JP2021101629A5 - - Google Patents

Description

This patent application relates to systems and methods for genomic and genetic analysis of human nucleic acid samples.

<Next generation sequencing (hereinafter also referred to as “NGS”)>
Next-generation sequencing, also known as high -throughput sequencing, is a common method of high -throughput, parallel sequencing of nucleic acid fragments familiar to those skilled in the art. Next-generation sequencing equipment and systems are commercially available from a variety of suppliers (see www.illumina.com).

Next-generation sequencing is a broad term used to describe many different modern sequencing technologies, including the following sequencing technologies.
・ Illumina (Solexa) Sequencing (registered trademark)
・ Ion Torrent: Proton/PGM Sequencing (registered trademark)
・ SOLiD Sequencing (registered trademark)

NGS technology produces high quality DNA sequences (“reads”). Reads generated by NGS technology are shorter than those generated by Sanger sequencing technology by capillary electrophoresis (650-1000 bp) developed in 1977 by Frederick Sanger and his colleagues. Sanger sequencing technology has been the most widely used method for about 30 years. Reads generated by Sanger sequencing technology have low throughput and high cost. On the other hand, the leads made by the NGS method are all much shorter and the cost is not very high. However, the total number of base pairs sequenced in a single run of NGS is orders of magnitude larger. These two factors pose new informatics challenges, including the ability to process millions or even billions of such short NGS reads. Sequenced reads are usually processed in one of two ways. That is, either these reads are mapped to the correct position in an existing backbone /reference sequence to create a sequence that is similar, but not necessarily identical, to the backbone (referred to as "read mapping"), or these reads are spliced together. Either a new array (called "de novo assembly").

A major advantage of read mapping back to the reference genome, as opposed to de novo assembly, is that it greatly simplifies the process of genome prediction. Assembly requires finding all genome sequences, which creates a lot of ambiguity , whereas resequencing based on a reference sequence only needs to find differences between the reference sequence and the sample. Given the complexity and time required, de novo assembly is several orders of magnitude slower and requires more memory than mapping assembly.

Read mapping is the first and most fundamental step in the NGS analysis pipeline, where a newly sequenced human genome (or exome or targeted A step aimed at finding differences in newly sequenced fragments of the human genome, such as fragments of genes.

In addition, read mapping is used to sequence millions or billions of short NGS reads to determine coverage (reads at a particular location/locus), an important quality parameter for NGS experiments and the conclusions drawn from those experiments. number).

<Human reference genome (hereinafter also referred to as “HRG”)>
The Human Genome Project, a United States federal government effort in collaboration with the private company Celera Genomics , completed a draft of the entire human genome in February 2001. This draft has since been revised several times (see Lander et al. 2001, Venter et al. 2001, Church et al. 2011). Over the years, genome assembly has made steady progress, with new versions (“builds”) being released one after another, and the latest Genome Reference Consortium (GRC) human genome assembly, GRCh38 (see Schneider et al. 2017), has been largely inaccurate. It is the best assembled mammalian genome that exists. GRCh38 has only 875 remaining assembly gaps, with fewer than 160 million unspecified 'N' nucleotides (GRCh38 onwards, p8). On the other hand, the first version had about 150,000 gaps (see Editorial (October 2010). "E pluribus unum". Nature Methods. 7 (5): 331. doi:10.1038/nmeth0510-331).

HRG is also the single most important resource in human genetics and genomics today. The HRG acts like a cosmic coordinate system and is therefore a space in which annotations (genes, promoters , etc.) and genetic variation are accounted for (Harrow et al. 2012; ENCODE, 2012; 1000 Genomes Project Consortium , 2012). HRG is also the reference for the read alignment step in the next-generation sequencing analysis pipeline, and downstream of this mapping, HRG is used for functional assays and variant calling pipelines (Li H & Durbin 2009; DePristo et al., 2011).

The first type of HRG consisted of multiple DNA sequences from a small group of 13 anonymous DNA donors of predominantly European origin (see Snyder et al) volunteers in Buffalo, NY. Donors were solicited in the Buffalo News on Sunday, March 23, 1997. The first 10 men and 10 women were invited to meet with the project's genetic counselors, donate blood, and DNA was extracted from the donated blood. Due to the way these DNA samples were processed, approximately 80% of the reference genomes are from 8 individuals. A single male DNA sample, designated RP11, constitutes 66% of the total reference genome.

A resource of sequence data from new genome mapping techniques and single haplotypes from new donors to identify and determine complex regions containing large-scale duplications and structural variations of larger population problems, e.g. has been put into the latest build. At the time of filing of this application, GRCh38 contains sequences from 50 different humans (http://www.bio-itworld.com/2013/4/22/church-on-reference-genomes-past-present- future.html).

Limitations of HRG 1. HRG is linear.
All human DNA is carried in physically discrete units called chromosomes. Humans are diploid organisms containing two sets of genetic information, one set inherited from the mother and one set inherited from the father. As a result, each somatic cell has 22 pairs of chromosomes called autosomes (one chromosome of each pair is from one parent) and two sex chromosomes (males have an X and a Y chromosome; females have have two X chromosomes). Each chromosome contains a single, very long, linear DNA molecule. The DNA molecule in the smallest human chromosome consists of about 50 million nucleotide pairs, and the largest human chromosome contains about 250 million nucleotide pairs.

The diploid human genome thus consists of 46 single DNA molecules of 24 different types. Since human chromosomes exist in nearly identical pairs, it is necessary to sequence 3 billion nucleotide pairs (the haploid genome) to obtain complete information about a typical human genome . The human genome is thus said to contain 3 billion nucleotide pairs, whereas most human cells contain 6 billion nucleotide pairs. The human haploid genome consists of 22 autosomes plus Y and X chromosomes.

Each of all chromosomes corresponds to a single DNA molecule, ie millions of contiguous nucleotide bases. One might think that these DNA molecules are linear and that each chromosome corresponds to a single contiguous/linear sequence of nucleic acids. Unfortunately, this is incorrect for two reasons.
(1) Due to the nature of genomic DNA and the limitations of sequencing, some parts of the genome remain unsequenced.
(2) Some regions of the genome vary so much between individuals that they cannot be represented as a single continuous sequence.
However, HRG consists of regular bases (A, C, T and G) and is a linear sequence of 24 with gaps denoted as consecutive multiple 'N's that mark the positions of the gaps in the assembly. is represented as

The main goal of the Human Genome Project was to create a single representative sequence, with some regions of uncertainty, ie a single scaffold for each physical chromosome. The Human Genome Project also included a small amount of alternative scaffolds representing allelic variations (multiple different types of DNA bases present at a SNP locus are called alleles). However, these alternate scaffolds had no formal relationship with the primary scaffold. When it was found that a single reference sequence did not adequately represent regions of the human genome with some highly diverse forms, a formal model for introducing representative alternatives for hypermutable regions was GRCh37 (Church et al. 2011) was added as a start . Sequences in the form of "alternate locus scaffolds " of a thousand bases to several million bases are associated with "primary" (haploid) assembly along the primary scaffold. attached and explained. In this assembly (GRCh38, p9) at the time of filing of this application, these sequences occupy 178 regions and a total of 261 linear sequences (see Paten et al. 2017).

Another complicating factor is that HRGs are inferred from pools of DNA from multiple anonymous individuals in international genome sequencing projects. Thus, the resulting HRG is in fact a randomly mixed conglomerate, a collection of multiple different haploid DNA sequences that cannot be represented as a single linear sequence. There is

2. HRG is definitely not disease-free Chen and Butte (2011) identified 3556 disease-predisposing mutations in HRG, including 15 rare mutations (major allele frequency <1%). Using a curated database of high-quality, quantitative human disease SNP data, the authors found that healthy people were at risk of developing 104 diseases for the reference genome. We investigated the possibility of increasing The results showed an increased risk of type 1 diabetes, hypertension and other diseases. This is evidence that HRG is definitely not representative of the average person and is not immune to disease. Although HRG has significantly accelerated the analysis of sequencing efforts in the human genome, focusing on variants that differ from the reference genome is likely to miss many disease-causing variants, including rare variants (see Chen & Butte 2011). ).

3. Reference allele bias over European ancestry A major challenge in using HRG assembly in prior art NGS analysis pipelines is the relatively small number of anonymous individuals whose HRGs have a bias over European ancestry. The fact that it is extracted from a donor's DNA sample and thus represents a small sample from a large group of human genetic diversity.

Although the reference genome as a coordinate system representing the majority of genomes is relatively valid and relatively widespread, the use of HRGs to study all other human genomes accordingly is extremely unlikely. It is of great concern that it eliminates much human variability and introduces bias into the broad canonical allele (see Petrovski et al. 2016, Paten et al. 2017). Reference allele bias is the tendency for alleles present in the reference genome to be reported with emphasis and other alleles whose underlying DNA does not match the reference allele to be suppressed (Degner et al. 2009). , Brandt et al. 2015).

This bias occurs mainly during the read mapping and alignment steps when re-sequencing. For proper mapping, multiple reads must be represented in the reference genome and must be from genomic sequences sufficiently similar to the reference sequence to be identified as the same genomic element. If these conditions are not met, errors in mapping regularly diverge from the true sequence (see Paten et al. 2017). Depending on the ancestry history of the reference genome in which each locus is biased, reference allele biases may be biased in people of certain genetic subsets relative to others, and in certain regions of the genome. It may have more impact than other regions (see Petrovski et al. 2016, Paten et al. 2017]). Highly polymorphic regions, such as HLA genes, are particularly susceptible to reference allele bias, especially when a single reference genome is used as an index to locate NGS reads (Nielsen et al. 2011). reference). In such cases, many of the true mutations cannot be identified because they are present in haplotypes different from the genome used as index. Thus, reads made in such regions are misaligned and lost (see Brandt et al. 2015).

