KR20190137012A - Method for predicting disease risk based on analysis of complex genetic information - Google Patents

Abstract

The present invention relates to a method for predicting a disease risk based on analysis of complex genetic information networks. The method for predicting a disease risk based on analysis of complex genetic information networks developed in accordance with the present invention can derive stable correlations with diseases from a small number of genetic information combinations through the introduction of optimization or learning method, and provides genetic information correlations based on a network model. By using the correlations between the genetic information and the disease derived from the present invention, it is expected to secure a diagnostic technology that satisfies the level of accuracy and economic feasibility that is commercially available in the real medical industry. In addition, a biomarker derived from the present invention is expected to be efficiently used for the prognosis of diseases through the manufacture of medical devices including a diagnostic chip and a terminal and commercialization as a disease diagnosis service. The method includes a step of extracting complex genetic information; a step of building a complex genetic information library; a step of deriving a disease state specific biomarker; and a step of predicting a risk.

Description

Method for predicting disease risk based on analysis of complex genetic information}

The present invention relates to a method for diagnosing disease risk based on analysis of complex genetic information networks.

To date, technology trends for diagnosing diseases are characterized by microarrays using human covalent polymorphisms (monobase polymorphisms, gene copy number variation, base insertion / deletion, etc.) of specific genes or expression information across gene populations. Researches are being conducted in order to search for genes related to specific diseases by measuring expression changes of genes and proteins using protein chips or the like, and to study the functions of the genes.

However, the existing researches focused on examining the associations between one kind of specimens and diseases, and there is a lack of understanding on the relationship and correlation between various genetic information and diseases. In addition, due to the lack of technology for analyzing the relationship between the complex genetic information and the disease, it is difficult to find a variant specific to a new disease that is not known in the past, and the accuracy of the diagnostic technique was also significantly low.

The technique of extracting biomarkers from genetic information is a method of extracting markers by statistically analyzing genetic information related to disease. However, the conventional biomarker extraction technique is performed only within the existing information range obtained from the bottom-up, and remains at the level of extracting markers based on some genetic information including genes. There is a limit to the 1: 1 relationship between diseases.

In addition, a disease diagnosis service based on a biomarker uses a method of performing a diagnosis service by calculating the degree to which specific genetic information contributes to a disease or a trait. However, conventional diagnostic service technology relies on deriving a simple relationship between a disease and a kind of genetic information, and fails to perform a complex analysis between the disease and genetic information. There is a limit in that it is not possible to reflect changes in characteristics due to high-dimensional variables such as treatment, treatment, and relapse. For this reason, the accuracy of diagnosis is low, and there are limitations in obtaining different results depending on the type of service platform.

International Publication No. 2014-052909.

In order to solve the above problems, the present invention derives disease state specific information from the relationship between the complex genetic information, and utilizes a network model-based optimization and artificial intelligence-based machine learning techniques. An attempt was made to develop a biomarker for accurate disease diagnosis and a method for predicting disease risk.

To achieve this goal, the accuracy of diagnosis is improved by analyzing the relationship between complex and diverse genetic information and genetic information to understand the relationship between diseases, and extracting the optimal combination of genetic information and introducing analysis techniques based on network models. High and economical biomarkers were also sought.

The present invention,

Extracting complex genetic information from a sample of a diseased patient and a normal person;

Comparing and analyzing the information between the complex genetic information to construct a complex genetic information library;

Deriving a disease state specific biomarker by applying an optimization method or a learning method to the complex genetic information library; And

Constructing a network model for predicting disease risk from the disease state specific biomarker and predicting risk;

Provides a disease risk prediction method, including a complex genetic information relationship analysis, including.

The present invention also provides a disease state specific biomarker derived through the disease risk prediction method through the complex genetic information relationship analysis.

The disease risk diagnosis method based on the analysis of the complex genetic information network in the blood developed in accordance with the present invention can derive a stable correlation with the disease from a small number of genetic information combinations through the introduction of a learning method, and a network model. Provide genetic information correlation based on By using the correlation between the genetic information and the disease derived from the present invention, it is expected that it is possible to secure a diagnostic technique that satisfies the level of accuracy and economic feasibility that can be commercially available in the medical industry.

In addition, the biomarkers derived from the present invention are expected to be efficiently used for the prognosis of diseases through the manufacture of medical devices including a diagnostic chip and a terminal and commercialization as a disease diagnosis service.

