symptoms used to describe stage 1 ; stage 1 refers to mild COPD and is characterized by a FEV l /FVC ratio less than 70% and an FEVl greater than 80%; stage 2 refers to moderate COPD and is characterized by a FEV l/FVC ratio less than 70% and an FEV l between 50% and 80%; stage 3 refers to sever COPD and is characterized by a FEV l /FVC ratio less than 70%) and an FEVl between 30% and 50%o and stage 4 refers to very severe COPD and is characterized by a FEVl/FVC ratio less than 70% and an FEV l less than 30% or the presence of chronic renal failure or right heart failure.
[0046 j The term "in silico analysis" refers to analysis performed on a computer or via computer simulation. Gene signature analysis involves detection of gene expression based on identity and expression level for a multitude of genes. In silico analysis may apply one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample, e.g., as COPD or non-COPD. Comparisons between expression levels from test samples and control samples may require computer analysis to determine the degree and significance of any changes observed. See, e.g. , U.S. Provisional Patent Application entitled "Systems and Methods relating to Network-based Biomarker Signatures," filed concurrently with the instant application, incorporated herein by reference in its entirety and having the attorney docket no. 106500-0022-001 ; U.S. Provisional Patent Application entitled "Systems and Methods for Generating Biomarker Signatures," filed concurrently with the instant application, incorporated herein by reference in its entirety, incorporated herein by reference in its entirety and having the attorney docket no. 106500-0028-001 ; U.S. Provisional Patent Application entitled "Systems and Methods for Generating Biomarker Signatures with
Integrated Bias Correction and Class Prediction," filed concurrently with the instant application, incorporated herein by reference in its entirety and having the attorney docket no. 106500-0032-001 ; and U.S. Provisional Patent Application entitled "Systems and Methods for Generating Biomarker Signatures with
Integrated Dual Ensemble and Simulated Annealing Techniques," filed
concurrently with the instant application, incorporated herein by reference in its entirety and having the attorney docket no. 106500-003 1 -001 .
[0047] The term "individual" refers to a vertebrate, preferably a mammal. The mammal can be, without limitation, a mouse, a rat, a cat, a dog, a horse, a pig, a cow, a non-human primate or a human. [0048] The term "individual at risk for COPD" refers to an individual who is predisposed to COPD. Predisposition to COPD may be due to one or more genetic or environmental factors. For example, an individual related to a COPD patient is more likely to get COPD than an individual who is not related to a COPD patient. Further, exposure to environmental factors such as radon gas, asbestos, tobacco smoke, and air pollution can increase the risk for COPD and predispose an individual to COPD.
[0049] The term "individual having COPD" or "individual suffering from COPD" refers to an individual experiencing progressive loss of lung function, typically characterized by airway narrowing. COPD can be bronchial or emphysematous and may be detected by analyzing clinical, functional, and radiological findings or detecting relevant biomarkers.
[0050] The term "MALDI-TOF" refers matrix-assisted laser
desorption/ionization time of flight mass spectroscopy. Matrix-assisted laser desorption/ionization (MALDI) is a two step process that uses laser-triggered desorption of protonated and deprotonated matrix materials to protonate or deprotonate analyte molecules (e.g., DNA, RNA, and proteins). Time-of- flight (TOF) mass spectrometry refers to a method in which an ion's mass-to-charge ratio is determined via the time that it takes an ionized particle to reach a detector at a known distance.
[0051 ] The term "machine learning methods" refers to methods that allow a machine, such as a programmable computer, to improve its performance at a certain predictive task that is based on the known properties of examples or training data. Machine learning methods include, without limitation, support vector machines (SVMs), network-based SVMs, ensemble classifiers, neural network-based classifiers, logistic regression classifiers, decision tree-based classifiers, classifiers employing a linear discriminant analysis technique, a random-forest analysis technique, or both.
[0052] The term "network-based methods" refers to methods for identifying biomarkers that is based on the properties of groups of functionally interrelated genes that form a network in a biological system, instead of treating individual genes in the biological system a priori as completely independent and identical. [0053] The term "numerical biomarker score" refers to a number that is representative of the result(s) of one or more of the network-based analysis or machine learning methods.
[0054] The term "polynucleotide hybridizing to" refers to a polynucleotide molecule that binds to a target nucleic acid molecule through complementary base pair sequencing. Hybridization typically requires two nucleic acids that contain complementary sequences, although depending on the stringency of the
hybridization, mismatches between bases are possible. The appropriate stringency for hybridizing nucleic acids depends on the length of the nucleic acids and the degree of complementation, variables well known in the art. Exemplary high stringent hybridization conditions are equivalent to about 20-27 °C below the melting temperature (Tm) of the DNA duplex formed in about 1 M salt. Many equivalent procedures exist and several popular molecular cloning manuals describe suitable conditions for highly stringent hybridization and, furthermore. provide formulas for calculating the length of hybrids expected to be stable under these conditions (see, e.g., Current Protocols in Molecular Biology, John Wiley & Sons, N.Y. (1989), 6.3.1 6 or 13.3.6; or pages 9.47-9.57 of Sambrook, et al. (1989) Molecular Cloning, 2nd ed., Cold Spring Harbor Press). "High stringency" refers to hybridization and/or washing conditions at 68 °C in 0.2 x SSC, at 42 °C in 50 % formamide, 4 x SSC, or under conditions that afford levels of hybridization equivalent to those observed under either of these two conditions. The greater the degree of similarity or homology between two nucleotide sequences, the greater ιι ιυ v al ue u i m, lui !i » ui !UJ ui i iue i L iu at m^ iia v n ig u iG SC i i n_ i i a u v C stability (corresponding to higher Tm) of nucleic acid hybridizations decreases in the following order: RNA:RNA, DNA:RNA, DNA: DNA.
[0055] The terms "protein," "polypeptide" and "peptide" are used
interchangeably and indicate at least one molecular chain of amino acids linked through covaient or non-covalent bonds. The terms do not refer to a specific length of the molecular chain. Peptides, oligopeptides and proteins are included within the definition of "polypeptide". The terms include post-translationai modifications of the molecule, e.g., phosphorylation, glycosylation and acetylation. The terms also include protein fragments, fusion proteins, mutant proteins and variant proteins.
[0056] The term "SELDI-TOF" refers surface-enhanced laser
desorption/ionization time of flight mass spectroscopy. Surface-enhanced laser desorption/ionization (SELDI) is a variant of MALDI that uses a target with a biochemical affinity for the analyte. Time-of-flight (TOF) mass spectrometry refers to a method in which an ion's mass-to-charge ratio is determined by measuring the time that it takes an ionized particle to reach a detector at a known distance.
[0057] The term "test sample" refers to a sample obtained from an individual at risk for, having or suffering from COPD. A test sample may be any sample suspected of containing or exhibiting a biomarker. The test sample is analyzed and compared to a control sample, including medical standards developed from control samples, to diagnose, prognose, classify or grade COPD in the individual. A test sample may be obtained from lung tissue, such as tissue obtained by biopsy from a tumor, or other biological tissue. For example, a test sample may be blood, blood cells, serum, plasma, sputum, saliva, tissue, bronchial washing, bronchial aspirates, bronchia brushings, exhaled breath, lymph fluids, and urine. Tissue specimens, such as those obtained by biopsy, may be fixed (e.g., formaldehyde-fixed paraffin- embedded (FFPE)).
[0058] As used herein, to "train" a data set means to generate a classifier that can accurately predict classifications of a set of test samples. For example, a training data set includes a set of samples, and each sample may correspond to a measurement from a different patient. A machine learning technique is applied to the training data set to generate a "classifier," which corresponds to a way of assigning each sample in the training data set to a category (such as "disease positive" or "disease free"). In addition to the training data set, a training class set is known. The training class set includes a known category assigned to each sample (or person). The categories predicted by the classifier are compared to the known categories. If the predicted categories mostly match the known categories, the classifier has performed well. However, if there are substantial differences between the predicted categories and the known categories, the parameters of the machine learning technique may be updated, and the updated machine learning technique is applied. These steps are repeated until the performance of a classifier exceeds a threshold, and the final classifier is provided. The final classifier may then be applied to a test data set. The test data set may correspond to measured samples from different patients, but the patients in the test data set may have unknown categories (disease states). Thus, applying the final classifier to the test data set thus allows for prediction of the disease states of the patients.
Gene Signatures rnnrm r - „f : f„„
|υυ.ν/| wuc aspect ui uic liivciiuuii piOviQcs yCnc sigiiatLucs usciui tui diagnosing, prognosing, classifying or grading COPD. In some embodiments, the gene signature comprises at least 2 genes selected from the genes listed in Table 1. In some embodiments, the gene signature comprises at least 2, at least 3, at least 4, at 1 of ^ it 1 at Ipoct 1 ^ at lpacf at Ip cf at l^acf ^n af l a † ^^ at
/in i„„„+ I c i„„„+ cn i„„„+ cc „+ i„„„< c „+ t„„„+ : „+ in ic isi tu, at icdsi at least , at least , at least uu, at least uj, at tcast / υ, at least 75, at least 80 or at least 84 genes selected from the genes listed in Table 1. In some embodiments, the gene signature comprises each of the genes listed in Table 1. Optionally, said at least 2, at least 3, at least 4, at least 5, or at least 10 genes are selected from the group consisting of: PROS1, IRAKI, VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR1A, PAPD4, VCL, PPP2R5C, DGKA, and CYP51A1, which are the genes that appear in 4 of the 5 lists generated in Example 1.
ΓΠΠίίΠ! In e m r»rnHririirnf* tc at lr»¾ct at 1 j at l^act A at 1 > < at 1 p* Q
10, at least 15, at least 20, at least 25, at least 30 or at least 35 of the genes selected from the genes listed in Table 2 have increased expression compared to a control sample. In some embodiments, each of the genes listed in Table 2 has increased expression compared to a control sample. In some embodiments, at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, or at least 45 of the genes selected from the genes listed in Table 3 have decreased expression compared to a control sample. In some embodiments, each of the genes listed in Table 3 has decreased expression compared to a control sample. [0061] In some embodiments, the gene signature includes a degree of up- regulation of a subset of genes in the gene signature compared to the control sample. For example, each up-regulated gene in the gene signature may, independently, be up-regulated at least 1.5-fold, at least 2-fold, at least 2.5-fold, at least 3-fold, at least 3.5-fold, at least 4-fold, at least 4.5-fold, at least 5-fold, at least
6- fold, at least 7-fold, at least 8-fold, at least 9-fold, at least 10-fold, at least 100- fold, at least 1 ,000-fold or more compared to the control sample. Similarly, in some embodiments, the gene signature includes a degree of down-regulation of a subset of genes in the gene signature compared to the control sample. For example, each down-regulated gene in the gene signature may, independently, be down-regulated at least 1 .5-fold, at least 2-fold, at least 2.5-fold, at least 3 -fold, at least 3.5-fold, at least 4-fold, at least 4.5-fold, at least 5-fold, at least 6-fold, at least
7- fold, at least 8-fold, at least 9-fold, at least 10-fold, at least 100-fold, at least 1 , 000-fold or more compared to the control sample.
Γ0062] The present invention encompasses the following gene signatures;
i. A and B;
ii. A, B, and C;
iii. A, B, C, and D;
iv. A, B, C, D, and E;
v. A, B, C, D, E, and F;
vi. A, B, C, D, E, F, and G;
vii. A, B, C, D, E, F, G, and II;
viii. A. B. C. D. E. F. G. H. and I:
ix. A, B, C, D, E, F, G, H, I, and J;
X. A, B, C, D, E, F, G, H, I, J, and K
xi. A, B, C, D, E, F, G, H, I, J, K, and L
xii. A, B, C, D, E, F, G, PI, I, J, , L and M
xiii. A, B, C, D, E, F, G, FI, I, J, K, L, M, and N
xiv. A, B, C, D, E, F, G, H, I, J, K, L, M, N, and O;
XV. A, B, C, D, E, F, G, I I, I, J, K, L, M, N, O, and P;
xvi. A, B, C, D, E, F, G, H, I, J, K, L, M, N, 0, P, and Q
xvii. A, B, C, D, E, F, G, II, I, J, K, L, M, N, 0, P, Q, and R; xviii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, and S;
xix. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, and T;
xx. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, O, P, Q, R, S, T, and U;
xxi. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, and V;