As mentioned above, canonical allele bias is a known problem when resequencing the human genome to find mutations using HRG, and modification of the canonical allele improves the accuracy and interpretation of mutation calling. (see Fakhro et al. 2016). One way to alleviate this problem is to modify the mutation rate early in the genome interpretation process by modifying the reference genome, so that mutations found in the genome become minor alleles in the population. (see Dewey et al., 2011). Such modifications to the reference genome reduce the number of false positives and simplify the analysis workflow by reducing the number of mutations that must be interpreted (see Fakhro et al. 2016).

The Future: A Graph-Based Reference Structure/Genome Graph As a common reference for human genetics and genomics research, since the single haploid reference genome represents only a small fraction of human diversity. There is a growing perception that it is inadequate. There are mutations and annotations that cannot be easily explained with respect to the reference genome (see Horton et al. 2008, Pei et al. 2012). Furthermore, targeting a single haploid reference genome for read mapping and interpretation introduces reference allele bias as described above. To alleviate such problems, it is believed that current reference genome assemblies, such as the human genome assembly at the time of filing (GRCh38, see p. 9), will be of " alternate locus " sequences ("alts"), i.e. highly variable. Additional multiple sequence representations of the regions of the human genome that are represented, including sequences attached at their ends to positions within the "primary" (haploid) canonical assembly. Such a structure contains partially overlapping sequence pathways and can be thought of as a form of mathematical graph, the genome graph (see Novak et al. 2017).

Graphs have long occupied a niche in the field of biological sequence analysis, and in biological sequence analysis graphs have been used to conveniently represent "collections of possible sequences". Normally all arrays are themselves indirectly encoded as paths in the graph. Because sequences are encoded in this way, graphs are well suited to represent a reference population, which is a collection of naturally related sequences (see Paten et al. 2017). The graph contains not only the approximate sequence of samples, but also the specific mutations of many samples.

Genome graphs are expected to improve read mapping, variant identification processing and haplotyping. Graph-based criteria are expected to replace one-dimensional (linear) criteria in humans and other applications where sequenced individual cohorts are possible (Novak et al. 2017). reference). Various projects are underway to create and use such genome graphs. Genome graphs can now be made from libraries of multiple common mutations, and although still experimental, there are tools that show the potential of graph-based methods.

Despite its theoretical advantages, research into mutation identification using genome graphs is still in its infancy. Many issues must be addressed. How should replication and repetition be represented? How should we classify multiple short mutations that may or may not be concordant? How can we classify diversity more comprehensively using graphs? The answers to these questions depend on future research.

For genome graphs to be useful in practice, the expected low standard bias must lead to an objective improvement of the variant identification process over established methods. Therefore, developing mutation identification processing algorithms for genome graphs has become an important research frontier.

Qatar Genome (QTRG)
Qatar is a peninsula on the Persian Gulf coast, with a total population of about 300,000 Qatari citizens. Qatari consanguineous marriage rates are among the highest in the world and are still rising. The rate of intraracial marriage in Qatar is nearly 100%. These factors, along with large families, explain the high incidence of congenital genetic diseases that weigh heavily on Qatar's budget. Factors such as these have prompted the Qatari government to find ways to protect its citizens from the threat of genetic diseases (see Zayed 2016).

Government officials decided to launch the Qatar Genome Project (QGP, see http://www.gulf-times.com/story/374345/Qatarlaunches-genome-project) in 2013. The aim of this project is to map the disease-causing/rare mutations and define the Qatari human genome as a therapeutic strategy for individuals, thereby freeing Qatari people from high rates of congenital genetic diseases. The goal is to sequence the genome of each Qatari citizen to protect. The ultimate goal of this project is to apply the information obtained to clinical practice, making this technique part of the routine work of the Qatari healthcare system (see Zayed 2016). In order to realize clinical application of QGP, several important tasks must be achieved, including making the mutation identification process highly sensitive and accurate (see Koboldt 2010).

Indigenous Qatari Arabs by combining allele frequency data from genomic sequencing of all 1161 Qatari people, who represent 0.4% of the population, to facilitate accurate treatment in the region of the Middle East and North Africa. A population-specific genome was assembled (QTRG) dedicated to disease studies in populations of the local population of . A total of 20.9 million single nucleotide polymorphisms and 3.1 million InDel (insertions and deletions) were found in Qatar. This includes an average of 1.79% novel mutations per genome (see Fakhro et al. 2016).

1000 Genomes Project (1kG)
In 2008, the genomes of at least 1000 people worldwide were sequenced, resulting in the 1000 Genomes Project (hence the name 1000 Genomes Project). The current Phase III analysis of this project includes 26 community populations and 2,504 individuals from five so-called super-regional populations, each delimited by stitching four to seven community populations (1000 Genomes Project Consortium et al. 2015). This smaller-scale haplotype resource facilitates understanding of genetic diversity at the genomic and geographic level (see Baye, 2011).

Recent advances in NGS technology have enabled DNA and RNA sequencing to be performed quickly and inexpensively, thereby revolutionizing the discipline of genomics and molecular biology. Genome sequencing projects in healthy and diseased populations have identified functional or disease-linked genomic variants. These genomic variations provide clues about therapeutic targets or genomic markers for new clinical applications.

Mutation identification processes are generally based on the alignment of multiple raw read sequences against a reference genome (read mapping). This alignment-based approach has many limitations. Such limitations include imperfect genome assembly (see Meyer, LR et al., 2013) and structural alterations within normal individual human genomes (Sudmant et al., 2015). ), sequencing errors in the reads, and multiple single nucleotide polymorphisms (SNPs) interfering with read mapping.

Currently, read mapping to linear HRG is standard practice at the time of filing of this application, and is standard practice in clinical NGS analysis pipelines and individual human resequencing . This is because HRG is relatively effective and widely used as a coordinate system for most genomes. In addition (unlike in situations where genomic interference using genome graphs occurs), a number of methods have been published for identifying mutations using linear reference genomes (Nielsen et al. 2011).

However, as mentioned above, one major problem is the bias in HRG to ignore previous information about genetic diversity within a species. Currently, this problem is solved by modifying the reference genome so that the identified mutations compared to the modified reference genome are minor alleles in the population.

Successful clinical genomic studies using NGS technology require highly accurate and consistent identification of individual human genome mutations. A prerequisite for such purposes is the correct implementation of the read mapping ( alignment ) and subsequent mutation identification procedures.

One object of the present invention is to find new biomarkers, specifically single nucleotide variations (SNV), insertions and deletions (InDel), copy number variations (CNV) and, for example, chromosomal translocations, Finding mutations in genes used for next-generation sequencing in human genome research, such as structural variations such as inversions, duplications, large insertions and deletions.

Another objective is to increase the accuracy and reliability of biomarkers used for cancer therapy where current NGS-based biomarker technology, such as biomarker technology, is used to analyze cancer cells and damaged DNA of cancer cells. It is to raise the degree.

A method according to a first aspect of the present invention is a method for genomic and/or genetic analysis of a human nucleic acid sample, said method comprising the steps of:
a) Providing a set of multiple human reference genomes.
b) Testing human nucleic acid samples to determine gender and/or ancestry.
c) selecting one or more population-specific human reference genomes (PHREGs) from said set of plurality of human reference genomes based on the results of said gender and/or ancestry test of step b);
d) aligning said human nucleic acid sample against the PHREG selected in step c);

In the following, "population-specific human reference genomes" (PHREGs) are understood as ancestry-specific reference genomes and sex -specific reference genomes. PHREGs sufficiently reduce the bias of the criterion to improve the accuracy of the alignment and, if followed by mutation identification, also improve the accuracy of the mutation identification process. It is an advantage of the present invention that it not only improves the accuracy of alignment , but also the speed of calculation, the number of correctly aligned reads and the number of calculation steps for positioning. The advantage of using PHREGs for genomic and/or genetic analysis of human nucleic acid samples is increased read coverage (depth) , and this advantage can be appreciated by the increased sensitivity of the mutation identification process.

In the context of the present invention, the term "human nucleic acid sample" generally means any nucleic acid sample that has been isolated from a human sample. This human nucleic acid sample may specifically include NGS reads as defined in detail below.

A human nucleic acid sample is typically a sample resulting from any biochemical, molecular and cell biological technique suitable for producing a human nucleic acid sample. Such procedures include punctures, biopsies, cell-free DNA kits, and the like. Human nucleic acid samples may be extracted from any suitable source, including bodily fluids, mucous membranes, tissues, tissue extracts or cells, or combinations thereof. A human nucleic acid sample may be a comparative reference sample extracted from any suitable source. Human nucleic acid samples may include, for example, blood samples, blood plasma samples, urine samples, tumor samples, and may also be subjected to the tissue processing technique FFPE (formalin-fixed paraffin-treated tissue or formaldehyde-fixed paraffin-treated tissue). ) may include undesirable artifacts resulting from solidification by ).

The human nucleic acid sample may specifically be DNA, RNA and/or size-fractionated total DNA or RNA. Providing DNA from a subject sample may involve one or more biochemical purification steps. Such biochemical purification steps may include, for example, centrifugation, lysis and/or fractionation steps, i.e. cell lysis by mechanical or chemical disruption steps, which mechanical or chemical disruption steps Examples include, but are not limited to, multiple freeze and/or thaw cycles, (multiple) salt treatments, phenol-chloroform extraction, sodium dodecyl sulfate (SDS) treatment and proteinase K digestion. Optionally, preparing the DNA from the sample of interest further removes large RNA such as ribosomal RNA abundant in the presence of polyethylene or salt, or salt, preferably in a potassium chloride solution. This may include precipitating out any interfering sodium dodecyl sulfate (SDS) present. Methods for purifying total DNA or RNA from cells and/or tissues are well known to those of skill in the art, such as guanidine thiocyanate-acidic phenol chloroform extract (eg, TRizol®, Invitrogen, USA). ), including standard techniques such as using However, it is equally preferred to provide the DNA of interest without the biochemical precipitation and/or purification steps described herein.