1 illustrates an example of a concept of a disease risk diagnosis method based on the analysis of the complex genetic information network of the present invention.
Figure 2 illustrates the concept of step-by-step genetic information based on the gene expression process.
3 illustrates an example of a concept of a method for deriving and verifying a disease state specific biomarker using a learning method.
4 shows an example of the characteristic modeling of disease state specific biomarkers.
5 shows an example of a method for CNN analysis of protein expression data.
6 shows an example of an algorithm for predicting risk of digestive cancer from mi-RNA information.
7 is a verification result using only basic CNN analysis.
8 is a result of the change in the result of learning by extracting the important mi-RNA candidate combination.
9 is a result of the simultaneous check and precise diagnosis possible confirmed in the protein.

Hereinafter, a disease risk diagnosis method based on the analysis of the complex genetic information network of the present invention will be described in detail with reference to the accompanying table or drawings.

When the drawings are described, they are provided as examples in order to ensure that features of the present invention to those skilled in the art will fully convey. Therefore, the present invention is not limited to the drawings presented and may be embodied in other forms, and the drawings may be exaggerated to clarify the spirit of the present invention.

At this time, if there is no other definition in the technical terms and scientific terms used, it has the meaning that is commonly understood by those of ordinary skill in the art to which the present invention belongs, the gist of the present invention in the following description and the accompanying drawings Descriptions of well-known functions and configurations that may be unnecessarily blurred are omitted.

In the present invention, "sample sample" or "sample" refers to genetic information obtained for analysis and is used in the same sense throughout the specification.

The present invention relates to a method for diagnosing disease risk based on analysis of complex genetic information networks in blood.

The present invention compares and analyzes general life phenomena and disease-related information based on the extracted complex genetic information to help understand genetic information functions, and further, to derive disease state specific biomarkers with high accuracy and A disease risk prediction model can be built.

In the present invention, in order to derive a disease state specific biomarker and build a disease risk prediction model, a large amount of genetic information is used in a big data processing technology, an artificial intelligence-based deep learning method, and an example of a machine learning technology ( machine learning method) may be used in combination.

Hereinafter, a method for predicting disease risk through the analysis of complex genetic information relationships will be described in detail.

The present invention

Extracting complex genetic information from a sample of a diseased patient and a normal person;

Comparing and analyzing the information between the complex genetic information to construct a complex genetic information library;

Deriving a disease state specific biomarker by applying an optimization method or a learning method to the complex genetic information library; And

Constructing a network model for predicting disease risk from the disease state specific biomarker and predicting risk;

Provides a disease risk prediction method, including a complex genetic information relationship analysis, including.

Hereinafter, each step will be described in detail.

First, the step of extracting complex genetic information from specimens of disease patients and normal people will be described in detail.

In the step of extracting the complex genetic information from the specimens of the disease patients and normal people, information related to DNA, RNA, protein, etc. for the entire genome of the sample can be obtained. The method of acquiring the information is not limited as long as the object of the present invention is not impaired. For example, the method of acquiring the information may be obtained from a genetic information database, and more specifically, provided by the National Institutes of Health (NIH). Databases are available, and more specific examples of cancer-related information can be obtained from the full genome information provided by TCGA (The Cancer Genome Atlas) on disease by type. As another example, genome sequencing may be requested from a hospital or a patient directly collected to obtain information. As another example, it is possible to secure and use a whole exome sequence set that directly plays a role in protein synthesis in a gene, but is not limited thereto.

In the present invention, the genomic sequence information of the sample may have some changes depending on the type of genetic information database, the equipment used for sequencing, the sequencing method, and the like. In addition, the genomic sequence information is not limited so long as it does not impair the object of the present invention, for example, may be based on information provided in the human genome map found from the human genome project.

In the present invention, the entire genome sequence information of the specimens of disease patients and normal people may be the basis for detecting other biomarkers in the present invention, such as cf-DNA, ct-DNA, etc., which can be obtained from such genome sequence information. The analysis is performed based on the difference of the genomic sequence information of the sample, including DNA information, mRNA expression information such as mRNA, mi-RNA, protein synthesis information, and the like. Although not limited to the above genome sequence information, chromosomal information, information related to the position of the base sequence in the chromosome, mutation information of the base sequence associated with addition, deletion or substitution of the base sequence, RNA information, protein expression information, protein 3 Information including dimensional structure and reliability can be used primarily for the detection of disease diagnostic biomarkers.