xxii. A, B, C. D, E, F, G, Id, I, J, K, L, M, N, 0, P, Q, R, S, T, U, V, and W; xxiii. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, O, P, Q, R, S, T, U, V, W, and X; xxiv. A, B, C, D, E, F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, V, W, X, and
Y;
xxv. A, B, C, D, E, F, G, Id, I, J, K. L, M, N, 0, P, Q, R, S, T, U, V, W, X, Y, and Z;
xxvi. A, B, C, D, E, F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, and AA;
xxvii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, and BB;
xxviii. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, 0, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB and AC;
xxix. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, and AD;
xxx. A, B, C, D, E, F, G, I t, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD and AE;
xxxi. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE and AF;
xxxii. A, B, C, D5 E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF and AG;
xxxiii. A, B, C, D, E, F, G, Id, I, J, , L, M, N, 0, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG and AH;
xxxiv. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH and AI;
xxxv. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI ana A J;
xxxvi. A, B, C, D, E, F, G, II, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ and AK; χχχνπ. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK and AL;
xxxvm. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL and AM; xxxix. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P. Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM and AN; xl. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN and AO; xli. A, B, C, D, E, F, G, II, I, J, K, L, M, N, O, P, , i , O, i, U , VV, Λ, I , /_.,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO and AP;
xlii. A, B, C, D, E, F, G, II, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP and AQ;
xhii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, p, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AL AJ, AK, AL, AM, AN, AO, AP, AQ and AR;
X I IV. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q. R, S, T, U, W, X, Y, Z.
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR and AS;
xlv. A, B, C, D, E, F, G, FI, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS and AT;
xlvi. A, B, C, D, E, F, G, II, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z.
AA, AB, AC, AD, AE. AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR. AS, AT and AU;
xlvii. A, B, C, D, E. F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU and AV;
xlviii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P. Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP. AQ, AR, AS, AT, AU, AV and AW; xlix. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AFI, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW and AX;
1. A, B, C, D, E, F, G, H, I, J, K, L, M, N, 0, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
AQ, AR, AS, AT, AU, AV, AW, AX and AY;
li. A, B, C, D, E, F, G, id, I, J, K, L, M, N, 0, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AFI, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY and AZ;
Hi. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, 0, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AFI, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ and BA;
liii. A, B, C, D, E, F, G, id, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA and BB;
liv. A, B, C, D, E, F, G, FI, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB and BC;
iv. A, B, C, D, E, F, G, i f. I, J, K, L, M, N, 0, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC and BD; lvi. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AFI, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD and BE; lvii. A, B, C, D, E, F, G, Id, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE and BF;
lviii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF and BG; lix. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG and BH;
lx. A, B, C, D, E, F, G, H, I, J. K, L, M, N, 0, P, Q, R, S, T, U, W, X, Y, Z.
AA, AB, AC, AD, AE, AF, AG, AFI, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH and BI;
Ixi. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
AQ, AR, AS, AT, AU, AV. AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI and BJ;
ixii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AO, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ and BK;
Ixiii. A, B, C, D, E, F, G, H, I, J. K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, ovj, on, m, dJ , I *> iv ariu o ,
lxiv. A, B, C, D, E, F, G, II, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE. AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP. AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BI I. BI, BJ, BK, BL and BM;
lxv. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC. BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM and BN;
ixvi. A, B, C, D, E, F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN and BO; ixvii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O. P, Q, R, S, T, U, W, X, Y, Z.
AA, AB, AC, AD, AE, AF, AG, AH, Ai, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BE, BM, BN, BO and BP;
lxviii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR. AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE. BF, BG, BH, BE BJ, BK, BE, BM, BN, BO, BP and BQ;
lxix. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BE, BM, BN, BO, BP, BQ and BR;
Ixx. A. B, C, D, E, F, G, H, I, J, K, L, M, N. O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR and BS;
lxxi. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD. AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH. B !, BJ, BK , BL, BM, BN, BO, BP, BQ, BR, BS and BT;
Ixxii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT and BU ; Ixxiii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE. BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO. BP, BO, BR, BS, BT, BU and BV; Ixxiv. A, B, C, D, E. F, G, H, I, J, K, L, M, N, O. P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP,
AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV and BW;
Ixxv. A, B, C, D, E, F, G, IT, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW and BX;
Ixxvi. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AO, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX and BY;
Ixxvii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AO, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV,
BW, BX, BY and BZ;
lxxviii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV,
BW, BX, BY, BZ and CA;
Fxxix. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV,
BW, BX, BY, BZ, CA and CB;
Sxxx. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF,
BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV,
BW, BX, BY, BZ, CA, CB and CC; lxxxi. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AFI, AI, AJ, AK, AL, AM, AN, AO, ΛΡ, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC and CD;
Ixxxii. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV. AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BU, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD and CE;
lxxxiii. A, B, C, D, E, F, G, FI, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG. BH, BI, BJ, BK, BL, BM, BN, BO, BP, BO, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE and CF; and
lxxxiv. A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z,
AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS. AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE. BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE, CF and CG;
wherein each of A, B, C, D, E, F, G, H, I, J, K, L. M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP. AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY. BZ, CA, CB, CC, CD, CE, CF and CG are independently selected from the genes listed in Table 1 and each of A, B, C, D, E, F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN. AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE. BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE, CF and CG are different.
[0063] In some embodiments, each of A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, ΒΧ, BY, BZ, CA, CB, CC, CD, CE, CF and CG are independently selected from the genes listed in Table 2 and each of A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE, CF and CG are different.
[0064] In some embodiments, each of A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE, CF and CG are independently selected from the genes listed in Table 3 and each of A, B, C, D, E, F, G, I I, I, J, K, L, M, N, O, P, Q, R, S, T, U, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, BH, BI, BJ, BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, BW, BX, BY, BZ, CA, CB, CC, CD, CE, CF and CG are different.
Methods of Using Biomarkers and Gene Signatures
[0065] The biomarkers and gene signatures of the invention may be used in methods of diagnosing, prognosing, classifying or grading COPD in biological sample or an individual. One aspect of the invention provides a method of diagnosing, classi fying or grading COPD in an individual at risk for or having
COPD. In some embodiments the method comprises classifying a test sample as COPD or non-COPD. In some embodiments, the method comprises measuring the expression levels of at least 2 genes listed in Table 1 in a test sample; and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as COPD or non-COPD. In some embodiments, the expression levels of at least 2, at least 3, at least 4, at least 5, at - J J - least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1 are measured. In some embodiments, a differential pattern of expression levels of said at least 2 genes in the test sample diagnoses, classifies or grades the COPD.
[0066] In one embodiment, the methods of the invention can be used to identify a gene signature and a classifier (e.g., a gene-signature-based classifier) that can distinguish datasets obtained a COPD sample from those datasets obtained from a non-COPD or healthy sample. In some methods of the invention, control data is not collected or used. Instead, a classifier or a previously established standard may be used to determine whether a test sample is a COPD sample. For example, a classifier that is obtained by training with network-based or machine-learning based methods using datasets obtained from subjects with COPD and datasets from subjects without COPD can be used. Alternatively, one or more numerical scores (e.g., average fold change or rank abs tvai as shown in Table 1 ) generated by the algorithms described herein may be used as a previously established standard. The levels of expression of one or more of the genes listed in Table 1 in a test sample may be compared to the previously established standard, and the comparison may be used to classify the test sample as a COPD sample or a normal sample.
[0067] in one embodiment, the invention provides a method of diagnosing COPD in a biological sample, wherein the method comprises determining the properties (for example, absence, presence or expression level) of one or more genes listed in Table 1 in the biological sample; and applying in silico analysis with a classifier obtained from a network-based method, a machine-learning based method, or a combination of the foregoing methods. The classifier can be obtained from the network-based methods, the machine-learning based methods, or a combination of the foregoing methods by training with datasets obtained from subjects with COPD, subjects with COPD of a certain determined GOLD stage, or healthy subjects. In another embodiment, a classifier may be obtained, given appropriate class(es) of training datasets, to identify a specific prognosis of the
COPD, to indicate transition from a more severe or less severe GOLD stage or to indicate that a particular treatment regimen should be used to treat the individual who provided the biological sample.
[0068] In one embodiment, the methods of the invention comprise obtaining a test sample (such as bronchial brushings) from an individual, determining the absence, presence or expression level of one or more of the genes listed in Table 1 in the test sample, comparing said absence, presence or expression level to the absence, presence or expression level of the same gene(s) in a control sample, and selecting a COPD treatment regimen based on the comparison. In a further embodiment, the invention provides a method for monitoring the progress of a COPD treatment in an individual, said method comprising determining at suitable time intervals before, during, or after therapy (for example, at different time points during the treatment) in a sample taken from said individual differential expression of a panel of at least 2 genes selected from the genes listed in 'fable 1 .
[0069] In one embodiment, the invention encompasses a method that comprises collecting data on the properties of one or more genes in the gene signature without generating a gene signature. For example, the method of the invention comprises obtaining a test sample from an individual, and detecting the absence, presence or the expression level of one or more of the genes listed in Table 1 in the sample. In one embodiment, the invention encompasses a method that comprises using data on the properties of one or more genes in a gene signature that are already collected as training data to generate an improved gene signature using one or more network-based methods, one or more machine learning methods, or a combination of the foregoing methods. In one embodiment, the invention encompasses a method that comprises collecting data on the properties of one or more genes in a biological system which is included in a gene signature, and using the data to predict a classification of the state of the biological system associated with the collected data.