In the context of this invention, the term "nucleic acid" refers to any oligonucleotide molecule composed of either single- or double-stranded deoxyribonucleotides or ribonucleotides, or both, including genomic DNA, nuclear DNA, somatic includes (somatic) DNA, germline ( germline ) DNA and/or artificially designed and/or manufactured DNA, where artificially designed and/or manufactured DNA is derived from messenger RNA profiles It includes, but is not limited to, in vitro produced DNA, preferably in the form of cDNA. The term "nucleic acid" generally refers to single- or double-stranded oligonucleotide molecules of identical or similar length, ie, of identical or similar number of nucleotides.

Human nucleic acid samples are genomic sequences useful for assessing, analyzing, aligning , indexing , and/or profiling specific mutations in genes at the genomic, transcriptional, or post-transcriptional level. It can be anything that says Thus, human nucleic acids according to the present invention may include any coding region, non-coding region, exon, intron, chromosomal and/or intrachromosomal region, promoter region, enhancer region, small and /or long regulatory RNA ( small /long Regulatory RNA) , active transcribed and/or untranscribed regions, transposons, hotspot mutated regions, frameshift mutated regions and the like.

A "set of human reference genomes" has at least two human reference genomes, preferably a plurality of human reference genomes. The test for sex and/or ancestry in step b) is to select one or more best matching human reference genomes from the set of human reference genomes described above in step c). In a preferred case, the test for sex and/or ancestry of step b) automatically classifies sex and/or ancestry, and a single PHREG for use in a subsequent alignment step d) of said plurality of human reference genomes. Allows you to choose from the set consisting of However, it is also possible to select one or more extra PHREGs for later analysis.

Testing for sex and/or ancestry in step b) is performed on a sex- and/or-specific subset of sequence variants for sex and/or ancestry extracted from a database of curated data. based is preferred. Such sequence variations are preferably single nucleotide polymorphisms (SNPs) and/or single nucleotide variations (SNVs). Such subsets of sequence variations used in tests for sex and/or ancestry are also referred to as population-dependent patterns of human ancestry and sex (PHASP). Preferably, the database of curated data described above contains all known sequence variations for all populations. The PHASP dataset is an excerpt from the database of curated data described above. This PHASP dataset is a much smaller dataset than the PHREG dataset and is the most discriminating dataset when classifying. The approach used to create PHASP is a computational method based on machine learning that involves reducing features that are genotypes. Such machine learning may be tested against standard classification results.

A test to determine gender and/or ancestry includes a preliminary alignment step that detects distinct sequence variation patterns in said human nucleic acid sample. In this step, the human nucleic acid sample is aligned to a single human reference genome, eg GRCh37 or GRCh38. The single human reference genome used here in the test of step b) is not ancestry-specific or gender -specific. The patient's ancestry and gender are determined by comparing the sample's sequence variation pattern to the PHASP dataset.

According to one embodiment, the testing in step b) may include testing for gender . According to another embodiment, the testing of step b) may include testing for ancestry. According to yet another embodiment, the testing of step b) may include testing for gender and testing for ancestry.

In one exemplary embodiment, the set of human reference genomes has both male and female reference genomes. If the gender test in step b) determines that said human nucleic acid sample is a male reference genome or a female reference genome, then in step c) the corresponding male or female reference genome or both are tested in subsequent step c). is selected as the PHREG used in the positioning of the .

Since multiple sex chromosomes contain homologous sequences, by using a sex -corrected reference genome (having X and Y chromosomes in males and no Y chromosome in females) Prevents lead misalignment . Therefore, using a gender -specific reference genome reduces later false positive and false negative variant identification.

In another exemplary embodiment, the set of human reference genomes comprises multiple ancestral reference genomes. The ancestral test of step b) determines the best matching reference genome or genomes from a number of ancestral specific reference genomes. Then in step c) the closest reference genome or genomes are chosen as the PHREG or PHREGs to be used in subsequent step d). The ancestral test of step b) determines the best matching reference genome or genomes from a number of ancestral specific reference genomes. Then in step c) the closest reference genome or genomes are chosen as the PHREG or PHREGs to be used in subsequent step d).

Choosing the wrong ancestor can lead to many false-positive mutation identifications and many false-negative mutation identifications. Using an ancestral specific reference genome can effectively increase the number of correctly aligned reads, reduce the identification of false positive mutations and reduce the identification of many false negative mutations.

Similarly, if the set of human reference genomes has an ancestry-specific male reference genome and an ancestry-specific female reference genome, the combined test for sex and test for ancestry is error-free.

The term "testing" in step b) should be understood as comprising testing at least one gene or genome of a human nucleic acid sample. Testing for genes and/or genomes is more reliable than any information derived from "self-reports." Individual-reported and investigator-specified ancestry is usually based on subjective interpretation of a complex combination of both genetic and non-genetic information, including behavior, culture, social norms, skin color and other influences. ing. Research participants or patients rarely report their ethnicity correctly. There are many reasons for misreporting one's ethnicity. Some people do not know their true ancestry, or only know their most recent ancestry (or their geographical origin). Others, on the other hand, identify themselves with one ethnic group despite having mixed backgrounds (see Mersha & Abebe 2015). The literature (see Ainsworth, 2015 and Mersha & Abebe, 2015) shows that self-reported ancestry and gender are often inaccurate. In fact, Ainsworth goes so far as to explain that 1 in 100 people are affected by abnormalities in sexual development, resulting in a physical appearance that is inconsistent with their human genome.

The method of the present invention also has the advantage that it can be used for additional quality checks to find sample replacements based on gender and ancestry. Discrepancies between self-declared sex and ancestry and sex and ancestry predicted by the sequencing run, e.g., due to sample shuffling or mishandling by other laboratories may be found.

The term " alignment " generally refers to the computational step of comparing a sequenced sample to a reference sequence to correspond to matching positions in the reference sequence. For this purpose, for each read in the sequence data generated, the corresponding portion of the reference sequence to which that read corresponds must be found. In other words, alignment or read mapping is the process of determining, for a detected nucleic acid sequence read, the portion of the genomic sequence that most likely originated the read. In an exemplary embodiment, the reads are NGS reads, but reads from other sequencing methods are also encompassed by the present disclosure.

Aligned reads obtained from human nucleic acid samples may be displayed, stored, printed, transmitted over a communication network, and otherwise processed further. Additional applications and uses of aligned human nucleic acid samples specifically include one or more of the following.

1) Local realignment around insertions and deletions (InDel)
The term "InDel" refers to base pair insertions or deletions in the genome, typically involving small genetic alterations from 1 base pair to 10,000 base pairs in length. Realignment around insertions and deletions improves subsequent data analysis, particularly mutation identification.

2) Base Quality Score Correction (BQSR)
The term "base quality score" is a base-by-base error rating and represents the particular confidence of a base as determined by a sequencing instrument. Base quality scores may be used, for example, to assess specific evidence of subsequent mutations. BQSR can modify base quality scores to account for regularly occurring technical errors due to the physics or chemistry of the method of sequencing.

3) Machine learning to separate truly distinct variants from machine artifacts common to next-generation sequencing technologies.

4) Mutation discovery and genotyping to find all possible mutations. Also referred to herein as mutation identification processing. Mutation discovery may include discovery of SNP/SNV, InDel, CNV and SV (chromosomal translocations, inversions, duplications, large InDels).

5) Evolutionary Analysis Studies Evolutionary analysis studies may include tools that measure nucleotide diversity, population-to-population differences, linkage disequilibrium, and the frequency spectrum of mutations from one or more populations. Evolutionary analysis may typically involve computational tools that calculate evolutionary sequence statistics. The computational tool may perform a sliding window analysis over the entire chromosome or scaffold. The computational tool may, for example, generate a phylogenetic tree of human nucleic acid samples.

Such evolutionary analysis can be performed, for example, by the "POPBAM" software, for example described at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3767577/.

6) Testing for wild-type biomarkers Additionally, aligned human genomic samples may be tested to determine if wild-type biomarkers are present. A wild-type biomarker is a biomarker that is contained in the PHREG and therefore not detected during the mutation specific procedure. Therefore, the post- alignment computational step includes testing to find each known biomarker. This test indicates whether there are biomarkers in aligned human genomic samples regardless of the PHREG information at the location of interest.

According to one embodiment, the method of the present invention further comprises performing mutation identification processing of the human nucleic acid sample aligned to the selected PHREG. The present invention has the advantage of improving the accuracy of the mutation identification process by introducing tests for sex and/or ancestry first to determine the correct PHREG for use in subsequent alignment and mutation identification steps. There is

Aligned human nucleic acid samples, more specifically NGS reads extracted and localized from human nucleic acid samples, are further processed by one or more so-called variant callers which are computational modules. This variant caller has several different mutation identification processing algorithms that detect any type of mutation (SNV, InDel, copy number variation as somatic mutation , structural variation). Subsequent method steps may include interpreting the mutations. Variant identification processing and/or variant interpretation may be displayed, stored, printed, transmitted over a communication network, or otherwise processed further. The method of the present invention has the advantage of detecting previously undetected biomarkers by removing the bias of the reference genome used. Specifically, the methods of the present invention can distinguish between mutations in a variety of genes, including SNVs, multiple nucleotide variations (MNVs), complex events, as well as large mutations, specifically specifically hotspot mutations, frameshift mutations, non-silent mutations, stop codon mutations, nucleotide insertions, nucleotide deletions, copy number variations, copy number variations as somatic mutations , and/or splice sites , including but not limited to.