In the present invention, the analysis of the information contained in the genomic sequence information may be made by adding or subtracting information according to the type, version, environment of use of the program used.

Next, a step of constructing a complex genetic information library by comparing and analyzing the information between the complex genetic information will be described in detail.

In the step of constructing a complex genetic information library by comparing and analyzing the information between the complex genetic information, the complex relationship existing between the genetic information obtained in the step of extracting the complex genetic information from the disease patients and normal samples is analyzed. Important genetic information can be extracted and libraryed.

In the present invention, the genetic information is not limited so long as the object of the present invention is not impaired. For example, DNA information such as cf-DNA, ct-DNA, RNA expression information such as mRNA and mi-RNA, Protein synthesis information may be mentioned (FIG. 2).

Extraction of important genetic information factors for analysis is not limited as long as the object of the present invention is not impaired, but may include the following processes.

First, the classification accuracy for the case where the normal group and the disease group can be distinguished using a single genetic information factor can be extracted. The single genetic information factor is not limited to the type and number when the information can be distinguished from the normal group and the disease group, but for example, a single base polymorphism including a variation of the base sequence associated with addition, deletion or substitution of the base sequence Variations, gene copy number variation, protein amino acid sequence polymorphism variation, and the like, but are not limited thereto. For example, if the variation of the nucleotide sequence is common in the sample of the disease group, and the variation is not the same in the sample of the normal group, the corresponding genetic information is identified, and the positional information and the variation information of the nucleotide sequence are extracted. It is good to save it.

Next, by measuring the difference between the actual expression amount and the reference amount for each genetic information factor can be set on / off tag to determine whether the genetic information factor affects the selection of the disease group. As an example of the method for setting as described above, the reference value of the expression amount for each step associated with the important gene gene expression process is called Th ₁ , Th ₂ , Th ₃ , respectively, and the amount of expression of the genetic information increased or decreased by disease. In this case, an increase reference value (Th ₁ ^up , Th ₂ ^up , Th ₃ ^up ) and a decrease reference value (Th ₁ ^down , Th ₂ ^down , Th ₃ ^down ) may be defined and used, respectively. By using the variables defined as described above, the genetic information for which the expression level changes due to disease can be extracted while satisfying each expression level criteria for the obtained specimen sample. At this time, if necessary, the base sequence information of the relevant genetic information can be secured and used, and a single base polymorphic variation, gene copy number variation, etc., including a variation of the base sequence associated with addition, deletion or substitution of the above-described base sequence Extracting the variant may utilize variation information of the nucleotide sequence caused by the disease, but is not limited thereto.

In the steps illustrated in FIG. 2 using the extracted genetic information factors, a correlation between the complex genetic information can be determined by constructing a library by analyzing the expression amount change, the nucleotide sequence variation, etc. between the genetic information corresponding to the different stages. It can then be used in the derivation of biomarkers.

[Table 1] Example of Genetic Information Relationship Library

As an example of library construction through genetic information relationship analysis, mi-RNA1 is expressed below the threshold Th ₂ ^{down in} gastric stage 1 male patients and liver cancer stage 2 male patients as shown in Table 1 above, and at the same time protein5 When expressed above Th ₃ ^up ; SNP1 of mi-RNA5 was found in stage 1 men and stage 1 women but did not find a relationship with other specific genetic information; Genetic information can be recorded and libraryed.

For example, the information analyzed through the method may be stored or managed by converting it into a certain platform, that is, the same form as shown in Table 1 above.

Next, a step of deriving a disease state specific biomarker by applying an optimization method or a learning method to the complex genetic information library will be described in detail.

In the step of deriving a disease state specific biomarker by applying an optimization method or a learning method to the complex genetic information library, the complex genetic information library constructed through the method may be optimized or Analysis can be performed through learning methods to derive biomarkers specific to disease states.

The method of extracting disease state specific biomarker candidates is not limited as long as the object of the present invention is not impaired, but from the derived genetic information library, there is a relationship between the same genetic information in the sample sample and the library in the disease state to be identified. After confirming that the information is established, extracting the increase / decrease relationship of the genetic information, the variation information of the nucleotide sequence, and the number of genetic information can be extracted and selected as a candidate group for deriving a disease state specific biomarker. In selecting the candidate group, it is preferable that the disease state specific biomarker satisfies at the same time minimizing the number of genetic information to be considered while maximizing the accuracy of representing the disease state. After defining a multi-variable optimization form, a disease state specific marker may be derived through the application of a mathematical algorithm, but is not limited thereto.