[0070] In some embodiments, the method comprises detecting the expression level of at least 2 genes listed in Table 1 in a test sample obtained from the individual; and comparing the expression level of the genes listed in Table 1 in the test sample to the expression level of the genes listed in Table 1 in a control sample. In some embodiments, if the expression level of the genes listed in Table 1 is different in the test sample than in the control sample, then the individual suffers from COPD. In some embodiments, the expression level of the genes listed in Table 2 is higher in the test sample than in the control sample. Optionally, the expression level of the genes listed in Table 3 is lower in the test sample than in the control sample. In some embodiments, the method further comprises detecting the expression level of the genes listed in Table 1 in the control sample. In some embodiments, the expression levels of at least 2, at least 3 , at least 4, at least 5, at least 10, at least 1 5, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1 are detected.
[0071] In some embodiments, the test sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine. Optionally, the test sample is obtained from a large airway of the individual, such as from a bronchial brush inserted into the large airway of the individual.
[0072] In some embodiments, the control sample is selected from blood, serum, plasma, sputum, saliva, tissue, bronchia brushings, exhaled breath, and urine. Optionally, the control sample is obtained from a large airway of an individual not affected with COPD, such as from a bronchial brush inserted into the large airway of the individual not affected with COPD. In some embodiments, the control sample is obtained from the individual at risk for or having COPD prior to onset of COPD. In other embodiments, the control sample is obtained from an individual that does not suffer from COPD.
[0073] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring mRNA levels. For example, mRNA level is measured by amplification, hybridization, mass spectroscopy, serial analysis of gene expression, or massive parallel signature sequencing. Optionally, the amplification is reverse transcription PGR, real time quantitative PGR, differential display or TaqMan PGR. In some embodiments, the hybridization is a dot blot, a slot blot, an RNase protection assay, microarray hybridization, or in situ hybridization. The mass spectroscopy may be MALDI-TOF mass spectroscopy. In some embodiments, the expression level of the genes listed in Table 1 in the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1 .
[0074] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring the level of proteins encoded by the genes. Optionally, the protein level is measured using an antibody assay or by mass spectroscopy. In some embodiments, the antibody assay is selected from Western analysis, immunofluorescence, ELISA, and immunohistochemistry. The mass spectroscopy may be MALDI-TOF mass spectroscopy or SELDI-TOF mass spectroscopy.
[0075] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control sample are detected by measuring both mRNA levels and the level of proteins encoded by the genes. In some embodiments, expression levels are measured using the ampliiication, hybridization, mass spectroscopy, serial analysis of gene expression, massive parallel signature sequencing, and antibody assays discussed above.
[0076] In some embodiments, the expression level of the genes listed in Table 1 in the test sample and the expression level of the genes listed in Table 1 in the control non-tumor biological sample are compared by in silico analysis. The in silico analysis may be network based analysis or a machine-learning method. Methods of Biomarker Detection, Arrays and Panels
[0077] Detection of the nucleic acid and/or protein biomarkers described herein in a test sample or a control sample may be performed in a variety of ways.
[0078] In one aspect, the methods of the invention rely on the detection of the presence or absence of biomarker genes and/or biomarker gene expression, or the qualitative or quantitative assessment of either over- or under-expression of a biomarker gene in a population of cells in a test sample relative to a standard (for example, a control sample). Such methods utilize reagents such as biomarker polynucleotides and biomarker antibodies. [0079] In particular, the presence, absence or level of expression of a biomarker gene may be determined by measuring the amount of biomarker messenger RNA (mRNA), for example, by DNA-DNA hybridization, RNA-DNA hybridization, reverse transcription-polymerase chain reaction (PGR), real time quantitative PCR, differential display or TaqMan PCR; followed by comparing the results to a reference based on a control sample (for example, samples from clinically- characterized patients and/or cell lines of a known genotype/phenotype). In one embodiment, microRNA expression or turnover may be measured. Hybridization, mass spectroscopy (e.g., MALDI-TOF or SELDI-TOF mass spectroscopy), serial analysis of gene expression or massive parallel signature sequencing assays can also be performed. Non-limiting examples of hybridization assays include a singleplex or a multiplexed aptamer assay, a dot blot, a slot blot, an RNase protection assay, microarray hybridization, Southern or Northern hybridization analysis and in situ hybridization (e.g. , Fluorescent in situ hybridization).
[0080] For example, these techniques find application in mi croarray-based assays that can be used to detect and quantify the amount of biomarker gene transcript using cDNA- or oligonucleotide-based arrays. Microarray technology allows multiple biomarker gene transcripts and/or samples from different subjects to be analyzed in one reaction. Typically, mRNA isolated from a sample is converted into labeled nucleic acids by reverse transcription and optionally in vitro transcription (cDNAs or cRNAs labelled with, for example, Cy3 or Cy5 dyes) and hybridized in parallel to probes present on an array. See, for example, Schulze et al, Nature Cell Biol. , 3 :E190 (2001 ); and Klein et al, J Exp Med, 194: 1 625- 1638 (2001 ), which are incorporated herein by reference in their entirety. Standard Northern analyses can be performed if a sufficient quantity of the test cells can be obtained. Utilizing such techniques, quantitative as well as size-related differences between biomarker transcripts can also be detected. In some embodiments, the expression level of the genes listed in Table 1 in the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1.
[0081] In some embodiments biomarkers are detected using reagents that specifically detect the biomarker. Such reagents may bind to a target gene or a target gene product (e.g., mRNA or protein), such that levels of the gene product may be quantified. Such reagents may be nucleic acid molecules that hybridize to the mRNA or cDNA of target gene products. Alternatively, the reagents may be molecules that label mRNA or cDNA for later detection, e.g., by binding to an array. The reagents may bind to proteins encoded by the genes of interest. For example, the reagent may be an antibody or a binding protein that specifically binds to a protein encoded by a target gene of interest. Alternatively, the reagent may label proteins for later detection, e.g., by binding to an antibody on a panel. In some embodiments, reagents are used in histology to detect histological and/or genetic changes in a sample.
[0082] The present invention provides isolated biomarker polynucleotides or variants thereof, which can be used, for example, as hybridization probes or primers ("biomarker probes" or "biomarker primers") to detect or amplify nucleic acids encoding a biomarker polypeptide, particularly a biomarker polypeptide encoded by a biomarker gene or polynucleotide selected from the group depicted in Table 1 , Table 2, or Table 3.
[0083] Nucleic acid molecules comprising nucleic acid sequences encoding the biomarker polypeptides or proteins of the invention, or genomic nucleic acid sequences from the biomarker genes (e.g., intron sequences, 5 ' and 3 ' untranslated sequences), or complements thereof (i.e. , antisense polynucleotides), are collectively referred to as "biomarker genes," "biomarker polynucleotides" or "biomarker nucleic acid sequences" of the invention. The present invention also provides isolated biomarker polynucleotides or variants thereof which can be used, for example, as hybridization probes or primers ("biomarker probes" or "biomarker primers") to detect or amplify nucleic acids encoding a biomarker polypeptide of the invention. The term "biomarker gene product" thus
encompasses both mRNA as well as translated polypeptide as a gene product of a biomarker.
[0084] The isolated biomarker polynucleotide according to the invention may comprise flanking sequences (i. e. , sequences located at the 5' or 3' ends of the nucleic acid), which naturally flank the nucleic acid sequence in the genomic DNA of the organism from which the nucleic acid is derived. For example, in various embodiments, the isolated biomarker polynucleotide can comprise less than about 5 kb, 4 kb, 3 kb, 2 kb, 1 kb, 0.5 kb or 0.1 kb of nucleotide sequences which naturally flank the coding sequence in genomic DNA of the cell from which the nucleic acid is derived. In other embodiments, the isolated biomarker
polynucleotide is about 10-20, 21 -50, 51 -100, 101 -200, 201 -400, 401 -750, 75 1 - 1000, or 1001 - 1500 bases in length.
[0085] In various embodiments, the biomarker polynucleotides of the invention are used as molecular probes in hybridization reactions or as molecular primers in nucleic acid extension reactions as described herein. In these instances, the biomarker polynucleotides may be referred to as biomarker probes and biomarker primers, respectively, and the biomarker polynucleotides present in a sample which are to be detected and/or quantified are referred to as target biomarker
polynucleotides. Two biomarker primers arc commonly used in DNA
amplification reactions and they are referred to as biomarker forward primer and biomarker reverse primer depending on their 5 ' to 3 ' orientation relative to the direction of transcription.
[0086] In one embodiment, the invention encompasses methods of detecting genetic change in a biomarker gene (e.g., a mutation or a change in copy number). In another embodiment, the invention encompasses methods of detecting a change in the rnethvlatiosi of a biomarker gene.
[0087] A biomarker probe or a biomarker primer is typically an oligonucleotide which binds through complementary base pairing to a subsequence of a target biomarker polynucleotide. The biomarker probe may be, for example, a DNA fragment prepared by amplification methods such as by PGR or it may be chemically synthesized. A double-stranded fragment may then be obtained, if desired, by annealing the chemically synthesized single strands together under appropriate conditions or by synthesizing the complementary strand using DNA polymerase with an appropriate primer. Where a specific nucleic acid sequence is given, it is understood that the complementary strand is also identified and included as the complementary strand will work equally well in situations where the target is a double stranded nucleic acid. A nucleic acid probe is
complementary to a target nucleic acid when it will anneal only to a single desired position on that target nucleic acid under proper annealing conditions which depend, for example, upon a probe's length, base composition, and the number of mismatches and their position on the probe, and must often be determined empirically. Such conditions can be determined by those of skill in the art.
[0088] In one aspect of the invention, biomarkers may be detected in the test sample or the control sample by gene expression profiling. In these methods, mRNA is prepared from a sample and mRNA expression levels are measured by reverse transcription quantitative polymerase chain reaction (RT-PCR followed with qPCR). RT-PCR is used to create a cDNA from the corresponding mRNA. The cDNA may be used in a qPCR assay to produce fluorescence as the D A amplification process progresses. By comparison to a standard curve, qPCR can produce an absolute measurement such as number of copies of mRNA per cel l. Northern blots, microarrays, Invader assays, and RT-PCR combined with capillary electrophoresis may be used to measure expression levels of mRNA in a sample. Further details are provided, for example, in ''Gene Expression Profiling: Methods and Protocols," Richard A. Shimkets, editor, Humana Press, 2004 and US patent application 2010/0070191 .
[0089] The invention encompasses an array comprising polynucleotides that hybridize to genes listed in Table 1 , Table 2, or Table 3. The array may comprise polynucleotides that hybridize to at least 2, at least 3, at 'least 4, at least 5, at 'least 10, at least 1 5, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, at least 84 genes or all genes listed in Table 1 , Table 2, or Table 3. In one
embodiment, the polynucleotides are immobilized on a solid surface. Examples of solid surfaces include paper, filter, nylon or other type of membrane, slide including glass slide, and chip (e.g., silicon, microarray chip). The polynucleotides may be single-stranded nucleic acid molecules (e.g., antisense oligonucleotides or iragments of cDNA). In some embodiments, the array is not a human genome- wide array. Examples of human genome-wide array include, but are not limited to, Exon 1.0 ST, Gene 1.0 ST, U 95, U 133, U 133A 2.0, and U 133 Plus 2.