A human nucleic acid sample donor is a patient, ie, a person suffering from or suspected of suffering from a given disease. The methods of the present invention should not be considered to apply only to patients.

Mutation identification processing and interpretation of mutations involves analysis of genomic sequences that indicate the presence or absence of a given disease. Based on the interpretation of the mutation, patients are divided into a first group in which a given therapy is not recommended or a second group in which a given therapy is recommended. Thus, the method of the present invention has the advantage that it can be used as part of a disease screening procedure by assessing the presence or absence of a given disease in a patient.

The methods of the present invention may also or alternatively comprise detecting disease symptoms associated with or associated with the human nucleic acid sample. Symptoms of illness may be found, for example, from an electronic health record, or may be entered by the patient himself/herself or a primary care physician via input means of a computing device. Such disease symptoms are identified according to a disease ontology, eg, ISD-10, MeSH, or MeDRA. In finding disease symptoms of a given classification, there are special ontologies that offer advantages such as more accurate classification of disease symptoms. In oncology, it is beneficial to use the ICD-O-3 and/or TNM classification system.

Based on the results of the mutation identification process and mutation interpretation, and taking into account the patient's disease, the method of the present invention may comprise preparing a treatment plan for the patient. In this case, the treatment plan may specifically be a personalized treatment plan. Here, a treatment plan is specifically a personalized treatment plan for a patient, such personalized treatment plan adapted to the patient's genetic data, specifically the patient's clinical, molecular and/or or may include treatment options adapted to genetic status.

In order to determine a potential patient therapy, the method of the present invention may, for example, examine multiple mutations found in the patient, i.e., multiple mutations found in the patient's tumor or in the patient's healthy control tissue. is indicative of the outcome of treating the patient. The method of the present invention may further comprise determining any therapeutics corresponding to any of the mutations found. The method of the present invention scores the determined treatments and ranks them according to their scores to prioritize treatment options or prioritize treatment contraindications for the patient. may include doing

For the purposes of the present invention, the term "treatment" includes prescribing a therapeutic drug or pharmaceutically active compound to prevent, ameliorate, or cure symptoms associated with disease symptoms. The term "treatment" also includes surgery, radiotherapy and/or chemotherapy or combinations thereof.

The improved alignment and mutation identification processes of the present invention provide physicians with improved diagnostic capabilities when performing two alternative therapeutic modalities: disease screening methods or personalized treatment planning. Capabilities, such as improved treatment decisions, can be provided .

According to one embodiment, alignment positioning is performed at the major allele level for PHREG. The major allele level uses the nucleic acid base codes (A, C, G, T) in PHREG to modify the reference sequence for the population. At a given position in the population, the single most common nucleotide is selected. Alleles present in the original reference sequence (eg, GRCh37 or GRCh38) may be used if allele frequencies are the same.

According to another embodiment, the alignment is performed at the non-rare allele level for PHREG. Non-rare allele levels use nucleic acid fuzzy codes that follow established IUPAC nomenclature, eg, "R" for "A" or "G" (see Cornish-Bowden, 1985). The non-rare level encodes 2 or 3, preferably 2 alleles with considerable frequency in the population. A significant frequency is 30%, 20%, 15%, 10%, 5%, 3%, 1% or 0.1% or more, especially 5% or more. It is believed that more than one mutated allele is incorporated into the PHREG for a genomic location, allowing for more accurate read alignment . In one embodiment, only single nucleotide variations (SNVs) are considered at the non-rare allele level. In other embodiments, insertions and deletions (InDel) and other structural alterations are also contemplated.

According to one embodiment, mutation specific processing for PHREG is performed at the major allele level. In certain embodiments, alignments may be performed at the non-rare allele level and variant identification processing may be performed at the major allele level. In an alternative example, variant identification processing is performed at the non-rare allele level.

According to one embodiment, the human reference genome provided in step a) is a published human reference genome. The published human reference genome may include in particular HRG builds, in particular GRCh37 and GRCh38 builds. Additionally or alternatively, the published human reference genome may contain QTRGs. Additionally or alternatively, the published human reference genome may comprise the genome obtained in the 1000 Genomes (1 kG) Project. VCF files of all chromosomes for the 1kG project are available for download from the latest release of the 1kG FTP site, ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ . . If more individual and ethnic datasets were available (e.g., the Thousand Arab Genomes Project examining the population of the United States of America (see Al-Ali, M. et al., 2018)), the You can use them in your method.

Additionally or alternatively, the human reference genome provided in step a) was obtained from a published human reference genome. Here, what is "obtained from"? It may involve error correction and/or modifying the human reference genome for major or non-rare allele levels.

When error correction is performed, the reference nucleotide observed in 0 individuals of a given population is replaced with the corresponding high frequency nucleotide.

In one embodiment, step a) comprises modifying the plurality of human reference genomes to a predetermined coding level, where the coding level comprises either a nucleic acid base code or a nucleic acid fuzzy code . . Coding levels, including base code codes for nucleic acids, are used to define PHREGs, particularly at the major allele level. Coding levels, including ambiguous codes for nucleic acids, are used to define PHREGs, particularly at the non-rare allele level.

In one embodiment, single nucleotide mutations are considered to correct for coding levels. For each population (or superpopulation), use all reported SNVs and their allele frequencies. InDel, CNV and/or SV are also considered in other embodiments.

According to one embodiment, four different levels are proposed for modifying a reference sequence to a given population, two of these four levels being based on the nucleic acid base code (A, C , G, T), the other two use ambiguous codes according to IUPAC (see Cornish-Bowden, 1985), eg, an encoding where "R" stands for "A" or "G". The code levels of such PHREG are defined as follows.

1. Most conservative error correction: A reference nucleotide that is not found in any one person in the population is replaced by the corresponding frequently occurring nucleotide, eg, the corresponding 1 kG frequently occurring nucleotide.

2. Major allele: choose the single most frequently occurring nucleotide at a given locus in the population (if allele frequency is relevant, use the allele present in the original reference sequence (e.g. GRCh37 or GRCh38) ).

3. Non-rare alleles: Code up to 2 alleles of significant frequency (eg, 5% or more) in the population using IUPAC codes if necessary.

4. Complete modeling of detected alleles: Code at each position all (up to 4) reported alleles for at least one person in the population.

However, the full expression of the 1 kG mutation in level 4 PHREG is achieved by disproportionately large number of genome modifications that make the genome unique. , making it extremely difficult to find seed progeny by readmappers. Therefore, in one embodiment, level 3 is adopted and the alignment is performed using an IUPAC ambiguity -aware alignment algorithm. Even the most sophisticated variant callers today are not designed to handle ambiguous nucleic acid codes , so unless a more sophisticated IUPAC ambiguity -aware alignment algorithm is available, the subsequent variant identification process will be at Level 2. Use PHREG.

Thus, the method of the present invention has the advantage that PHREG can be modified for genomic variations in the population at user-defined levels, depending on the target population and subsequent analysis.

According to one embodiment, the human reference genome provided in step a) is PHREG. Thus, step a) may for example involve downloading PHREG from a public data source.

As mentioned above, PHREG is primarily understood as an ancestral-specific reference genome and/or a sex -specific reference genome. In one embodiment, the human reference genome provided in step a) is already population specific, as it contains metadata indicating the ancestry and/or gender of the population. For example, at the time of filing this application, the current Phase 3 analysis of the 1kG project includes 2504 individuals from 26 cohorts and 5 so-called supergroups formed by combining 4 to 7 cohorts. Results for the 5 superpopulations (AFR: Africa, AMR: Mixed America, EAS: East Asia, EUR: Europe, SAS: South Asia) associated with the 26 populations of these 1 kG Phase 3 studies are available at http: Find out at //www.internationalgenome.org/faq/which-populations-are-part-your-study.

In one embodiment, the 1 kG project data is used to generate genomes unique to each of the optimized populations representing each of the 31 (super) populations described above and additional superpopulations including all other populations. To construct.

In the case of the human reference genome provided in step a), PHREG metadata can be provided, eg by downloading from public data sources. Such metadata is useful for quality control of the present invention. If this metadata matches the gender classifier data and the ancestry classifier data, then the quality control can be considered successful. If not, the software may issue an alarm or warning that is displayed to the user, and additionally or alternatively, the software may stop progressing the sequence of steps, for example prior to the alignment step. .

According to one embodiment, testing for gender includes at least one of the following steps. examining at least one location in a gender -specific gene on the X and/or Y chromosome; utilizing differences in localization of multiple human genomic samples on the X and/or Y chromosome; Directly or indirectly determining the sex of a test, FISH analysis, CGH analysis, or human nucleic acid sample.

As noted above, testing for sex may be the result of a by-product of FISH analysis (fluorescence in-situ hybridization analysis) of human nucleic acid samples (see Gall JG 1969). Also, tests to determine gender may be a by-product of CGH analysis (comparative genomic hybridization ) (see Kallioniemi A. et al. 1992).

Gender -probing tests allow efficient and reliable discrimination of male or female nucleic acid samples.

Since multiple individuals from one ancestry or ethnicity share many SNPs that distinguish them from other ancestry or ethnicity, examining multiple SNPs that determine a given range of ancestry can be beneficial to the read alignment and variant identification process. The most appropriate PHREG to use can be identified. Thus, based on the results of the ancestry paper, PHREGs can be selected from multiple sets of human reference genomes.