The mathematical algorithm for optimizing the multivariate function can be introduced without limitation as long as it can solve the problem of the multivariate function. For example, a simulated annealing technique, a genetic algorithm, a tap search technique, a simulation Tied evolution, probabilistic evolution techniques, and the like, and preferably genetic algorithms. When the disease state-specific biomarker is extracted by the above method, it is not necessary to end the whole process, but may be interrupted in the middle of finding the optimal solution, and the best solution among the solutions obtained until then may be used.

The genetic algorithm is based on the biogenetics of the natural world, and it is a method of creating a better solution by expressing possible solutions to a problem in a form of data structure using a parallel and global search algorithm, and then gradually modifying them. . Here, the data structures representing the solutions are genes, and the process of producing better and better solutions by modifying them can be represented by evolution. In other words, the genetic algorithm is a simulated evolution search algorithm to find a solution x that optimizes some unknown function Y = f (x). The genetic algorithm is more of an approach to solving a problem than an algorithm for solving a specific problem, and can be applied to any problem that can be expressed in a format that can be used in the genetic algorithm. In general, if a problem is too complex to be calculated, it is preferable to use genetic algorithms as a way to obtain an answer close to the optimal solution even though the actual optimal solution is not obtained.

As an example of a method for deriving a disease state specific biomarker, as shown in FIG. 3, a learning sample to be analyzed and a verification sample for verifying the accuracy of the learning sample may be provided. Analysis may include only genetic information specific to the disease state, but is not limited thereto. As described in the embodiment of the present invention, learning is performed by randomly separating an analysis target library into a learning sample and a verification sample. It is possible to improve the accuracy by repeating the learning process several times.

If the size of the library is large, it is desirable to perform a process to reduce the complexity because it is difficult to calculate the classification accuracy for all subsets and the complexity increases. If the size of the library is N, all subsets have 2 ^ N cases. Accordingly, as the size of the library increases, it is difficult to calculate the classification accuracy for all subsets and the complexity increases, so to solve this problem, it is necessary to reduce the complexity by using a heuristic algorithm. For example, for the case where the subset size is N, check the likelihood of the marker and consider only the case where it is most likely, and then reduce the size of the set in stages. The number of will be reduced to N (N + 1) / 2.

Genetic information selection, which is a variable for optimizing the multivariate function, is not limited so long as the object of the present invention is not impaired. For example, genetic information may be arbitrarily selected according to the heuristic algorithm, and preferably genetic having maximum accuracy. You can select a combination of information. For example, when there is a characteristic that increases the genetic information mi-RNA1 and ct-DNA5 at the same time, the information related to the increase or decrease of each expression level of mi-RNA1 and ct-DNA5 are used as two features each for learning. It can be used, and whether the mi-RNA1 and ct-DNA5 are simultaneously increased in the sample, whether there is a feature can be used as a feature in learning.

In the present invention, the artificial intelligence-based learning method used for the learning for biomarker derivation is not limited in kind as long as the object of the present invention is not impaired. For example, a neural network Deep learning, deep learning, and the like may be used. Examples of the neural network may include a convolutional neural network (CNN) and a recurrent neural network (RNN). In one embodiment of the present invention, The CNN may be used as is, but is not limited thereto, and an appropriate learning technique may be selected and used according to the acquired data and characteristics of the biomarker.

In the present invention, preferably, a process for verifying the performance of the disease state specific biomarker derived through the above method may be further performed. To this end, the accuracy of the derived biomarker can be verified by applying the disease state specific biomarker to the sample or normal sample not used for biomarker detection, and then calculating the classification accuracy.

Next, the steps for constructing a network model for predicting disease risk from the disease state specific biomarker and predicting the risk will be described in detail.

In constructing a network model for predicting disease risk from disease state specific biomarkers, a disease derived by using a complex genetic information library and an optimization technique or learning method obtained by analyzing the relationship between the complex genetic information From state specific biomarkers, state changes such as the invention of disease, progression and recurrence can be constructed in the form of a network.

The method of constructing the network is not limited as long as the object of the present invention is not impaired, but a change in information in a disease state specific biomarker derived according to a change in a specific disease state using a genetic information library constructed through the method. It may include a method of analyzing. As an example of the analysis, as shown in Figure 4 can be modeled in the form of a mathematical function by tracking the discontinuous expression change of the genetic information ct-DNA1 or mi-RNA5. Although the form of the mathematical function is not particularly limited, it is preferable to select a regression function that can satisfactorily satisfy the data of discontinuous expression changes.