[0090] In another aspect of the invention, detection of the biomarkers described herein may be accomplished by an immunoassay procedure. The immunoassay typically includes contacting a test sample with an antibody that specifically binds to or otherwise recognizes a biomarker, and detecting the presence of the antibody/biomarker complex in the sample. The immunoassay procedure may be selected from a wide variety of immunoassay procedures known to those skilled in the art such as, for example, competitive or non-competitive enzyme-based immunoassays, immunoprecipitation, enzyme-linked immunosorbent assays (ELiSA), radioimmunoassay (RIA), immunofluorescence, immunohistochemistry (IHC), cytological assays and Western blots. Further, multiplex assays may be used, including antibody panels or arrays, wherein several desired antibodies are placed on a support, such as a glass bead or plate, and reacted or otherwise contacted with the test sample or the control sample.
[0091] Antibodies used in these assays may be monoclonal or polyclonal, and may be of any type such as IgG, IgM, IgA, IgD and IgE. Monoclonal antibodies may be used to bind to a specific epitope offered by the biomarker molecule, and therefore mav provide a more specific and accurate result. Antibodies may be produced by immunizing animals such as rats, mice, rabbits and goats. The antigen used for immunization may be isolated from the samples or synthesized by recombinant protein technology. Methods of producing antibodies and of performing antibody-based assays are well-known to the skilled artisan and are described, for example, more thoroughly in Antibodies: A Laboratory Manual (1988) by Harlow & Lane; Immunoassays: A Practical Approach, Oxford
University Press, Gosling, J. P. (ed.) (2001 ) and/or Current Protocols in Molecular Biology (Ausubel et al.) which is regularly and periodically updated.
[0092] In certain embodiments, the present invention also provides "biomarker antibodies" including polyclonal, monoclonal, or recombinant antibodies, and fragments and variants thereof, that immunospecifically bind the respective biomarker proteins or polypeptides encoded by the genes or cDNAs (including polypeptides encoded by mRNA splice variants) as listed in Tables 1 , 2, and 3.
[0093] Various chemical or biochemical derivatives of the antibodies or antibody fragments of the present invention can be produced using known methods. One type of derivative which is diagnostically useful as an immunoconjugate comprising an antibody molecule, or an antigen-binding fragment thereof, to which is conjugated a detectable label. However, in many embodiments, the biomarker antibody is not labeled but in the course of an assay, it becomes indirectly labeled by binding to or being bound by another molecule that is labeled. The invention encompasses molecular complexes comprising a biomarker antibody and a label , as well as immunocomplexes comprising a biomarker polypeptide, a biomarker antibody, and immunocomplexes comprising a biomarker polypeptide, a biomarker antibody, and a label.
[0094] Examples of detectable substances or detectable labels include various enzymes, prosthetic groups, fluorescent materials, luminescent materials, bioluminescent materials, and radioactive materials. Examples of suitable enzymes include horseradish peroxidase, alkaline phosphatase, beta-galactosidase and acetylcholinesterase. Examples of suitable prosthetic group complexes include streptavidin/biotin and avidin/biotin. Examples of suitable fluorescent materials include umbelliferones, fluoresceins, fluorescein isothiocyanate, rhodamines, dichlorotriazinylarnine fluorescein, dansyl chloride, phycoerythrins, Alexa Fluor 647, Alexa Fluor 680, DilCi9(3), Rhodamine Red-X, Alexa Fluor 660, Alexa Fluor 546, Texas Red, YOYO- 1 + DNA, tetramethylrhodamine, Alexa Fluor 594, BODIPY FL, Alexa Fluor 488, Fluorescein, BODIPY TR, BODiPY TMR, carboxy SNARF- 1 , FM 1 -43 , Fura-2, Indo- 1 , Cascade Blue, NBD, DAPI, Alexa Fluor 350, aminomethylcoumarin, Lucifer yellow, Propidium iodide, or dansylamide. An example of a luminescent material is luminol. Examples of bioluminescent materials include green fluorescent proteins, modified green fluorescent proteins, luciferase, luciferin, and aequorin. Examples of suitable
. . . . . . , 125 13 1 35 3
radioactive material include I, I, S or 1 1.
[0095] Immunoassays for biomarker polypeptides will typically comprise incubating a sample, such as a biological fluid, a tissue extract, freshly harvested cells, or lysates of cells, in the presence of a detectably labeled antibody capable of identifying biomarker gene products or conserved variants or peptide fragments thereof, and detecting the bound antibody by any of a number of techniques well- known in the art. One way of measuring the level of biomarker polypeptide with a specific biomarker antibody of the present invention is by enzyme immunoassay (EIA) such as an enzyme-linked immunosorbent assay (ELISA) (Voller, A. et ctl , J. Clin. Pathol. 57 :507-520 (1978); Butler, I.E., Melh. Enzymol. 75:482-523 (1981 ); Maggio, E. (ed.), Enzyme Immunoassay, CRC Press, Boca Raton, FL, 1980). The enzyme, either conjugated to the antibody or to a binding partner for the antibody, when later exposed to an appropriate substrate, will react with the substrate in such a manner as to produce a chemical moiety which can be detected, for example, by spectrophotometric, or fluorimetric means.
[0096] The biological sample may be brought in contact with and immobilized onto a solid phase support or carrier such as nitrocellulose, or other solid support which is capable of immobilizing cells, cell particles or soluble proteins. The support may then be washed with suitable buffers followed by treatment with the detectably labeled biomarker antibody. The solid phase support may then be washed with the buffer a second time to remove unbound antibody. The amount of bound label on solid support may then be detected by conventional means. A well known example of such a technique is Western blotting.
[0097] In various embodiments, the present invention provides compositions comprising labelled biomarker polynucleotides, or labelled biomarker antibodies to the biomarker proteins or polypeptides, or labeled biomarker polynucleotides and labeled biomarker antibodies to the biomarker proteins or polypeptides according to the invention as described herein.
[0098] Antibodies and other reagents may also be used to detect post- translational modifications (e.g., methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycosylation, sugar modification, lipidation, lipid modification, ubiquitination, sumolation, disulphide bonding, cysteinylation, oxidation, glutathionylation, carboxylation,
glucuronidation, and deamidation) of biomarker proteins or biomarker
polypeptides.
[0099] The invention encompasses a panel comprising antibodies that bind to proteins encoded by genes listed in Table 1 , Table 2 or Table 3. The panel may comprise antibodies that bind to proteins encoded by at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, at least 84 genes or all genes listed in Table 1 , Table 2 and Table 3. In one embodiment, the panel of antibodies is immobilized on a solid surface. Examples of solid surfaces include microspheres, plates, wells, slides, and beads (e.g., protein A or protein G agarose).
[0100] In addition to antibody-based techniques, the biomarkers described herein may also be detected and quantified by mass spectrometry. Mass spectrometry is a method that employs a mass spectrometer to detect ionized protein markers or ionized peptides as digested from the protein markers by measuring the mass-to- charge ratio (m/z). Labelling of biomarkers (along with other proteins) with stable heavy isotopes (deuterium, carbon- 13, nitrogen- 1 5, and oxygen- 18) can be used in quantitative proteomics. These are either incorporated metabolically in sample cells cultured briefly in vitro, or directly in samples by chemical or enzymatic reactions. Light and heavy labelled biomarker peptide ions segregate and their intensity values are used for quantification. For example, analytes may be introduced into an inlet system of the mass spectrometer and ionized in am ionization source, such as a laser, fast atom bombardment, plasma or other suitable ionization sources known to the art. The generated ions are typically collected by an ion optic assembly and introduced into mass analyzers for mass separation before their masses are measured by a detector. The detector then translates information obtained from the detected ions into mass-to-charge ratios.
[0101] The invention also encompasses methods that involve measuring the activity of a biomarker (e.g., enzymatic activity). Examples of enzymatic activity include, without limitation, kinase, phosphatase, protease, ubiquitination, oxidase and reductase activity.
[0102] The invention also provides compositions comprising biomarker polynucleotides, biomarker polypeptides, or biomarker antibodies according to the invention as described herein in the various embodiments. The invention further provides diagnostic or detection reagents for use in the methods of the invention, for example, reagents for flow cytometry and/or immunoassays that comprise fluorochrome-labeled antibodies that bind to one of the biomarker polypeptides of the invention. [0103] In one embodiment, the invention provides diagnostic or detection reagents that comprise one or more biomarker probes, or one or more biomarker primers. A diagnostic reagent may comprise biomarker probes and/or biomarker primers from the same biomarker gene or from multiple biomarker genes. In another embodiment, the invention also provides diagnostic compositions that comprise one or more biomarker probes and target biomarker polynucleotides, or one or more biomarker primers and target polynucleotides, or biomarker primers, biomarker probes and biomarker target polynucleotides. In some embodiments, the diagnostic compositions comprise biomarker probes and/or biomarker primers and a sample suspected to comprise biomarker target polynucleotides. Such diagnostic compositions comprise biomarker probes and/or biomarker primers and the nucleic acid molecules (including NA, mRNA, cRNA, cDNA, and/or genomic DNA) of a subject in need of a diagnosis/prognosis of COPD.
In Siiico Analysis and Computer Readable Media [0104] Biomarkers and gene signatures of the invention may be predicted based on gene expression patterns in COPD. In some embodiments, biomarker and gene signature prediction comprises gene expression patterns in control (e.g., non- COPD) biological samples. A heterogeneous ensemble learning approach may be used to classify samples based on their gene expression profiles. Such an approach may combine predictions from different approaches that use genes, gene set- derived features and/or causal network-derived features in order to get a classification and a prediction confidence for each classified sample. Methods that may be used to generate biomarkers and gene signatures of the invention include shrunken centroids, factor rotation, logistic regression models, network-based approaches, disease module-based approaches, linkage methods, modularity or pathway-based methods and diffusion-based methods.
[0105] The biological data (such as training data and test data) used in these methods may be drawn from the literature, databases (including data from preclinical, clinical and post-clinical trials of pharmaceutical products or medical devices), genome databases (genomic sequences and expression data, e.g. , Gene Expression Omnibus by National Center for Biotechnology Information or ArrayExpress by European Bioinformatics Institute (Parkinson et al . 2010, Nucl. Acids Res., doi : 10.1093/nar/gkql 040. Pubmed ID 21071405)), commercially available databases (e.g., Gene Logic, Gaithersburg, MD, USA) or experimental work. In one embodiment, the REACTOME, EGG or BIOCARTA pathway gene set collections from the Broad Institute (Cambridge, MA) may be used. The data may be related to nucleic acid (e.g. , absolute or relative quantities of specific DNA or RNA species, changes in DNA sequence, RNA sequence, changes in tertiary structure, or methyl ation pattern as determined by sequencing,
hybridization - particularly to nucleic acids on microarray, quantitative polymerase chain reaction, or other techniques known in the art), protein/peptide (e.g. , absolute or relative quantities of protein, specific fragments of a protein, peptides, changes in secondary or tertiary structure, or posttranslational modifications as determined by methods known in the art) and functional activities (e.g., enzymatic activities, proteolytic activities, transcriptional regulatory activities, transport activities, binding affinities to certain binding partners) under certain conditions, among others. Modifications, including posttranslational modifications of protein or peptide, can include, but are not limited to, methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycosylation, sugar modification, lipidation, lipid modification, ubiquitination, sumolation, disuiphide bonding, cysteinylation, oxidation, glutathionylation, carboxylation,
glucuronidation, and deamidation. In addition, a protein can be modified posttranslationally by a series of reactions such as Amadori reactions, Schiff base reactions, and Maillard reactions resulting in glycated protein products.
[0106] The test data sets may be processed and have their quality controlled together if they are obtained from the same technology platform (e.g., an
Affymetrix platform). For example, raw data files may be read by the ReadAffy function of the affy package (Gautier et al., Bioinformatics, 20:307-3 1 5 (2004)) belonging to Bioconductor (Gentleman et al., Genome Biol, 5(1 0):R80 (2004)) in R (R Development Core Team, R: A Language and Environment for Statistical Computing. 2007). The quality may be controlled by: 1. Generating RNA degradation plots (using the AffyRNAdeg function of the affy package (Gautier, 2004)), NUSE and RLE plots (using the function affyPLM) (Brettschneider et al., Technometrics, 50(3):241 -264 (2008)), calculating the MA(RLE) values;
2. Excluding arrays from the training datasets that fall below a set of thresholds on the quality control checks or that do not correspond to the test parameters;