Multiple different experimental setups can be used in the upstream genomic analysis pipeline steps that verify the ancestry of individuals before proceeding with the alignment to determine the best-matching PHREG criteria and prevent errors.

1) A test for ancestry may be based on a machine learning algorithm used on human nucleic acid samples or another classification scheme that utilizes ancestral specific mutations. The ancestry test may in particular be machine learning based using genotypes at multiple exon positions, eg more than 100, more than 500, more than 1000, more than 2000, preferably more than 5000 exon positions.

2) Appropriate genotyping can be based on NGS data or alternative experimental approaches, such as SNP arrays performed in forensic studies (see Fondevila et al. 2013). Here, using non- coding SNPs helps determine ethnicity.

3) Appropriate genotypes can also be determined in addition to an NGS panel presenting the same non- coding SNPs (and flanking regions) tested in the forensic SNP array from the alternative experimental procedure in 2). .

In particular, testing for ancestry may involve using genotypes for at least one genomic location.

In one specific embodiment, testing for ancestry may involve testing at least one gene selected from the sequence protocols contained herein. 249 genes from the sequence protocol included here for accurate results were shown.

Additionally or alternatively, testing for ancestry may include testing multiple SNP arrays and/or multiple SNP chips, and/or testing markers from Sanger sequencing or mass spectrometry, or to determine appropriate genotypes. may include any other experimental method of

In one specific embodiment, the test examining ancestry is ABL2, ATP1A3, CIC, CYP2C8, CYP2C9, EPHA3, EPHA7, ERBB3, ERG, ETV1, F2, FAS, HFE, IL11RA, IL2RA, ITGB6, KIF11, KIT, KLK3, LRP6, MDM4, NAT2, NTRK2, PDGFB, PIK3R1, PLA2G3, PLAU, PRKCB, RICTOR, SLC7A11, STAT3, T, TSC1, VCAM1, VDR, VEGFB, ACVRL1, AXL, CA9, CALCR, CASP9, ENG, EPHB1, ERBB4, ESR1, FGFR2, HPSE, HSP90AA1, ITK, MRE11A, PLK1, PTPRC, SERPINE1, SMC4, TERT, TLR3, WISP3, WT1, XRCC1, ANGPT2, ARID2, BARD1, CBR3, CDH2, CYP1B1, DDR2, DNMT3A, EPCAM, ERCC2, FANCG, FANCL, GSTP1, IRS2, ITGB1, JAK3, LHCGR, MSH6, NCF2, RNF43, SLC5A5, TMPRSS2, TNFRSF8, AKT1, CD248, CD4, ESR2, EZH2, IGF1R, ITGAV, ITGB2, KLHL6, MAP3K1, MET, MLL, MTHFR, NFKB1, NUP93, PARP8, RB1, RPE65, TSHR, ABL1, BLM, CYP19A1, DPP4, EPHA6, ERBB2, EWSR1, FOXP4, ITGAM, KDM5A, LPA, LTK, MLH1, PBRM1, PHLPP2, SF3B1, TNFRSF10 ABCG2, ACPP, ADAM15, DPYD, EPHA5, EPHB6, FOLH1, KDR, MSH3, MST1R, NTRK1, ROCK2, SLC6A2, TET2, TGM2, TH, ABCB1, CD22, CD40, CD44, CDH20, CYP11B2, ERCC5, GPR124, IL7R, ITGB3, ITGB5, NCL, NOD2, NR4A1, PGR, PLCG1, PPP2R1A, PRAME, PTCH2, RET, SETD2, XPC, ASXL1, EPHB4, PLA2G6, SYK, TET1, EP300, FLT1, ITGA1, LOXL2, PDGFRB, PIK3CD, SSTR5, TEC, APC, ATR, CLU, CREBBP, C YP2D6, EML4, MMP2, PARP2, PDGFRA, TRPM8, CSF1R, DOT1L, FGFR3, FGFR4, GLP2R, IKBKE, JAK1, NOTCH2, SPEN, SPG7, BRCA1, CYP11B1, GNAS, ITGA5, LTF, NRP2, PTK2B, TNKS, ABCC1, CEACAM5, CYP4B1, EGFR, FLT3, INSR, PTCH1, SMARCA4, ZNF217, BCR, EEF2, SELP, SLCO1B1, ABCC2, FLT4, MTR, IL4R, MTOR, RPTOR, TEK, ATM, CARD11, FANCD2, MEFV, NF1, TP73, BRCA2, CD109, PTPRD, ABCC6, IGF2R, P2RX7, ROS1, ACE, PARP1, PRKDC, CENPE, TSC2, ALK, NOTCH1, TNC, NOTCH3, POLE, MLL2, MYH11, POLD1, GRIN3B, F5, FANCA, LRP1B, LRP2, It comprises testing at least one gene selected from the gene group consisting of VWF.

In a more specific embodiment, testing for ancestry comprises testing at least one genomic coordinate selected from the group of genomic coordinates listed in Appendix 1. Appendix 1 lists the GRCh37-based genomic coordinates of the features used in the ancestral classifier. The first three columns are formatted according to the BED file standard (see https://www.ensembl.org/info/website/upload/bed.html) and are (from left to right) chromosomes, with 0 at the left end of the feature. and the end coordinate (that is, the first position after the end position of the characteristic portion) when the left end of the characteristic portion is 0. The fourth column shows the appropriate bases for the feature position classifier. Column 5 shows the corresponding gene name.

Gene names are those accepted by the HUGO Gene Nomenclature Committee (HGNC, see https://www.genenames.org/). HGNC is responsible for authorizing unique codes and names for human loci, including protein-coding genes, ncRNA genes and pseudogenes, to enable unique scientific communication. Gene names used herein were read in August 2013.

In another specific embodiment, the test examining ancestry is in Appendix 2 (see Fondevila et al. 2013)
contains at least one of the SNPs listed in . Appendix 2 shows the chromosomal number (left column) where the SNP is located, the exact chromosomal location (middle column) and the corresponding rs number (right column). where rs number is the grant number given by NCBI (National Center for Biotechnology Information) in the SNP database (dbSNP, see https://www.ncbi.nlm.nih.gov/projects/SNP/), Widely used to refer to specific SNPs across multiple genomic databases. When multiple researchers identify a SNP, they submit a report (including the sequences immediately surrounding the SNP) to the dbSNP database. If duplicate reports are sent, they are merged into the same non-duplicate reference SNP cluster assigned a unique rsid. Further information is available at the URL http://www.ncbi.nlm.nih.gov/sites/books/NBK44406/.

Such ancestral studies include genetic and/or genomic studies that allow discrimination of multiple ancestral categories. Such multiple ancestral categories are defined as AFR, AMR, EAS, EUR and SAS according to the 1kG project. However, the method of the present invention is not limited to the data of the 1kG project, and if more comprehensive data sets are available, e.g. can also

According to one embodiment, the human nucleic acid sample comprises a plurality of sets of reads obtained in a next generation sequencing process . Alignment involves mapping these reads to selected PHREGs. Additionally or alternatively, the human nucleic acid sample includes multiple sets of reads obtained in a targeted sequencing process, eg, panel sequencing .

Advantageously, the method of the present invention can be successfully combined with any existing NGS analysis workflow based on read mapping for HRG.

Before mapping the reads of a human nucleic acid sample to a selected PHREG and aligning the human nucleic acid sample to the selected PHREG, a sequencing library is prepared by randomly fragmenting the DNA or cDNA sample. , followed by 5'- and 3'-adapter ligation . In some embodiments, the fragmentation and ligation reactions are combined in a single step, followed by PCR amplification of the adaptor- ligated fragments.

To map the reads of a human nucleic acid sample to a selected PHREG and align the human nucleic acid sample to the selected PHREG, a plurality of sets of DNA fragments as described above are sequenced to generate about 28 base pairs. It may be assumed to generate multiple reads of 1000 base pairs in length from (see Goodwin S. et al. 2016). This set of DNA fragments contains a sufficient number of reads (usually a few to several thousand) to reach a given coverage of the target region, commensurate with the experimental task at hand.

In one embodiment, the next generation sequencing method comprises whole exome sequencing . In another embodiment, the next generation sequencing method comprises whole genome sequencing . The term "whole exome sequencing " usually refers to a technique for determining the sequence of all protein- coding genes (known as exomes) in a genome. The method consists of first selecting a subset of the protein-encoding DNA (known as exons) and sequencing this DNA using any high -throughput DNA sequencing technique . Humans have approximately 180,000 exons comprising approximately 1.5% of the human genome, or approximately 30 million base pairs. In particular, exome sequencing may be performed by next generation sequencing techniques. "Whole Genome Sequencing " (also known as WGS, full genome sequencing , complete genome sequencing or whole genome sequencing ) is the one-time, laboratory process of determining the complete DNA sequence of an organism's genome. This process entails determining not only the DNA of the organism's chromosomes, but also all of the DNA contained in the mitochondria.

According to another aspect of the invention is a computer system for genetic analysis of human genome samples, the computer system comprising:
a) a first module having computer instructions for providing a set of a plurality of human reference genomes;
b) a second module that tests a human nucleic acid sample for gender and/or ancestry;
c) a third module having computer instructions for selecting one or more population-specific human reference genomes or PHREGs from said plurality of human reference genome sets based on the results of said sex and/or ancestry tests; When,
d) a fourth module having computer instructions for aligning said human nucleic acid sample to said selected one or more PHREGs.

Specifically, the computer system may be adapted or configured to perform any of the methods described above. Thus, the features described with respect to the methods above are disclosed for the computer system, and vice versa, the features described with respect to the computer system are also disclosed for the methods above.