The regression method used in constructing the regression function is divided into simple regression analysis and multiple regression analysis. Simple regression analysis is used to analyze the relationship between one dependent variable and one independent variable. Can be used to identify the relationship between a variable and several independent variables. The expression change illustrated in FIG. 4 may be composed of one dependent variable and one independent variable to obtain a regression function through simple regression analysis. For example, the expression of ct-DNA1 of FIG. 4 may be an exponential function. Expression of mi-RNA5 can be modeled as a regression function in the form of a step function.

Genetic information, which is a network model for predicting disease risk composed of genetic information, to mathematically model the characteristics of disease state specific biomarkers through the above method, and to track the process of change of genetic information according to the major state of disease. Genetic-information relation network model can be established.

The form of the genetic information relationship network model is not limited as long as the object of the present invention is not impaired, but the specific disease such as a static disease network or a time course, a habit, etc. which applies only correlations between complex genetic information It may be in the form of a dynamic disease network in which information is added as a variable, and preferably in the form of a dynamic disease network. The use of this type of network model is desirable because it can track constantly changing genetic information characteristics and diagnose and predict diseases.

In the present invention, the accuracy of the genetic information relationship network model, which is a network model for predicting biomarkers and disease risk, is not limited but may be evaluated using the following indicators.

Sensitivity: A measure of how well patients are classified as having a real disease, and can be defined as TP / (TP + FN) to prevent false-based diagnosis failures, where TP is a disease The number of cases classified as FN is the number of cases classified as normal. For network models for predicting biomarkers and disease risk, preferably at least 95%, more preferably at least 99%, and most preferably at least 99.9% can increase test costs and commercialization, and important genetic information related to multiple diseases. It is good to increase the number of cases that can be confirmed by one test using.

Specificity: A measure of assessing whether a person is classified as normal, and defined as TN / (TN + FP) to prevent unnecessary follow-up due to false disease diagnosis, where TN is used to classify normal as normal. The number of cases, FP, is the number of cases in which a normal person is classified as a diseased patient. For network models for biomarker and disease risk prediction, preferably 90% or more, more preferably 95% or more, 99% or more can increase test costs and commercialization, and key genetic information related to multiple diseases It is good to increase the number of cases that can be confirmed by one test using.

In the prediction of disease risk, sensitivity or specificity may be used, or a combination of the two may be used. Of these, sensitivity is more important than specificity in predicting disease risk.

In the present invention, any disease can be applied as long as the biomarker can be derived. For example, the disease may be a disease requiring rapid diagnosis, and more specifically, the cancer may include bladder urothelial carcinoma or breast cancer. (Breast invasive carcinoma), cervical and endocervical cancers, colon cancer, colon cancer (Colon adenocarcinoma), esophageal cancer (Esophageal carcinoma), glioblastoma multiforme, head and neck squamous cell carcinoma, Kidney Chromophobe, Kidney renal clear cell carcinoma, Kidney renal papillary cell carcinoma, Acute Myeloid Leukemia, Brain Lower Grade Glioma, Liver hepatocellular carcinoma, Lung adenocarcinoma, Lung squamous cell carcinoma, Ovarian serous cystad enocarcinoma, Pancreatic adenocarcinoma, Adrenal cancer (Pheochromocytoma and Paraganglioma), Prostate adenocarcinoma, Rectum adenocarcinoma, Sarcoma, Malignant melanoma (Skin Cutaneous Melanoma), adenocinotomoma Testicular Germ Cell Tumors, Thyroid carcinoma, Thymoma and Uterine Corpus Endometrial Carcinoma may be at least one selected from the group consisting of, bladder cancer, breast cancer, colon cancer, colon cancer 1 type selected from the group consisting of cervical cancer, liver cancer, lung adenocarcinoma, anaerobic renal cell carcinoma, clear cell type renal cell carcinoma, papillary renal cell carcinoma, serous ovarian epithelial cancer, prostate cancer, lung squamous cell carcinoma and gastric cancer It may be more than one, more preferably may be one or more selected from the group consisting of breast cancer, colon cancer and stomach cancer, but is not limited thereto.

The present invention also provides a disease state specific biomarker derived through the disease risk prediction method through the complex genetic information relationship analysis.