or both.
Arrays passing quality control checks may be normalized using the gcrma algorithm (Wu et al., Journal of the American Statistical Association, 99:909
(2004)). If the datasets were obtained from a database, the samples classification may be obtained from the series matrix file of the same database for each dataset. The output of this part of the method may consist of: a gene expression matrix on training samples and test samples, probesets, and the class information for the training samples.
[0107] Non-limiting examples of methods that may be used to generate predictions are: transformation invariant (Tranlnv) (U.S. Provisional Patent Application entitled "Systems and Methods for Generating Biomarker Signatures with Integrated Bias Correction and Class Prediction," filed concurrently with the instant application and having the attorney docket no. 10(5500-0032-001 ), dual ensemble (Yang et al., Current Bioinformatics, 5(4):296-308 (2010)), generalized simulated annealing (Tsallis and Stariolo, Physica A: Statistical Mechanics and Its Applications, 233( 1 ):395-406 ( 1996); Xiang and Gong, Physical Review E, 62(3):4473 (2000); Xiang et al ., Physics Letters A, 233(3):216-220 ( 1 997); Xiang et al, The Journal of Physical Chemistry A, 104(12):2746-2751 (2000)), T-filter, CORG (Chuang et a!., Mol Syst Biol, 3 : 140 (2007)), single and pairs, dual bagging, forward learning, NPA (network perturbation amplitude) (see, e.g., International Patent Application No. PCT/EP2012/061 035, filed June 1 1 , 2012 and U.S.
Provisional Patent Application entitled "Systems and Methods Relating to
Network-Based Biomarker Signatures," filed concurrently with the instant application and having the attorney docket no. 1 06500-0022-001 ) and Laplacian based learning. Each of the foregoing patent applications and publications is incorporated herein by reference in its entirety.
[0108] Generalized simulated annealing may be modified for binary functions. In one embodiment, a dual binary generalized simulated annealing based method may be used (DualGensemble) (U.S. Provisional Patent Application entitled "Systems and Methods for Generating Biomarker Signatures with Integrated Dual Ensemble and Simulated Annealing Techniques," filed concurrently with the instant application, incorporated herein by reference in its entirety and having the attorney docket no. 106500-0031 -001 ). T-filter is a method of filtering genes based on the t-test by setting P-value and fold-change thresholds. CORG may be modified by calculating activity scores by leveraging the F-test instead of the T- test. CORG may also be combined with SVM. Dual bagging is a combination of bagging (Breiman, The Elements of Statistical Learning: Data Mining, Inference, and Prediction. 2nd ed., ed. T. Hastie, R. Tibshirani, and J. Friedman, (2009)) and the random subspace method (Bryll, Pattern Recognition, 20(6): 1291 - 1302 (2003) Ho, IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(8):832- 844 ( 1998); Skurichina, Pattern Analysis and Applications, 5(2): 121 - 135 (2002)).
[0109] The single and pairs method may include the following steps:
1 . Select a threshold ao,
2. For each pair of genes compute the leave-one-out cross-validation of a quadratic discriminant analysis and record the accuracy. Save the pair of genes if this accuracy is less than ao.
3. For each gene, compute the an and 1 - ao quantile of the values in each class to discriminate the classes A and B, Qao(A), Q i.a0(A), Qa0(B). Q i ao(B). Select the gene if either: Qa0(A) > Q,-a0(B) or Q,.a0(A) < Qa0(B).
4. Use the obtained list to train a classification algorithm on the reduced feature space.
5. Choose ao by cross-validation.
[01 101 The forward learning method may include the following steps:
Set IN to the empty list, choose N (for example, 20, 100 or 200).
For n =1 , . . . N do a. For each gene, g, not in IN, compute a randomForest (ntree=500) on the subspace corresponding to {IN,g} , record the out-of-the bag true positive rates (TPr) and true negative rates (TNr) and compute the g- performance ^TPr * TNr.
b. Select the gene, gmax, for which the g-performance is maximum and add it to the list lN:={IN,gmax} .
c. Then train a classification algorithm on the subspace given by IN. N is chosen by cross-validation.
[0111] The Laplacian based learning method may include the following steps:
1. Compute Spearman or Pearson correlation between the samples based on their gene expression profiles for both test and training data. Normalize the distance matrix obtained from the correlation matrix (Kij= ij/sqrt(Kii* jj)).
2. Compute the k-nearest neighbors of each sample (k chosen by cross-validation, usually k=2, 3.4,5)).
3. Define a graph with samples as nodes and put an edge between neighbors.
4. Create the (combinatorial) Laplacian of the graph and get its generalized inverse G, which is a positive definite kernel.
5. Extract main kernel principal components (KPC) from G and train a
Svm on it. The number of KPC's is chosen with rdetools package function rde (Braun et al., The Journal of Machine Learning Research, 9: 1875- 1908 (2008)).
6. Train a SVM with the training samples and get performance by cross-validation.
7. Predict the test cases.
[01 12] Network-based analysis can be combined with machine learning methods to generate predictions, for example, combining any one of CORG, dual bagging or T-filter with a network-based analysis.
[0 13] In some embodiments, methods used to generate predictions are further combined with another classification method (e.g., a method that is used for cross- validation). Non-limiting examples of classification methods include PAMR (Tibshirani et al, Proc Natl Acad Sci USA, 99(10):6567-6572 (2002)),
RandomForest (Breiman, Machine Learning, 45(l):5-32 (2001)), Linear
Discrimination Analysis (LDA), Eigengene-based Linear Discrimination Analysis (ELDA), Principal Components Analysis (PCA), Recursive Partitioning Tree (RPART), Support Vector Machine (SVM), K-Nearest Neighbors (KNN) (Bishop, Neural Networks for Pattern Recognition, ed. O.U. Press, 1995) and Partial Least Squares Discriminant Analysis (PLS.DA). In one embodiment, a network-based analysis that uses NPA may be combined with SVM (U.S. Provisional Patent Application entitled "Systems and Methods Relating to Network-Based Biomarker Signatures," filed concurrently with the instant application, incorporated herein by reference in its entirety and having the attorney docket no. 106500-0022-001.
[0114] In one embodiment, these methods may further include a step of oversampling to balance classes. The methods may include a step of filtering genes based on a simple T-test between the categories to be classified. The filtering step may reduce the number of genes to less than 1 ,500 or less than 2,000.
[0115] After predictions are generated by several methods, a vote may be made to obtain the classification as well as the confidence for the prediction of each sample of the sample set. If a method provides cross-validation results far below the other methods, it may be excluded. Such additional steps are contemplated in the methods of the invention.
[0116] The union of the gene signatures extracted by these methods may be considered as the larger gene signature. A weight may be given to genes to take into consideration the number of times they appear in a list. See, for example, in Table 1 the column "present/total lists" which shows the number of times each gene appears in one of the predicted gene signatures.
[0117] The genes obtained by these methods may be mapped to gene symbols using any suitable platform, for example, the Confero platform (Hermida et al., Confer o: an integrated Contrast and Gene Set Platform for Computational Analysis and Biological Interpretation of Omics Data, submitted, 2012).
[0118] The numerical methods for generating the gene signatures of the invention may include a testing step and confidence statistics for the genes. The testing step (or phase) is an exemplary use of the gene signature in carrying out the embodied method.
[0119] The invention encompasses a method for classifying a test sample as a COPD sample or a non-COPD sample, the method comprising: measuring the expression levels of at least 2 genes listed in Table 1 in a test sample; and applying one or more network-based methods, one or more machine-learning based methods, or a combination of the foregoing methods to the expression levels to obtain a classification of the test sample as cither a COPD sample or a non-COPD sample. In some embodiments, the expression levels of at least 2, at least 3, at least 4, at least 5, at least 10. at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, at least 84 genes or all genes listed in Table 1 are measured. In some embodiments, the classifier has been trained by in silico analysis or one or more feature selection and classification algorithms.
1 (1 1 70 1 On e fl snerf nf f li p i n ven t i on !irnm na a l i ¾t o f nn f nr mnr? lu nm nA prc or gene signatures of the invention stored on a computer readable medium. The absence, presence, activity or expression level of a biomarker in a biological sample (such as a control sample or a test sample) may also be stored on the computer readable medium. The computer readable medium may also include information that identifies the sample. The computer readable medium may also include a computer program product.
[0121 ] The computer program product may include a classifier based on at least two genes listed in Table 1. The classifier may be based on at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50. at least 55, at least 60, at least 65, at least 70, at least 75, at least 80, least about 85, at least 87 or all genes listed in Table 1.
[0122] Optionally, the classifier is trained by in silico analysis or one or more feature selection and classification algorithms. In some embodiments, the classifier is trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T-filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude. The classifier may be trained with at least the data in Gene Expression Omnibus datasets GSE10106, GSE10135, GSE 1 1906, GSE 1 1 952, GSE13933, GSE19407, GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545. Devices and Kits
[0123] One aspect of the invention encompasses devices useful for performing methods of the invention. For example, the devices may be used for diagnosing, classifying and/or grading COPD. The devices can comprise means for detecting the expression level of at least 2 genes listed in Table 1 or the level of at least 2 gene products of such genes in a test sample. Such means may include
components for performing one or more methods of nucleic acid extraction, nucleic acid amplification, nucleic acid detection, protein isolation and/or protein detection. Such components may include one or more of an amplification chamber (for example a thermal cycler), a plate reader, robotic sample handling
components, a capillary electrophoresis apparatus, a spectrophotometer, a mass spectrometer and/or a chip reader. These components can obtain data that reflects the expression level of the genes being analyzed. In some embodiments, the devices can comprise means for detecting the expression levels of at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1 . In some embodiments, the devices can comprise means for detecting the expression levels of the gene products of at least 2, at least 3, at least 4, at least 5, at least 1 0, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, at least 70, at least 75, at least 80 or at least 84 genes listed in Table 1.
[0124] The devices optionally comprise a means for identi fying a given test sample, and of linking the results obtained to that sample. Such means can include manual labels, barcodes, and other indicators which can be linked to a sample container or receptacle. Identification means may optionally be included in the sample itself, for example where an encoded particle is added to the sample. The results may be linked to the sample, for example in a computer memory that contains a sample designation and a record of expression levels obtained from the sample. Linkage of the results to the sample can also include a linkage to a particular sample container or receptacle in the device, which is also linked to the sample identity.
[0125] The devices may comprise an excitation and/or a detection means. Any instrument that provides a wavelength that can activate a label (e.g., fluorophore, fluorochrome and fluorescent dye) used on a detection reagent and is shorter than the emission wavelength(s) to be detected can be used for excitation. Examples of excitation sources include a broadband ultraviolet light source such as a deuterium lamp with an appropriate filter, the output of a white light source such as a xenon lamp or a deuterium lamp after passing through a monochromator to extract out the desired wavelcngth(s), a continuous wave (cw) gas laser, a solid state diode laser, or any pulsed lasers. Emitted light can be detected through any suitable component or technique; many suitable approaches are known in the art. For example, a fluorimeter or spectrophotometer may be used to detect whether the test sample emits light of a wavelength characteristic of a label used in a method of the invention.
[0126] The devices may comprise a means for correlating the expression levels of the genes being analyzed with COPD status, prognosis, grade and/or
classification. Such means may comprise one or more of a variety of correlative techniques, including lookup tables, algorithms, multivariate models, and linear or nonlinear combinations of expression models or algorithms, such as any of the in silico and machine learning methods described above. The expression levels may be converted to one or more biomarker scores, indicating that the individual providing the sample is not suffering from COPD or is suffering from stage 1 , stage 2, stage 3 or stage 4 COPD. The models and/or algorithms can be provided in computer readable format.