The modules described above may be software modules, software routines or software subroutines stored on a machine readable storage medium, such as non-rewritable or rewritable storage means, or a storage medium used by computer means, e.g. CD-ROM, Stored on portable storage means such as DVDs, Blu-ray discs, sticks or memory cards. Additionally or alternatively, such modules may be provided on a server or cloud server for download via a data network, such as the Internet, or via a communication link, such as a telephone line or radio.

Any of the modules disclosed herein may be multiple functional units, which are not necessarily physically separate from each other. Some of these modular units may be implemented in the form of a single physical unit, which is the case, for example, when several functions are implemented in one software package.

The multiple computer modules disclosed herein may not necessarily be part of an integrated system, but may be distributed among several separate systems interacting with each other through a communications network.

According to one embodiment, the second module for testing human nucleic acid samples for gender and/or ancestry is a computer module having a plurality of computer instructions. Additionally or alternatively, the second module may include wet-lab experiments, such as experiments performing FISH testing. The FISH test results may be analyzed electronically or visually to determine the sex of the sample.

According to another aspect of the present invention, there is provided a computer program which, when executed by a computer, causes said computer to perform said plurality of steps a), b), c) and d in any of said plurality of methods. ) contains multiple instructions that perform

According to yet another aspect of the present invention, a computer-readable storage medium, when executed by a computer, causes the computer to perform the steps a), b), c) and the steps a), b), c) and contains a plurality of instructions to perform d).

As already explained, the methods of the present invention are particularly suitable for identifying abnormalities in a patient's genome that are indicative of a given disease or that the patient is in compliance with a given therapy.

As used herein, the term "disease" includes any disease characterized by one or more genomic abnormalities. The term "disease" includes cancer, autoimmune disease, cardiovascular disease and any genetic disease. The patient may be of any species, preferably a mammal, more preferably a human .

Depending on the disease and its treatment, the person skilled in the art can select the individual treatment mode that will work for the patient.

As a result, yet another aspect of the present invention relates to a method of diagnosing a disease in a patient, the method eliciting an identification of a symptom of the disease in the patient, obtaining a nucleic acid sample from said patient, and performing the steps described herein. Genomic analysis and/or genetic analysis of said nucleic acid sample is performed according to the methods for genomic and/or genetic analysis of human nucleic acid samples described in the literature, and the analysis determines the disease status of said patient.

The identification of disease symptoms may be derived by any known method, for example, as user input, from electronic health or diagnostic records, or from patient databases containing diagnostic records.

With respect to this aspect of the present invention, the term "disease status" means in one embodiment that the patient's illness has been identified. In another embodiment, the term refers to more precisely diagnosing a disease, ie, identifying which subtype of the disease it falls under.

The present invention further relates to a method of treating a disease in a patient, comprising extracting a characteristic result of the patient's disease symptoms, obtaining a nucleic acid sample from said patient, and treating a human nucleic acid sample as described herein. Genomic analysis and/or genetic analysis of the nucleic acid sample is performed according to the method of genomic analysis and/or genetic analysis, and the analysis determines the disease status of the patient and treats the patient.

Yet another aspect of the present invention relates to a method of determining whether a patient is amenable to treatment with a given drug, the method comprising eliciting an indication of a patient's disease symptoms, obtaining a nucleic acid sample from the patient, Genomic analysis and/or genetic analysis of said nucleic acid sample according to the methods of genomic and/or genetic analysis of a human nucleic acid sample described herein to elicit possible treatments for disease symptoms in said patient. , perform mutation specific processing and interpretation of the mutation, and classify the derived potential therapies based on the interpretation of said mutation, each therapy as a desirable and recommended therapy for said patient, or contraindicated for said patient. Classified as a cure.

This method allows determination of treatments available to or effective for a patient. For example, it can be determined whether a given therapy is suitable for a patient or whether side effects of a given therapy are expected to be acceptable.

Again, the identification of disease symptoms may be extracted by any known method, for example, as user input, from electronic health records or electronic diagnostic records, or from patient databases containing diagnostic records.

Possible treatments for the patient's disease symptoms may be extracted from known methods, eg databases.

The present invention further relates to a method of treating a patient, the method eliciting an identification of disease symptoms in the patient, obtaining a nucleic acid sample from said patient, performing genomic analysis of the human nucleic acid sample as described herein and and/or performing genomic and/or genetic analysis of said nucleic acid sample according to methods of genetic analysis to elicit potential treatments for disease symptoms in said patient, mutation identification and interpretation of mutations, and interpretation of said mutations. and classifying each therapy as a desirable and recommended therapy for said patient or as a therapy contraindicated for said patient and a desirable and recommended therapy for said patient one of which is selected and the patient is treated according to the selected therapy.

Again, the identification of disease symptoms may be derived by any known method, for example, as user input, from electronic health or diagnostic records, or from patient databases containing diagnostic records.

Possible treatments for the patient's disease symptoms may be drawn from known methods, again for example from databases.

The foregoing as well as other objects, features, characterizing parts and advantages of the specification will become more apparent and understood by reference to the following detailed description of the invention in conjunction with the accompanying drawings.
FIG. 1 is a schematic flow diagram illustrating a method for genomic and/or genetic analysis of human nucleic acid samples according to the present invention. FIG. 2 is a flow diagram illustrating a method of data analysis in accordance with the present invention. FIG. 3 illustrates the multiple read mapping steps. FIG. 4 is a flow diagram illustrating a method for genomic and/or genetic analysis of human nucleic acid samples according to the present invention. FIG. 5 depicts the distribution of the features chosen for the gender classifier, calculated for the MH panel. FIG. 6 is a boxplot of memory usage and execution time for two gender -ancestor classifiers ( sex classifier and ancestry classifier) and EthSEQ.

DETAILED DESCRIPTION OF THE ACCOMPANYING FIGURES FIG. 1 illustrates a typical workflow for genomic and/or genetic analysis of a human nucleic acid sample, comprising the steps of extracting a human nucleic acid sample and preparing a sequence library. a step, a sequencing step, and a subsequent data analysis step. In the description of the present invention, the steps of extracting a human nucleic acid sample, preparing a sequence library, and sequencing are well-known and standard steps, and will not be described in detail. Details of the data analysis portion of the invention are shown in FIG.

FIG. 2 shows the data analysis steps of FIG. 1, which include a first gender and ancestry testing step followed by alignment (or read mapping), variant identification processing and annotation steps. The input files for the read mapping calculation module are raw sequence data, eg in the form of FASTQ files. The output file for the read mapping calculation module is, for example, a BAM file which is the input file for the mutation identification processing calculation module. An output file for the computational module of the mutation identification process is, for example, a VCF file. An annotation computation module for later use may annotate the data from the VCF file and output the annotated data in the required format, such as PDF or HTML. The file formats used here are only typical ones, and different formats may be used. For example, SAM files, CRAM files, etc. can be used instead of BAM. Also, the data analysis pipeline in FIG. 2 may include multiple computer modules that convert input or output files from one format to another.

FIG. 2 further compares the prior art situation with the present invention situation. The prior art method (indicated by "A" in Figure 2) does not test for gender and ancestry. Alignment and mutation identification procedures are therefore performed relative to standard HRG. A method according to the present invention (denoted by "B" in FIG. 2) performs a test for gender and ancestry that allows selection of one or more PHREGs. Subsequent alignment and mutation identification procedures are performed relative to this selected PHREG.

FIG. 3 outlines a representative read mapping step. In this example, the NGS read has an ancestral unique SNP 'A'. The ancestral unique SNP 'A' is located in close proximity to the previously undiscovered biomarker variation 'G'. Here, the near range is the range up to the length of the lead.

In the alignment step, the NGS reads are compared to the canonical HRG, yielding two mismatches , an ancestral SNP and a biomarker mutation. However, in the alignment step, when the same NGS read is compared to the corresponding PHREG, this PHREG has already been altered at the ancestral-specific position and is identical to the ancestral-specific SNP, so the mismatch caused by the NGS read, That is, there is only one biomarker.

The alignment algorithm uses a scoring system that penalizes any mismatches and/or gaps between the sequenced read and the selected reference genome. As a result, the read aligns to the position with the highest score, or does not align to any position either because all positions score low, or because too many genomic positions have the same alignment score . Reads are less likely to be aligned to HRG than to be aligned to PHREG because of mismatch penalties discovered during alignment algorithms. This is especially the case when another mutation falls within the read length. Therefore, this read is either discarded or, in the worst case, misaligned in the HRG.

Thus, comparison with PHREG has the effect of rescuing reads that originate from regions of the ancestral mutation site, especially if the reads have additional mutations (e.g., disease-causing) in addition to ancestral mutations. have this effect. This makes it possible to detect previously undiscovered biomarkers.

FIG. 4 is a flowchart illustrating a method for genomic and/or genetic analysis of human nucleic acid samples according to the present invention.

In a first step, a set of multiple human reference genomes is provided to a system having a processing unit. For this purpose, the first computer module of the system can download the reference genome from a remote device, eg an internet database. The processing unit may be any programmable computer, including at least a processor, having an internal memory such as RAM to enable storing and executing instructions. The processing unit has access to a non-volatile storage means capable of storing sets of data, such as clinical data and genetic profiles of patients, as well as genetic human reference genomes, as well as computer files. The system can access a communication network such as a LAN or the Internet.

In a second step, the computer system of said system adjusts the human reference genome, preferably to a code level determined by the user of said system. This coding level may comprise a nucleic acid base code or a nucleic acid fuzzy code. In certain embodiments, four different levels of fitting the human reference genome to the population are proposed, specifically maximally conservative error correction, major allele level, non-rare allele level and full modeling of all detected alleles. Two of these four levels use only the base codes (A, C, G, T) of the nucleic acids, and the other two use the fuzzy codes of the nucleic acids according to the IUPAC nomenclature.