The biomarker derived from the present invention is expected to be effectively used for the prognosis of diseases through the manufacture of medical devices including diagnostic chips and terminals and commercialization as disease diagnosis services.

Hereinafter, the content of the present invention will be described in more detail with reference to Examples. The examples are only for more specifically describing the present invention, and the scope of the present invention is not limited thereto.

[Experimental Material]

1. The following mi-RNA related data were obtained and used.

1) GSE54397

It was used in the research database of Kim Na-young, Seoul National University.

The database received 3523 mi-RNA data from microarray data of normal and cancer tissue samples of 16 gastric cancer patients.

2) GSE61741

It was provided by the database of Saarland University.

The database contains a total of 848 mi-RNAs, including 136 samples of 13, 29, and 94 gastric cancer patients, colorectal cancer patients, and normal persons, respectively, from microarray data of a total of 1049 blood samples including cancer patients and normal persons. Data was used and used.

3) TCGA NGS Data

It was provided from the TCGA database.

45,446 samples of 491 normal and cancer tissue samples of gastric cancer patients were downloaded from the database, and miRNA and NGS read information was obtained.

A total of 211 mi-RNA data were used.

2. The data other than the following mi-RNA was obtained and used.

1) TCGA protein expression array data

Protein expression database for breast cancer, thyroid cancer, liver cancer, kidney cancer, lung cancer patients and normal patients.

Breast cancer disease, breast cancer normal, thyroid cancer, liver cancer, kidney cancer 1 (Kidney Renal Clear Cell Carcinoma), kidney cancer 2 (Kidney Renal Papillary Cell Carcinoma), kidney cancer 3 (Kidney Chromophobe), lung cancer 1 (Lung Adenocarcinoma), Lung Squamous Cell Carcinoma data were obtained,

The number of samples is 1078, 45, 426, 183, 478, 215, 63, 365 and 327 respectively.

There is expression data for about 200 proteins in each sample, and only 146 protein data which are common in all samples were extracted and used.

Table 2 Example mi-RNA and Protein Expression Data

Example 1 Application of Learning Techniques to Protein Data

For the protein data, the relationship between the input data was derived using the CNN method of FIG. 5, passed through the fully connected layer, and finally classified using softmax.

Example 2 Accuracy Prediction

Learning was performed by applying the same CNN network method as in Example 1 above.

1.GSE54397

Of the 32 samples, 22 were trained and the remaining 10 were validated.

2. GSE61741

106 of 136 samples were trained and 30 were validated.

3.TCGA NGS data

391 of 491 samples were trained and the remaining 100 were verified.

The verification results are shown in FIGS. 7 and 8.

According to the results of FIG. 7, as the learning progresses, the GSE54397 model (tissue and microarray data) shows 100% classification accuracy, GSE61741 (blood and microarray data) is about 96.67% and TCGA NGS data is about 99% Converging to the classification accuracy, it was confirmed that all showed very high accuracy of 95% or more.

According to Figure 8, the sensitivity (sensitivity) as the progress of the extraction process of important mi-RNA converges to 1 for all 848 mi-RNA and 30 optimal mi-RNA extraction (BEST) cases as learning progresses In the case of specificity, fluctuation occurred near 1 while learning about 848 mi-RNAs, and converged to more than 0.95 for 30 optimal mi-RNA extractions (BEST). .

Example 3 Derivation of Biomarkers Using Clinical Data

Data on three diseases of breast, stomach and colorectal cancers were all obtained from a database of the TCGA (The Cancer Genome Atlas) project, which has been in progress since 2006 in NIH, USA. The detailed database names used to obtain each disease data are as follows.

Breast cancer: TCGA-BRCA, Gastric cancer: TCGA-STAD, Colorectal cancer: TCGA-COAD.

Among the mi-RNA genetic information data, 30 types of biomarkers were derived for each cancer type through learning through the CNN network method in the same manner as in the above example.

The results are shown in Table 3 below.