[0127] The devices may also comprise output means for outputting the COPD status, prognosis, grade and/or classification. Such output means can take any form which transmits the results to an individual and/or a healthcare provider, and may include a monitor, a display, and/or a printer. Output means may record the results to a computer readable medium. The device may use a computer system for performing one or more of the steps provided.
[0128] Another aspect of the invention encompasses kits for practicing the methods of the invention. Such kits may be used for diagnosing, classifying and grading COPD or for assessing the prognosis of COPD in an individual. The kits can be used for ciinical diagnosis and/or laboratory research. In one embodiment, a kit comprises in one or more containers one or more reagents that detect expression levels of genes that serve as biomarkers of COPD in a test sample. Preferably, the kit also comprises instructions in any tangible medium (e.g., written, tape, CD-ROM, DVD) on the use of the detection reagent(s) in one or more methods of the invention.
[0129] For nucleic acid-based methods (for example, amplification assays, hybridization assays, sequencing or polymerase chain reactions), a detection reagent in the kit may comprise at least one polynucleotide, probe, and/or primer specific for the COPD genes listed in Table 1 . The nucleic-acid based detection reagents may comprise sequences complementary to a portion of the signature genes or sequences that are portions of the signature genes. Such a kit may optionally provide in separate containers enzymes and/or buffers for reverse transcription, in vitro transcription, and/or DNA polymerization, nucleotides, and/or labeled nucleotides.
[0130] For protein-based methods, such as immunoassays, a detection reagent in the kit may comprise a biomarker antibody, which may be labeled or labelable. The antibodies may bind to proteins encoded by the COPD genes listed in Table 1 . In one embodiment, the detection reagents recognize a post-translational modification (e.g., methylation, acetylation, farnesylation, biotinylation, stearoylation, formylation, myristoylation, palmitoylation, geranylgeranylation, pegylation, phosphorylation, sulphation, glycosylation, sugar modification, lipidation, lipid modification, ubiquitination, sumolation, disulphide bonding, cysteinylation, oxidation, glutathionylation, carboxylation, giucuronidation, and deamidation) of a protein encoded by a gene selected from the genes listed in Table 1 . For protein-based methods that involve measuring the activity of a biomarker (e.g., enzymatic activity), the kit may include a substrate for the biomarker and a detection reagent that recognizes the products and/or byproducts of the activity being measured. Such a kit may optionally provide, in separate containers, buffers, secondary antibodies, signal generating accessory molecules, and/or labeled secondary antibodies, including fluorochrome-labeled secondary antibodies. The kit may also include unlabeled or labeled antibodies to various cell surface antigens which can used for identification or sorting of subpopulations of cells.
[0131] The detection reagents may be labeled or labelable by one or more detectable labels. Examples of detectable labels include, without limitation, radiolabels (e.g. radioactive nuclides), dyes, fluorescent proteins or materials (e.g., fluorochromes, fluorophores, fluorescein and rhodamine), luminescent proteins or materials, bioluminescent proteins or materials (e.g., luciferase, aequorin and luciferin), enzymes (e.g., beta-galactosidase, alkaline phosphatase, horseradish peroxidase and acetylcholinesterase) and prosthetic groups (e.g., biotin, streptavidin and avidin).
[01321 The detection reagents in the kit may be immobilized on a solid surface or packaged separately with reagents to immobilize them on a solid surface.
[0133] Also included in the kit may be positive and negative controls for the methods of the invention. The positive and/or negative controls included in a kit can be nucleic acids, polypeptides, ceil lysate, cell extract, whole cells from patients, or whole cells from cell lines.
Example 1 Generation of CO P I) Gene Signatures
[0134] This example is for the purpose of illustration only and is not to be construed as limiting the scope of the invention in any way. A heterogeneous ensemble learning approach aimed at classifying samples based on their gene expression profile is applied to extract genes whose expression levels allow distinguishing COPD from healthy samples. In summary, predictions from different approaches that use genes and gene sets-derived features are combined to get the most accurate classifier possible. Gene lists are extracted from these methods and combined to generate the list presented in Table 1 .
[0135] A schematic overview of the data and strategy used to generate the list given in Table 1 is given below: 1. Public datasets and a dataset downloaded from the SBV diagnostic signature challenge (http://sbvimprover.com) are used as the source of gene expression data.
The following public datasets are downloaded from the Gene Expression Omnibus (GEO) (http://www.ncbi.nlm.nih.gov/geo/) repository:
a. GSE 10106 (www.ncbi.nlm.nih. go v/geo/query/acc.cgi?acc=GSE 10106) b. GSE10135 (www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE 10135) c. GSE1 1906 (www.ncbi.nlm. nih.gov/geo/query/acc. cgi?acc=GSEl 1906) d. GSE1 1952 ( www.ncbi. nlm.nih.gov/geo/query/acc. cgi?acc=GSE l 1952) e. GSE 13933 (www.ncbi.nlm.nih.gov/gco/query/acc.cgi?acc=GSE 13933) f. GSE19407 (www.ncbi.nlm.nih. gov/geo/query/acc.cgi?acc=GSE 19407) g. GSE 19667 (www.ncbi.nlm. nih.gov/geo/query/acc. cgi?acc=GSE19667) h. GSE20257 ( www.ncbi.nlm. nih. go v/geo/query/acc.cgi?acc=GSE20257) i. GSE5058 ( www.ncbi.nlm. nih. gov/geo/query/acc.cgi?acc=GSE5058) j . GSE7832 (www.ncbi.nlm. nih. gov/geo/query/acc.cgi?acc=GSE7832) k. GSE8545 (www.ncbi.nlm.nih.gov/geo/query/acc,cgi?acc:=GSE8545).
2. As both training datasets are on the same Affymetrix platform as the test dataset (HGU- 133 + 2), they are processed and have their quality controlled together. In summary, raw data files are read by the ReadAffy function of the affy package (Gautier, 2004) belonging to Bioconductor (Gentleman, 2004) in R (R Development Core Team, 2007), and the quality is controlled by:
a. generating RNA degradation plots (with the AffyRNAdeg function of the a fly package), NUSE and RLE plots (with the function affyPLM
(Brettschneider, 2008)), and calculating the MA(RLE) values;
b. excluding arrays from the training datasets that fell below a set of thresholds on the quality control checks or that are duplicated in the above datasets; and
c. normalizing arrays that pass quality control checks using the gcrma algorithm (Wu, 2004). Training set sample classifications are obtained from the series matrix file of the GEO database for each dataset.
Arrays GSM 101 108.CEL, GSM252820.CEL, GSM252822.CEL,
GSM252824.CEL, GSM252826.CEL, GSM298246.CEL, GSN 252819.CEL, GSM252821 .CEL, GSM252842.CEL, GSM252843.CEL, GSM252844.CEL, GSM252845.CEL, GSM252846.CEL, GSM252847.CEL, ,GSM252848.CEL, GSM252849.CEL, GSM252850.CEL, GSM252851 .CEL, GSM252852.CEL, GSM252853.CEL, GSM252854.CEL, GSM1 14096.CEL, GSM 1 14098. CEL, GSM1 14099. CEL, GSMl 14100.CEL, GS 1 14101.CEL, GSM 1 14102. CEL, GSM l 1 4103. CEL, GSM1 14104. CEL, GSM 1 14105. CEL, GSM 1 14091 .CEL, GSM300871.CEL, GSM300872.CEL, GSM300873.CEL, GSM300874.CEL, GSM300875.CEL, GSM300876.CEL, GSM300877.CEL, GSM300878.CEL, GSM300879.CEL, GSM300880.CEL, GSM491043.CEL, GSM491044.CEL, GSM300862.CEL, GSM300863.CEL, GSM300864.CEL, GSM300865.CEL, GSM300866.CEL, GSM300867.CEL, GSM300868.CEL, GSM300869.CEL, GSM300870.CEL, GSM252856.CEL, GSM252857.CEL, GSM252861 .CEL, GSM252872.CEL, GSM252864.CEL, GSM l 01 1 1 1 .CEL, GSM252865.CEL, GSM252874.CEL, GSM252866.CEL, GSM190152.CEL, GSM252862.CEL, GSM252863.CEL, GSM252860.CEL, GSM252858.CEL, GSM252855.CEL, GSM252870.CEL, GSM252832.CEL, GSM252875.CEL, GSM252829.CEL, GSM252873.CEL, GSMl O l 109. CEL, GSM434064.CEL were not used for further analysis. The output of this part consists of a gene expression matrix X on 273 samples (233 training samples and 40 test samples) and 54675 probesets, and the class information for the training samples.
3. Features selection and classification algorithm(s) used for prediction of a gene signature follow the illustrated strategy of Figure 1 :
Briefly, a set of feature selection and classification algorithms are analyzed to obtain a number of classifications for each test sample. Each method has defined input and output:
INPUT: gene expression matrix Xnxp on n samples and p genes, training samples and test samples, and the class information for the training samples
OUTPUT: Class prediction for each test sample and a list of genes involved. [0136] Prior to applying features selection and classification methods, the following steps are performed: (1 ) oversampling is, optionally, used to balance classes in the training dataset; (2) mapping probe sets to gene symbols (Entrcz gene ids) using Confero platform (Hermida, 2012); and (3) optionally filtering the genes in the matrix based on simple T-test between the categories to be classified so that less than 1500 genes (for Dual Ensemble or T-filter methods) or less than 2000 genes (for the other methods) remain.
[0137] Cross validation of the results is performed using any of the following supervised methods
* PAMR (Tibshirani, 2002)
* RandomForest (Brciman, 2001 )
« Linear Discriminant Analysis (LDA)
* Support Vector Machine (SVM)
* K-Nearest Neighbors (KNN) (Bishop, i 995)
* Partial Least Squares Discriminant Analysis (PLS.DA)
[0138] The following methods are used to generate predictions:
<t. Dual Ensemble method.
This dual ensemble method builds ensemble of multiple classi fication algorithms applied in randomly perturbed data. The diversity of the ensemble classifier is imposed by using different classification algorithms and is further enhanced by data-level perturbation. See, e.g., Yang, 201 0. A molecular profile of a training dataset, TO. train and its associated phenotype cl. train (control and treatment) are used as input. The molecular profile of the test set TO. test is used to predict the phenotype cl.tcst.
b. T-filter method
Genes are filtered based on t-test to obtain a list of N genes, by setting P- value and fold-change thresholds. Thresholds are decreased (resp. increase) automatically if the list size is over N. Any M is trained on the resulting subspace, N is chosen by cross-validation,
e. CORG-modified method
This method is modified from CORG method (Chuang, 2007) as activity scores are calculated by leveraging F-test instead of T-test. It uses the c2.cp gene sets collection from the Broad Institute (Cambridge, MA) (Reactome, KEGG and Biocarta pathways). d. Single and Pairs method
In this method,
1 . A threshold (ao) is selected.
2. For each pair of genes the leave-one-out cross-validation of a
5 quadratic discriminant analysis is computed and the accuracy recorded. If this accuracy is less than ao, the pair of genes is saved.
3. For each gene, the ao and 1 - ao quantile of the values in each class to discriminating the classes A and B, Qa0(A), Q i-ao(A), Qa0(B), Q i -ao(B), is computed. If either: Qa0(A)> Qi-a0(B) or Qi-a0(A)< Qao(B), the gene is
10 selected.
4. The obtained list is used to train M on the reduced feature space.
5. ao is chosen by cross-validation.
e. Forward Learning method
In this method,
I J i . o t i lu ^n ip ^ l iS l, CnOO IN ^ L Vp lCdl i ZO , l UV , - 'J'J ) .
For n =1 , . . N do
a. For each gene, g, not in IN, compute a randomForest (ntree=500) on the subspace corresponding to { IN,g} , record the out-of-the bag true positive rates (TPr) and true negative rates (TNr) and compute the g- 0 performance V'TPr * TNr.
b. Select the gene, gmax, for which the g-performance is maximum and add it to the list IN:={IN,gmax} .
c. Then train M on the subspace given by IN. N is chosen by cross- validation.
5 [0139] The union of the gene signatures extracted from the results of the foregoing methods is considered as the larger gene signature. A weight is given to each of the genes in the union of the gene signature on the basis of the number of times it appears in a generated signature and the length of each generated gene signature. Genes in Table 1 are those that appear in at least 3 of 0 5 lists generated. A further reduction of the number of genes in the signature can be obtained by selecting the genes which has a fold-change value greater than a threshold (see Table 1 , the column "fc>2" wherein a value of 1 is ascribed to those genes that has a fold change value greater than 2). Further, the genes in Table 1 that appear in at least 4 of the 5 lists generated are more predictive of lung cancer status than those appearing in 3 of the 5 lists generated.
[0140] While implementations of the invention have been particularly shown and described with reference to specific examples, it should be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the disclosure.
Table 1 : Genes which are specifically regulated in airway brushings of COPD patients compared with control airway brushings
Figure imgf000062_0001
Figure imgf000063_0001
Figure imgf000064_0001
Figure imgf000065_0001
Figure imgf000066_0001
Figure imgf000067_0001
Table 2: Genes which are specifically upregulated in airway brushings of COPD patients compared with control airway brushings
Figure imgf000068_0001
Figure imgf000069_0001
Figure imgf000070_0001
Table 3: Genes which are specifically downregulated in airway brushings of COPD patients compared with control airway brushings
Figure imgf000071_0001
Figure imgf000072_0001
Figure imgf000073_0001