In a third step, a patient's human nucleic acid sample is provided. For this purpose, another computer module of the computer system can download raw sequence data, for example in FASTQ file format, from a sequencing laboratory performing sequencing of a sample of interest on a remote platform. In an alternative embodiment, sequencing may be performed by a department that performs analysis of nucleic acid samples and transfers the results internally. In connection with the third step, the system may receive clinical data of the patient from input sources, such as information about the disease the patient is suffering from and information about current treatments for it. The patient's clinical data may, for example, be received directly from the patient, typed at a keyboard, or inferred from text typed at a keyboard, or may be received from a multi-function selectable element in a GUI. Patient clinical data may be received from an electronic health record (EHR) or electronic medical record (EMR) and stored on a chip card or in a database searchable via a communication network.

In a fourth step, human nucleic acid samples are tested to predict gender and/or ancestry. Again, this test may be conducted in close proximity to the subject, or another computer module of the system may be used to read test results from an external service provider via a communications network. Gender and/or ancestry testing may be performed by a second computational module or another wet lab experiment.

In a fifth step, one or more PHREGs are selected from the set of human reference genomes based on the results of the gender and/or ancestry tests of the fourth step. This selection may be made by the third computing module.

In a sixth step, the human nucleic acid sample is aligned to the PHREG of choice. This involves mapping the set of reads obtained from this aligned NGS step to the PHREG of choice. This alignment may be performed in a fourth computer module and the output file may be a BAM file.

In a seventh step, mutation identification processing of the aligned human nucleic acid samples is performed relative to the selected PHREG. Prior to mutation identification processing, a predetermined computer module of the system may readjust the human reference genome, preferably to a predetermined code level set by the user of the system. This coding level may include the base code of the nucleic acid or the fuzzy code of the nucleic acid and may be different from the coding level used in the alignment step. Mutations are identified using the most suitable state-of-the-art algorithms. The mutation identification process is performed by a fifth computing module, the output of which may include sequence data in the form of mutations relative to PHREG in mutation identification processing format (VCF file).

In an eighth step interpretation of mutations is performed. The system may include another post-processing computational module that enables analysis of the identified mutations. In one embodiment, this post-processing computational module may analyze sets of genes and/or mutation sites that are indicative of the presence or absence of a given disease in a patient. Additionally or alternatively, the post-processing computational module considers additional clinical data of the patient to determine a plurality of therapy sets for the patient's disease, and further the patient's genetic data, specifically the identified Gene mutations may be used to determine the most appropriate therapy for the patient. In yet another embodiment, this post-processing computational module performs statistical analysis to determine mutation weights, nucleotide substitution rates and hotspot mutations from the identified mutations.

The mutations found can be used as classifiers predictive of therapeutic efficacy or safety or as classifiers for diagnostic or therapeutic purposes.

In a ninth step, diagnostic and/or therapeutic suggestions are made and provided. For this purpose, it may comprise an output interface operatively connected to the third, fourth, fifth and post-processing calculation modules such that the results of these modules are output. This output interface may be associated with any display means or printer that allows the information computed by the processing unit to be presented. In addition, there may be a link to a communication system for an intranet and/or a link to the Internet, such as a program for sending and receiving e-mail implemented via an output interface.

FIG. 5 is a chart showing the distribution of features for gender identification selected for each classification (F: female; M: male) calculated using MH panel data. Colored vertical lines represent class medians.
Graph of (a): Graph of X chromosome/Y ratio of localized reads (b): major alleles interrogated at 500 common SNP positions on the X chromosome, ranging from 0.8 to 1.0. Graph of Frequency (c): Percentage of Correctly Paired Reads on the Y Chromosome FIG.

FIG. 6 is a boxplot of memory usage and execution time for all exome samples of 300 TCGA with two gender -ancestral classifiers and EthSEQ. FIG. 6 should be viewed in light of the examples described below.

We present an antextry, a machine-learning-based tool that uses read alignments to determine sample sex and ancestry from whole -exome sequencing data. Subjects' own claims for both traits are known to be unreliable. Anxextri predictions are useful in terms of sample mix-up detection and also for use in unbiased interpretation of genomic mutation sites. It is especially useful when dealing with large groups. Performance evaluation tests of Ansectri using over 1300 samples have shown that Ansectri is highly accurate and has low time and memory requirements.

1. Introduction Next-generation sequencing of large populations has become increasingly common due to the rapid cost decline observed over the past decade (Cancer Genome Atlas Research Network et al., 2013; Rand et al., 2013). 2016), whole -exome approaches play a major role in large-scale studies. In particular, it is used in the fields of precision medicine and comprehensive characterization of disease. In this situation, knowing the correct ancestry and sex of the sample has many advantages. First, accurate knowledge of sample ancestry and sex facilitates quality control by helping to identify sample mix-ups caused by the complex procedures and manual work required for sample processing. Second, to avoid the strong European ancestry bias present in most genomic studies and in the human reference genome, and to improve clinical care in persons of diverse ancestry, ancestry should be considered the influence of mutations. (see Etrovski et al., 2016; Mersha et al., 2015; Fakhro et al., 2016). Finally, ancestry is widely used in genetic association studies to avoid false disease associations due to population stratification (see Wu et al., 2011). Because self-reports of sex and ancestry are often unreliable (see Mersha et al., 2015; Ainsworth, 2015), identification using genomic information is necessary.

"AnSextry," a machine learning method based on logistic regression analysis, was created to rapidly and reliably characterize sex and ancestry from whole-exome sequence pair-end read alignments . The algorithm relies on standard file formats and can be readily integrated into existing next-generation sequencing workflows. This algorithm provides a ready-to-use model and requires a simple BAM file as input. In addition, the algorithm has low memory requirements and runs on desktop computers. A comparative study with EthSEQ (see Romanel et al., 2017), the only other whole-exome BAM-file-based ancestry estimation tool, showed that 'Ansectri' was well comparable in accuracy, execution time and memory usage. indicates that there is To date, there are no other published methods for gender prediction.

2. Method 2.1 Algorithm
A set of two classifiers was prepared that estimated the most likely sex and ancestry of an individual based on whole-exome sequence pair-end read alignments . This tool utilizes read mapping and individual human genotype differences for prediction.

Gender and ancestry classifiers were based on logistic regression analysis using Python and Scikit-learn. Features corresponding to both of these classifiers were determined from the input BAM file. Paired-end reads were aligned using BWA 0.7.15 defaulted for alignment and no post-processing steps such as local realignment or duplicate elimination. A GRCh37 reference genome was used. The genome has no non-chromosomal supercontigs and has masked pseudo -autosomal regions PAR1 and PAR2 to avoid misalignment on the X and Y chromosomes. In the context of the present invention, the term "supercontig" is commonly understood as an ordered set of contigs, i.e., a contiguous length of genomic sequence in which the order of multiple bases is known with a high level of confidence. .

The gender classifier was worked by a two-class logistic regression analysis with L1 regularization to output the probability of each class. A 5-fold cross-validation was used to determine the appropriate regularization strength. The model that produced the highest area when a PR curve (Precision-Recall Curve) was drawn for the training data was selected as the optimal model to evaluate the set of test data.

The ancestry classifier was based on multinomial logistic regression analysis with L2 regularization and Principal Component Analysis and output probabilities for each of the five continental ancestry defined in the 1000 Human Genomes Project. The ancestry by five continents is African (AFR), Mixed American (AMR), East Asian (EAS), European (EUR) and South Asian (SAS) (The 1000 Genomes Project Consortium et al. , 2015). A 5-fold cross-validation was used to determine multiple suitable parameters. A model that gave the highest F1 score for training data was selected, and the model was applied to test data for evaluation.

2.2 Features The features used in the sex classifier were based on the alignment difference between the X and Y chromosomes (see Figure 5). The ratio of X-chromosome reads to Y-chromosome reads was used as well as the percentage of correctly paired reads on the Y chromosome. In addition, we combined the frequencies of major alleles at SNP positions in 500 well-known exonic regions on the X chromosome. SNPs with high frequency among the primary ancestors were chosen to eliminate population bias.

Common target regions of Agilent's All Exon Kit (Version 5, Version 6, Version 6 + COSMIC) and Molecular Health's Pan Cancer Gene Panel (target size 2.9 Mbp) for use in ancestry classifiers All autosomal SNPs with genomic locations within the segment were genotyped from the 1000 genomic data described in 2.3. Feature selection left valid SNPs showing differences among multiple ancestors, resulting in 10000 genotypes corresponding to 5040 genomic locations to be used as features for the ancestry classifier. The corresponding BED file is shown in Appendix 1 and can be used to determine overlap with any targeted sequencing kit.

2.3 Data To obtain data from diverse ancestry, an ancestry classifier was trained using genomic data from 1735 individuals from the 1000 Human Genomes Project Phase 3. Multiple ancestry by continent (AFR, AMR, EAS, EUR, SAS) was used for classification and multiple individuals were randomly selected to balance each classification. 694 individuals were part of the study set.

A study set of primary whole exome control data from 300 self-reported race and sex individuals corresponding to 3 types of cancer (bladder cancer, lung adenocarcinoma/squamous cell lung cancer, gastric cancer) Downloaded from TCGA (see cancergenome.nih.gov). All samples were sequenced using Agilent's SureSelect Human All Exon 50Mb kit. Data were randomly selected to bring the size of the balanced class corresponding to the TCGA categories. 150 men and 150 women, 100 Caucasians, 100 Asians and 100 Blacks or African Americans.