Optimal mi-RNA Biomarkers (BEST) by Cancer Type Cancer type mi-RNA biomarkers Breast cancer 'hsa-mir-30d', 'hsa-mir-145', 'hsa-mir-425', 'hsa-mir-203a', 'hsa-mir-452', 'hsa-mir-378a', 'hsa -mir-455 ',' hsa-mir-100 ',' hsa-mir-199b ',' hsa-mir-205 ',' hsa-mir-542 ',' hsa-mir-532 ',' hsa-mir -625 ',' hsa-mir-200c ',' hsa-mir-183 ',' hsa-mir-22 ',' hsa-mir-451a ',' hsa-mir-30a ',' hsa-mir-30e ',' hsa-mir-148a ',' hsa-mir-143 ',' hsa-mir-375 ',' hsa-mir-584 ',' hsa-mir-379 ',' hsa-mir-10a ', 'hsa-mir-182', 'hsa-mir-21', 'hsa-mir-486-1', 'hsa-mir-486-2', 'hsa-mir-10b' Colorectal cancer 'hsa-mir-6086', 'hsa-mir-3118-1', 'hsa-mir-1321', 'hsa-mir-548f-5', hsa-let-7c ',' hsa-mir-4752 ' , 'hsa-mir-183', 'hsa-mir-29a', 'hsa-mir-30e', 'hsa-mir-486-1', 'hsa-mir-194-1', 'hsa-mir- 194-2 ',' hsa-mir-30a ',' hsa-mir-28 ',' hsa-mir-25 ',' hsa-mir-486-2 ',' hsa-mir-182 ',' hsa- mir-30d ',' hsa-mir-203a ',' hsa-mir-10b ',' hsa-mir-148a ',' hsa-mir-145 ',' hsa-mir-378a ',' hsa-mir- 143 ',' hsa-mir-22 ',' hsa-mir-10a ',' hsa-mir-200c ',' hsa-mir-21 ',' hsa-mir-192 ',' hsa-mir-375 ' Stomach cancer 'hsa-mir-500b', 'hsa-mir-496', 'hsa-mir-2392', 'hsa-mir-5739', 'hsa-mir-4540', 'hsa-mir-6749', 'hsa -mir-1915 "," hsa-mir-202 "," hsa-mir-2467 "," hsa-mir-27b "," hsa-mir-583 "," hsa-mir-374c "," hsa-mir -219b "," hsa-mir-299 "," hsa-mir-142 "," hsa-mir-30d "," hsa-mir-3074 "," hsa-mir-147b "," hsa-mir-5009 ',' hsa-mir-624 ',' hsa-mir-181d ',' hsa-mir-489 ',' hsa-mir-581 ',' hsa-mir-29b-2 ',' hsa-mir-541 ',' hsa-mir-485 ',' hsa-mir-4519 ',' hsa-mir-20b ',' hsa-mir-486-1 ',' hsa-mir-527 '

From the above, there are eleven types of mi-RNA biomarkers common to breast cancer, colorectal cancer and gastric cancer, which can be interpreted as biomarkers having characteristics common to three cancers.

Biomarkers Common to Three Cancers Cancer type Common mi-RNA Biomarkers Breast, colon and stomach cancer 'hsa-mir-143', 'hsa-mir-148a', 'hsa-mir-182', 'hsa-mir-203a', 'hsa-mir-21', 'hsa-mir-22', 'hsa -mir-30a "," hsa-mir-30e "," hsa-mir-375 "," hsa-mir-486-1 "," hsa-mir-486-2 "

Among biomarkers derived from the analytical method of the present invention from the data of Example 3, from the biomarkers common to three cancers, it is known that the hsa-mir-486 series is associated with other cancers, and hsa The -mir-375 family is known to be involved in circulating the body, and the hsa-mir-30 family is known to be involved in cancer suppression.

That is, from the method of the present invention, it was confirmed that the biomarker, which was previously known to be a factor associated with cancer, was accurately extracted as a major factor, and also confirmed to be a correct conclusion about the known results.

In addition, it can be seen that in addition to the biomarkers described above, individual cancer specific biomarkers are novel biomarkers for the diagnosis of individual cancers.

Example 4 Prediction of the Accuracy of Biomarkers Derived Using Clinical Data

The result of performing the disease risk prediction operation using the biomarker derived in the same manner as in Examples 1 to 2 is as follows.

The measurement results of each sensitivity and specificity were performed to be a general result for the algorithm itself rather than a specific learning set through the 100-fold cross validation technique.

The risk prediction algorithm consists of a convolutional neural network, consisting of seven layers of convolutional layer and four layers of fully connected layer.

The convolutional layers are all composed of 1-dimensional filters. The first layer is 20 by 1, the second layer is 10 by 1, and the third and subsequent layers are 3 by 1.

Padding used the 'Valid' technique.

The fully connected layer is composed of 1024, 512, 256, and 128 nodes. Finally, the readout layer and softmax activation are used to classify disease probabilities.