Claims

We claim:
1 . A method of diagnosing an individual as being at risk for or having Chronic Obstructive Pulmonary Disease (COPD) comprising detecting the expression level of at least 2 of the genes listed in Table 1 in a test sample obtained from the individual; wherein a differential pattern of expression levels of said at least 2 genes in the test sample diagnoses the individual as suffering from COPD.
2. The method according to claim 1 , wherein the differential pattern of expression levels is identified by a classifier based on a plurality of genes listed in Table 1 , including said at least two genes, said classifier having been trained by in silico analysis or one or more feature selection and classification algorithms.
3. The method according to claim 1 or 2, wherein the differential pattern of expression levels is identified by a classifier based on a plurality of genes listed in Table 1 , including said at least two genes, said classifier having been trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T-filter, CORG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude.
4. The method according to any one of claims 1 -3, wherein said classifier having been trained with at least the data in the Gene Expression Omnibus datasets GSE 10106, GSE 101 35, GSE1 1906, GSE1 1952, GSE 13933, GSE 19407,
GSE19667, GSE20257, GSE5058, GSE7832, and GSE8545.
5. The method according to any one of claims 1 -4, wherein the method further comprises comparing the expression level of said at least 2 genes in the test sample and a control sample; or detecting the expression level of said at least 2 genes in the control sample and comparing the expression level of said at least 2 genes in the test sample and control sample, to identify the differential pattern.
6. The method according to any one of claims 1 -5, wherein said at least 2 genes are selected from the group consisting of: PROS 1 , IRAK I , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR1 A, PAPD4, VCL, PPP2R5C, DG A, and CYP51 A1.
7. The method according to any one of claims 1 -6, wherein the expression level of said at least 2 genes in the test sample are detected by using a human genome-wide array, a human lung tissue array or a custom array comprising polynucleotides of a plurality of genes in Table 1 and said at least 2 genes.
8. The method according to any one of claims 1 -6, wherein the expression level of said at least 2 genes in the test sample are detected by measuring the level of proteins encoded by the genes.
9. An array comprising polynucleotides hybridizing to at least 2 COPD signature genes immobilized on a solid surface, wherein the COPD signature genes are selected from the genes listed in Table 1 and said array is not a human genome-wide array.
10. A device comprising antibodies immobilized on a solid surface that bind to proteins encoded by at least 2 COPD signature genes, wherein the COPD signature genes are selected from the genes listed in Table 1.
1 1 . A computer readable medium or computer program product comprising a classifier based on at least two genes listed in Table 1 , said classifier having been trained by in silica analysis or one or more feature selection and classification algorithms.
12. The computer readable medium or computer program product according to claim 1 1 , wherein said classifier is trained by one or more algorithms selected from the group consisting of dual ensemble, generalized simulated annealing, T- filter, COPvG, CORG combined with support vector machine, dual bagging, single and pairs, forward learning, Laplacian based learning and learning method based on network perturbation amplitude.
1 3. The computer readable medium or computer program product according to claim 1 1 or 12, wherein said classifier is trained with at least the data in the Gene Expression Omnibus datasets GSE1 0106, GSE10135, GSE1 1906, GSE1 1952, GSE13933 , GSE19407, GSE19667, GSE20257, GSE5058, GSE7832, and
GSE8545.
14. The computer readable medium or computer program product according to any one of claims 1 1- 13 , wherein said at least two genes are selected from the group consisting of PROS 1 , IRAKI , VAV3, FUT3, SFN, ZBTB44, CLDN8, BMPR1 A, PAPD4, VCL, PPP2R5C, DGKA, and CYP51A 1 .
15. A kit for classifying and grading COPD or for assessing the prognosis of COPD, comprising one or more reagents that detects expression levels of the at least 2 of the genes listed in Table 1 in a test sample and instructions for using said kit for classifying and grading COPD in said individual or for determining the prognosis of the COPD in said individual.
PCT/EP2013/062996 2012-06-21 2013-06-21 Gene signatures for copd diagnosis WO2013190092A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201261662779P 2012-06-21 2012-06-21
US61/662,779 2012-06-21

Publications (1)

Publication Number Publication Date
WO2013190092A1 true WO2013190092A1 (en) 2013-12-27

Family

ID=48803505

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2013/062996 WO2013190092A1 (en) 2012-06-21 2013-06-21 Gene signatures for copd diagnosis

Country Status (1)

Country Link
WO (1) WO2013190092A1 (en)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014186036A1 (en) * 2013-03-14 2014-11-20 Allegro Diagnostics Corp. Methods for evaluating copd status
WO2015185656A1 (en) * 2014-06-05 2015-12-10 Medizinische Universität Wien Methods of diagnosing chronic obstructive pulmonary disease (copd) using novel molecular biomarkers
WO2015185653A3 (en) * 2014-06-05 2016-02-11 Medizinische Universität Wien Methods of diagnosing chronic obstructive pulmonary disease (copd) using novel molecular biomarkers
WO2015185658A3 (en) * 2014-06-05 2016-02-11 Medizinische Universität Wien Methods of diagnosing chronic obstructive pulmonary disease (copd) using novel molecular biomarkers
CN108257675A (en) * 2018-02-07 2018-07-06 平安科技（深圳）有限公司 Chronic obstructive pulmonary disease onset risk Forecasting Methodology, server and computer readable storage medium
WO2020145041A1 (en) 2019-01-11 2020-07-16 日本たばこ産業株式会社 In vitro evaluation method for risk of chronic obstructive pulmonary disease associated with smoking or inhalation
CN111553338A (en) * 2020-04-27 2020-08-18 南通智能感知研究院 Hyperspectral feature selection method based on simulated annealing algorithm
CN113633317A (en) * 2021-08-12 2021-11-12 生态环境部华南环境科学研究所 Exhaled air biomarker capable of identifying human health state
US11639527B2 (en) 2014-11-05 2023-05-02 Veracyte, Inc. Methods for nucleic acid sequencing
US11976329B2 (en) 2021-12-21 2024-05-07 Veracyte, Inc. Methods and systems for detecting usual interstitial pneumonia

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006105252A2 (en) * 2005-03-28 2006-10-05 The Regents Of The University Of Colorado Diagnosis of chronic pulmonary obstructive disease and monitoring of therapy using gene expression analysis of peripheral blood cells
WO2008109773A2 (en) * 2007-03-06 2008-09-12 Cornell Research Foundation, Inc. Chronic obstructive pulmonary disease susceptibility and related compositions and methods
US20090186951A1 (en) 2007-09-19 2009-07-23 Brody Jerome S Identification of novel pathways for drug development for lung disease
US20100070191A1 (en) 2008-09-09 2010-03-18 Somalogic, Inc. Lung Cancer Biomarkers and Uses Thereof
US20110160070A1 (en) 2008-03-10 2011-06-30 Lineagen, Inc. Copd biomarker signatures

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006105252A2 (en) * 2005-03-28 2006-10-05 The Regents Of The University Of Colorado Diagnosis of chronic pulmonary obstructive disease and monitoring of therapy using gene expression analysis of peripheral blood cells
WO2008109773A2 (en) * 2007-03-06 2008-09-12 Cornell Research Foundation, Inc. Chronic obstructive pulmonary disease susceptibility and related compositions and methods
US20100119474A1 (en) 2007-03-06 2010-05-13 Cornell University Chronic obstructive pulmonary disease susceptibility and related compositions and methods
US20090186951A1 (en) 2007-09-19 2009-07-23 Brody Jerome S Identification of novel pathways for drug development for lung disease
US20110160070A1 (en) 2008-03-10 2011-06-30 Lineagen, Inc. Copd biomarker signatures
US20100070191A1 (en) 2008-09-09 2010-03-18 Somalogic, Inc. Lung Cancer Biomarkers and Uses Thereof

Non-Patent Citations (40)

* Cited by examiner, † Cited by third party
Title
"Current Protocols in Molecular Biology", 1989, JOHN WILEY & SONS, pages: 6.3.1 6 - 13.3.6
"Product Insert of Proteome Profiler Array "Human Soluble Receptor Array Kit" antibody array described in XP055076598", 1 February 2012 (2012-02-01), XP055076601, Retrieved from the Internet <URL:http://www.rndsystems.com/pdf/ary012.pdf> [retrieved on 20130827] *
"R Development Core Team", R: A LANGUAGE AND ENVIRONMENT/OR STATISTICAC COMPUTING, 2007
"Shows that not genome-wide array HU6800 comprises probe for Fut3. The array is disclosed in XP008110295", 1 January 2001 (2001-01-01), XP055076507, Retrieved from the Internet <URL:http://www.affymetrix.com> [retrieved on 20130826] *
"Shows that not genome-wide array HU6800 comprises probe for IRAK1. The array is disclosed in XP008110295", 1 January 2001 (2001-01-01), XP055076505, Retrieved from the Internet <URL:http://www.affymetrix.com> [retrieved on 20130826] *
"Shows that not genome-wide array HU6800 comprises probe for PROS1. The array is disclosed in XP008110295", 1 January 2001 (2001-01-01), XP055076503, Retrieved from the Internet <URL:http://www.affymetrix.com> [retrieved on 20130826] *
AUSUBEL, CURRENT PROTOCOLS IN MOLECULAR BIOLOGY
BISHOP: "Neural Networks for Pattern Recognition", 1995, O.U. PRESS
BRAUN ET AL., THE JOURNAL OF MACHINE LEARNING RESEARCH, vol. 9, 2008, pages 1875 - 1908
BREIMAN, MACHINE LEARNING, vol. 45, no. 1, 2001, pages 5 - 32
BREIMAN: "The Elements o/Statistical Learning: Data Mining, Inference, and Prediction. 2nd ed.,", 2009
BRETTSCHNEIDER ET AL., TECHNOMETRICS, vol. 50, no. 3, 2008, pages 241 - 264
BRYLL, PATTERN RECOGNITION, vol. 20, no. 6, 2003, pages 1291 - 1302
BUTLER, J.E., METH. ENZYMOL., vol. 73, 1981, pages 482 - 523
CHUANG ET AL., MOL SYST BIOL, vol. 3, 2007, pages 140
ELISA PENNA ET AL: "microRNA-214 contributes to melanoma tumour progression through suppression of TFAP2C", THE EMBO JOURNAL, vol. 30, no. 10, 18 May 2011 (2011-05-18), pages 1990 - 2007, XP055076598, ISSN: 0261-4189, DOI: 10.1038/emboj.2011.102 *
GAUTIER ET AL., BIOINFORMATICS, vol. 20, 2004, pages 307 - 315
GENTLEMAN ET AL., GENOME BIOL, vol. 5, no. 10, 2004, pages R80
GEORGE L MUTTER ET AL: "Global Expression Changes of Constitutive and Hormonally Regulated Genes During Endometrial Neoplastic Transformation", GYNECOLOGIC ONCOLOGY, ACADEMIC PRESS, LONDON, GB, vol. 83, no. 2, 1 November 2001 (2001-11-01), pages 177 - 185, XP008110295, ISSN: 0090-8258, [retrieved on 20020304], DOI: 10.1006/GYNO.2001.6352 *
GOSLING, J. P.: "Immunoassays: A Practical Approach", 2001, OXFORD UNIVERSITY PRESS
HARLOW; LANE: "Antibodies Laboratory Manual", 1988
HERMIDA ET AL., CONFERO: AN INTEGRATED CONTRAST AND GENE SET PLATFÓRMFOR COMPUTATIONAL ANALYSIS AND BIOLOGICAL INTERPRETATION OF OMICS DATA, 2012
HO, IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, vol. 20, no. 8, 1998, pages 832 - 844
KLEIN ET AL., J EXP MED, vol. 194, 2001, pages 1625 - 1638
LIU L ET AL: "Detection of circulating cancer cells in lung cancer patients with a panel of marker genes", BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS, ACADEMIC PRESS INC. ORLANDO, FL, US, vol. 372, no. 4, 8 August 2008 (2008-08-08), pages 756 - 760, XP027181572, ISSN: 0006-291X, [retrieved on 20080602], DOI: 10.1016/J.BBRC.2008.05.101 *
MAGGIO, E.: "Enzyme Immunoassay", 1980, CRC PRESS
PANICI P ET AL: "Predictive value of multiple tumor marker assays in second-look procedures for ovarian cancer", GYNECOLOGIC ONCOLOGY, ACADEMIC PRESS, LONDON, GB, vol. 35, no. 3, 1 December 1989 (1989-12-01), pages 286 - 289, XP026231880, ISSN: 0090-8258, [retrieved on 19891201], DOI: 10.1016/0090-8258(89)90064-4 *
PARKINSON ET AL., NUCL. ACIDS RES., 2010
RICHARD A. SHIMKETS: "Gene Expression Profiling: Methods and Protocols", 2004, HUMANA PRESS
SAMBROOK ET AL.: "Molecular Cloning", 1989, COLD SPRING HARBOR PRESS, pages: 9.47 - 9.57
SCHULZE, NATURE CELL BIOL., vol. 3, 2001, pages EI90
SKURICHINA, PATTERN ANALYSIS ANDAPPLICATIONS, vol. 5, no. 2, 2002, pages 121 - 135
TIBSHIRANI ET AL., PROC NATL ACAD SCI USA, vol. 99, no. 10, 2002, pages 6567 - 6572
TSALLIS; STARIOLO, PHYSICA A: STATISTICAL MECHANICS AND ITS APPLICATIONS, vol. 233, no. I, 1996, pages 395 - 406
VOLLER, A. ET AL., J. CLIN. PATHOL., vol. 31, 1978, pages 507 - 520
WU ET AL., JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, vol. 99, 2004, pages 909
XIANG ET AL., PHYSICS LETTERS A, vol. 233, no. 3, 1997, pages 216 - 220
XIANG ET AL., THE JOURNAL OF PHYSICAL CHEMISTRY A, vol. 104, no. 12, 2000, pages 2746 - 2751
XIANG; GONG, PHYSICAL REVIEW E, vol. 62, no. 3, 2000, pages 4473
YANG ET AL., CURRENT BIOINFORMATICS, vol. 5, no. 4, 2010, pages 296 - 308

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10526655B2 (en) 2013-03-14 2020-01-07 Veracyte, Inc. Methods for evaluating COPD status
WO2014186036A1 (en) * 2013-03-14 2014-11-20 Allegro Diagnostics Corp. Methods for evaluating copd status
US10526656B2 (en) 2014-06-05 2020-01-07 Transgenion—International Institute for Regenerative Translational Medicine GmbH Methods of diagnosing chronic obstructive pulmonary disease (COPD) using novel molecular biomarkers
US11124835B2 (en) 2014-06-05 2021-09-21 Transgenion—International Institute for Regenerative Translational Medicine GmbH Methods of diagnosing chronic obstructive pulmonary disease (COPD) using novel molecular biomarkers
JP2017519521A (en) * 2014-06-05 2017-07-20 トランスゲニオン−インターナショナル インスティテュート フォー リジェネレイティヴ トランスレイショナル メディシン ゲーエムベーハー Method for diagnosing chronic obstructive pulmonary disease (COPD) using a novel molecular biomarker
JP2017519523A (en) * 2014-06-05 2017-07-20 トランスゲニオン−インターナショナル インスティテュート フォー リジェネレイティヴ トランスレイショナル メディシン ゲーエムベーハー Method for diagnosing chronic obstructive pulmonary disease (COPD) using a novel molecular biomarker
JP2017519522A (en) * 2014-06-05 2017-07-20 トランスゲニオン−インターナショナル インスティテュート フォー リジェネレイティヴ トランスレイショナル メディシン ゲーエムベーハー Method for diagnosing chronic obstructive pulmonary disease (COPD) using a novel molecular biomarker
WO2015185658A3 (en) * 2014-06-05 2016-02-11 Medizinische Universität Wien Methods of diagnosing chronic obstructive pulmonary disease (copd) using novel molecular biomarkers
US10508307B2 (en) 2014-06-05 2019-12-17 Transgenion—International Institute for Regenerative Translational Medicine GmbH Methods of diagnosing chronic obstructive pulmonary disease (COPD) using novel molecular biomarkers
WO2015185653A3 (en) * 2014-06-05 2016-02-11 Medizinische Universität Wien Methods of diagnosing chronic obstructive pulmonary disease (copd) using novel molecular biomarkers
WO2015185656A1 (en) * 2014-06-05 2015-12-10 Medizinische Universität Wien Methods of diagnosing chronic obstructive pulmonary disease (copd) using novel molecular biomarkers
US10604806B2 (en) 2014-06-05 2020-03-31 Transgenion—International Institute for Regenerative Translational Medicine GmbH Methods of diagnosing chronic obstructive pulmonary disease (COPD) using novel molecular biomarkers
US11639527B2 (en) 2014-11-05 2023-05-02 Veracyte, Inc. Methods for nucleic acid sequencing
CN108257675A (en) * 2018-02-07 2018-07-06 平安科技（深圳）有限公司 Chronic obstructive pulmonary disease onset risk Forecasting Methodology, server and computer readable storage medium
WO2020145041A1 (en) 2019-01-11 2020-07-16 日本たばこ産業株式会社 In vitro evaluation method for risk of chronic obstructive pulmonary disease associated with smoking or inhalation
KR20210102330A (en) 2019-01-11 2021-08-19 니뽄 다바코 산교 가부시키가이샤 Methods for In Vitro Assessment of Risk of Chronic Obstructive Pulmonary Disease Caused by Smoking or Aspiration
JPWO2020145041A1 (en) * 2019-01-11 2021-11-18 日本たばこ産業株式会社 In vitro assessment of the risk of chronic obstructive pulmonary disease from smoking or inhalation
JP7203124B2 (en) 2019-01-11 2023-01-12 日本たばこ産業株式会社 In vitro assessment method for risk of chronic obstructive pulmonary disease from smoking or inhalation
CN111553338A (en) * 2020-04-27 2020-08-18 南通智能感知研究院 Hyperspectral feature selection method based on simulated annealing algorithm
CN113633317A (en) * 2021-08-12 2021-11-12 生态环境部华南环境科学研究所 Exhaled air biomarker capable of identifying human health state
US11976329B2 (en) 2021-12-21 2024-05-07 Veracyte, Inc. Methods and systems for detecting usual interstitial pneumonia

Similar Documents

Publication Publication Date Title
WO2013190092A1 (en) Gene signatures for copd diagnosis
JP5405110B2 (en) Methods and materials for identifying primary lesions of cancer of unknown primary
CN103733065B (en) Molecular diagnostic assay for cancer
JP6309019B2 (en) Compositions and methods for diagnosing thyroid tumors
EP3071973B1 (en) Triaging of patients having asymptomatic hematuria using genotypic and phenotypic biomarkers
EP2864500B1 (en) Molecular malignancy in melanocytic lesions
Chen et al. Robust transcriptional tumor signatures applicable to both formalin-fixed paraffin-embedded and fresh-frozen samples
TWI582236B (en) Prognosis prediction for melanoma cancer
JP2011523049A (en) Biomarkers for head and neck cancer identification, monitoring and treatment
US8030060B2 (en) Gene signature for diagnosis and prognosis of breast cancer and ovarian cancer
WO2013086429A2 (en) Methods and compositions for classification of samples
Matsui Genomic biomarkers for personalized medicine: development and validation in clinical studies
CN109477145A (en) The biomarker of inflammatory bowel disease
WO2012125712A2 (en) Lung tumor classifier for current and former smokers
WO2012075069A2 (en) Signatures and determinants associated with cancer and methods of use thereof
US20190024184A1 (en) Distinguishing metastatic-lethal prostate cancer from indolent prostate cancer using methylation status of epigenetic markers
CN113430269A (en) Application of biomarker in prediction of lung cancer prognosis
EP3728630A1 (en) Compositions and methods for diagnosing lung cancers using gene expression profiles
WO2013190090A1 (en) Gene signatures for classifying and grading lung cancer
CN113388683A (en) Biomarker related to lung cancer prognosis and application thereof
CN114507717A (en) Method for predicting bile duct cancer recurrence by combining multiple mRNAs and application thereof
EP2607494A1 (en) Biomarkers for lung cancer risk assessment
US20230314408A1 (en) Immune profiling and methods of using same to predict responsiveness to an immunotherapy and treat cancer
Biasci et al. and Hepatology
CN113234831A (en) Model, product and system for predicting lung cancer prognosis

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13739154

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13739154

Country of ref document: EP

Kind code of ref document: A1