The results are shown in Table 5.

Disease risk prediction result by cancer type Cancer type Sensitivity Specificity Breast cancer 98.0% 95.5% Colorectal cancer 99.3% 96.0% Stomach cancer 99.0% 96.2%

From the above results, the disease risk prediction method through the composite genetic information relationship analysis according to the present invention not only provides a biomarker with a high accuracy of 95% or more, but also exhibited a sensitivity and specificity of 95% or more. From this, the present invention can increase the possibility of lowering the cost and commercialization of the test, and confirms that the number of cases that can be confirmed by a single test is increased by utilizing key genetic information related to multiple diseases, thereby satisfying the commercially available level of accuracy and economic efficiency. The present invention has been completed by showing that it can be used as a diagnostic technique.

Claims (14)

Extracting complex genetic information from a sample of a diseased patient and a normal person;
Comparing and analyzing the information between the complex genetic information to construct a complex genetic information library;
Deriving a disease state specific biomarker by applying an optimization method or a learning method to the complex genetic information library; And
Constructing a network model for predicting disease risk from the disease state specific biomarker and predicting risk;
Comprising a disease risk prediction method through a composite genetic information relationship analysis. The method of claim 1,
The complex genetic information is any one or two or more expression or synthetic information selected from the group consisting of DNA, RNA and protein, characterized in that the disease risk prediction method through a composite genetic information relationship analysis. The method of claim 1,
The complex genetic information library is derived by building a statistical analysis or optimization method (optimization method), characterized in that the disease risk prediction method through a composite genetic information relationship analysis. The method of claim 3, wherein
In constructing the complex genetic information library, setting an on / off tag for determining whether the genetic information factor affects the selection of a disease group by measuring the difference between the actual expression amount and the reference amount for each genetic information factor. Characterized in that the disease risk prediction method through a composite genetic information relationship analysis. The method of claim 4, wherein
The on / off tag setting is
a) important as genetic gene expression process, the standard value of the expression level of each step, each Th _1, Th _2, Th ₃ related to, and increase, respectively to increase or decrease the amount of genetic expression by the disease threshold value (Th ₁ ^up, Defining variables with Th ₂ ^up , Th ₃ ^up ) and reduction criteria (Th ₁ ^down , Th ₂ ^down , Th ₃ ^down ); And
b) using the variable, extracting genetic information for which the expression level is changed by the disease while satisfying each expression level criteria for the sample sample;
A disease risk prediction method through a complex genetic information relationship analysis, comprising a.
The method of claim 5,
Extracting a variant including a single nucleotide polymorphic variation or a gene copy number variation including addition, deletion or substitution of a nucleotide sequence by obtaining nucleotide sequence information of the genetic information when extracting the genetic information;
The disease risk prediction method further comprising a complex genetic information relationship analysis, characterized in that it further comprises. The method of claim 1,
Analyzing the relationship between the complex genetic information existing in the complex genetic information library and the disease through an optimization technique or a learning technique to derive a biomarker that can be used for disease analysis, through the complex genetic information relationship analysis How to predict disease risk. The method of claim 1,
Method of predicting disease risk through the analysis of the relationship between complex genetic information, characterized in that to build a static disease network (statistic disease network) model based on the disease state specific biomarker. The method of claim 1,
A method for predicting disease risk through a complex genetic information relationship analysis, characterized in that a dynamic disease network model is constructed in a network model for predicting disease risk and predicting a risk. The method of claim 7, wherein
The optimization technique is selected from the group consisting of simulated annealing technique, genetic algorithm, tap search technique, simulated evolution, probabilistic evolution technique, disease risk prediction method through a composite genetic information relationship analysis. The method of claim 7, wherein
The learning technique is selected from the group consisting of a neural network (neural network) and deep learning (deep learning), disease risk prediction method through a composite genetic information relationship analysis. The method of claim 11,
The neural network is selected from the group consisting of a convolutional neural network (CNN) and a recurrent neural network (RNN), disease risk prediction method through a composite genetic information relationship analysis. The method of claim 1,
The accuracy of the network model for predicting disease risk is characterized in that the sensitivity (sensitivity) of more than 95%, the specificity (specificity) of more than 90%, disease risk prediction method through a composite genetic information relationship analysis. A disease state specific biomarker derived through a method for predicting disease risk by analyzing the relationship between complex genetic information of any one of claims 1 to 13.

Images