Sequencing data from 988 self-reported gender cancer patients were sequenced using Molecular Health's pan-cancer gene panel, and the data was used to train and test a sex classifier. . Individuals were randomly selected to balance the female/male classification. 396 cases were randomly selected as test data for the sex classifier. The 300 TCGA cases described above were used as an additional test set.

3. 3. Results 3.1 Gender Classifier A gender classifier was trained using 592 data sets sequenced by the Molecular Health Pan Cancer Gene Panel. Paired-end reads were aligned and features were calculated as described in the methods section. After tuning the method by tolerance verification, the method was applied to two data sets for performance evaluation. The datasets used were 396 individuals sequenced by the gene panel described above and 300 TCGA individuals sequenced by available whole-exome data.

Based on panel test data, the gender classifier had an average accuracy of 97.5%, misclassifying 10 individuals (5 male, 5 female) (see Table 1). Misclassification was not associated with low coverage.

Table 1 provides detailed data for individuals sequenced by the Molecular Health Pan Cancer Gene Panel. This data shows when the predicted gender did not match the self-reported gender . The median coverage for all samples is 2116. The average coverage of all misclassified samples is close to or greater than this median, and misclassification is not considered related to lower than median coverage.

Since there is a 1% chance of abnormalities in gender development in the general population (see Ainsworth, 2015), some of the misclassified cases were in fact correctly classified, but self-reported The gender may have been incorrect.

Based on the TCGA test data, the accuracy of the sex classifier reached 100%. All 300 were correctly classified. In terms of execution time and memory usage, gender prediction took less than 1 minute in all cases, and the average memory usage was 526MB (see Figure 6).

3.2 Ancestral Classifier An ancestral classifier was trained on 1041 data sets from the 1000 Human Genomes Project. Each individual human genotype was used as a feature, as described in 2.2. The best performing model was determined on two test data sets. The two test data sets are 300 TCGA individuals whose whole exomes have been sequenced and the remaining 694 individuals from the 1000 Human Genomes Project.

The average accuracy of the ancestry classifier that classified the 1000 human genome test data was high, reaching 99%. The highest accuracy was for Asian ancestry, followed by African and South American ancestry (99% accuracy), followed by European ancestry (98% accuracy). % Accuracy). In total only 5 out of 694 were misclassified.

The ancestral classifier's classification results for the 300 TCGA exome test data set were slightly less accurate at 96.33%, with a total of 11 misclassified individuals. These results are comparable to EthSEQ (see Romanel et al., 2017). EthSEQ is the only other known ancestry prediction method that provides a suitable pre-computed model and is directly applicable to a single whole exome BAM file. Although these two results are in very good agreement, EthSEQ was slightly less accurate (94%), misclassifying a total of 18 individuals. Furthermore, EthSEQ requires longer execution time and requires more memory. For the ancestor classifier of the present invention, the average execution time is 28 seconds and the average memory usage is 540 MB, while for EthSEQ, it averages 4.5 seconds despite having multi-threading (4 cores). It took 8 minutes and used 14.7 GB on average (see Figure 6).

One important takeaway from the results is that the results of these two algorithms on the misclassified data sets are in very good agreement. Ten of the 11 who were not matched to the race given by TCGA were also misclassified by EthSEQ, and for 8 of these 10 cases both methods predicted the same ancestry. This suggests that the TCGA classification may have been incorrect for these misclassified individuals, where the TCGA race information is based on self-reports. Six of the ten commonly misclassified were predicted to be AFR or AMR. This result is consistent with Mersha et al.'s explanation that self-report errors are relatively common in African American and Latino populations. Table 2 shows the results for those who were misclassified.

Table 2 details TCGA individuals whose predictions (by Ansectry or EthSEQ or both) were not consistent with TCGA self-reported race. TCGA race includes “Black or African American (Black/African American),” “Caucasian,” and “Asian.” White rows corresponded to samples that were neither unsextri nor EthSEQ matched to TCGA race. Light gray rows were samples where only EthSEQ predictions were inconsistent with TCGA. In addition, the dark-colored samples were the only samples in which Ansectri's predictions did not agree with TCGA. If the coverage of a locus was insufficient, the genotype at that locus was inferred from the criteria for prediction of the unsector. The median coverage of all samples was 91x, and the coverage of the majority of misclassified samples was above this median, thus suggesting that misclassification is associated with coverage lower than the median. was inconceivable. Furthermore, the median number of estimated genotypes for the unclassified class of all samples was 390, which was close to the median number of misclassified unclassified samples (393). The number of inferred genotypes for all 300 TCGA samples varied between 227 (lowest) and 690 (highest), with 10-15% of inferred genotypes adversely affecting unsextri predictions. cannot be considered to have given Interestingly, the only individual that the unsextri misclassified but not the EthSEQ misclassified was classified as Caucasian according to the TCGA, whereas the unsextri classifier actually predicted that he was mixed race, with a probability of AMR54 .7% and EUR45,1% probability.

4. CONCLUSIONS: Ansectri, a novel method to reliably and easily determine an individual's sex and ancestry, based on aligned paired-end reads from whole exomes or, if target size permits, targeted sequencing experiments . showed . This tool provides two Python-based classifiers that rely on logistic regression analysis, and the prediction of ancestry by this tool is an alternative to the predominantly PCA-based methods used in the field of population genetics. Become. Ansectri provides an out-of-the-box reference model and requires minimal user input. Ansextry is fast, accurate, and easy to use.

Disclaimer
In this specification, the terms "ancestry-specific"/"ethnic-specific"/"group-specific" are interchangeable, as different authors use different terms for the same purpose. used for

References
1. Lander, ES et al. Initial sequencing and analysis of the human genome.
Nature 409: 860-921 (2001). [PMID: 11237011]
2. Church, DM et al. Modernizing reference genome assemblies. PLoS Biol. 9: e1001091 (2011). [PMID: 21750661]
3. Harrow, J. et al. GENCODE: the reference human genome annotation for The ENCODE Project. Genome Res. 22: 1760-1774 (2012). [PMID 22955987]
4. ENCODE Project Consortium. An integrated encyclopedia of DNA elements in the human genome. Nature 489: 57-74 (2012). [PMID: 22955616]
5. 1000 Genomes Project Consortium et al. A global reference for human genetic variation. Nature 526: 68-74 (2015). [PMID: 26432245]
6. Li H & Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25: 1754-1760 (2009). [PMID: 19451168]
7. DePristo, MA et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat. Genet. 43: 491-498 (2011). [PMID: 21478889]
8. Horton, R. et al. Variation analysis and gene annotation of eight MHC haplotypes: the MHC Haplotype Project. Immunogenetics 60: 1-18 (2008). [PMID: 18193213]
9. Pei, B. et al. The GENCODE pseudogene resource. Genome Biol.13: R51 (2012). [PMID: 22951037]
10. Degner, JF et al. Effect of read-mapping biases on detecting allele-specific expression from RNA-sequencing data. Bioinformatics 25: 3207-3212 (2009). [PMID: 19808877]
11. Brandt, DYC et al. Mapping Bias Overestimates Reference Allele Frequencies at the HLA Genes in the 1000 Genomes Project Phase I Data. G3 5: 931-941 (2015). [PMID: 25787242]
12. Novak A.; Hickey G.; Garrison E.; Blum S.; Connelly A.; Kural D.; Li H.; Lin M.; Miga K.; Ouyang N.; /www.biorxiv.org/content/biorxiv/early/2017/01/18/101378.full.pdf)
13. Paten B, Novak AM, Eizenga JM, Garrison E. Genome graphs and the evolution of genome inference. Genome Res. 5: 665-676 (2017) [PMID: 28360232]
14. Snyder M., et al. Personal genome sequencing: current approaches and challenges. Genes Dev. 5, 423-431 (2010) [PMID: 20194435]
15. Young, AL et al. A new strategy for genome assembly using short sequence reads and reduced representation libraries. Genome Res 2: 249-256 (2010) [PMID:20123915]
16. Flicek, P & Birney, E. Sense from sequence reads: methods for alignment and assembly. Nat Methods. 6: S6-S12 (2009) [PMID 19844229]
17. Chen R . & Butte AJ The reference human genome demonstrates high risk of type 1 diabetes and other disorders. Pac Symp Biocomput. 2011:231-242 (2011) [PMID: 21121051]
18. International Human Genome Sequencing Consortium. 2001. Initial sequencing and analysis of the human genome. Nature 409: 860-921 (2001) [PMID: 11237011]
19. International Human Genome Sequencing Consortium. 2004. Finishing the euchromatic sequence of the human genome. Nature 431: 931-945 (2004) [PMID: 15496913]
20. Schneider VA et al. Evaluation of GRCh38 and de novo haploid genome assemblies demonstrates the enduring quality of the reference assembly. Genome Res. 5:849-864. (2017) [PMID: 28396521]
21. [Editorial (October 2010). "E pluribus unum". Nature Methods. 5: 331. doi:10.1038/nmeth0510-331.
22. Nielsen R., Paul JS, Albrechtsen A., Song YS Genotype and SNP calling from next-generation sequencing data. Nat. Rev. Genet. 12: 443-45. (2011) [PMID: 21587300]
23. Fakhro, KA, Staudt MR, Ramstetter MD, Robay A., Malek JA, Badii R., et al. The Qatar genome: a population-specific tool for precision medicine in the Middle East. Hum. Genome Var. 16016 Human Genome Variation (2016) 3, 16016 doi:10.1038/hgv.2016.16; published online 30 June 2016 (2016) [PMID: 27408750]
24. Zayed H. The Qatar genome project: translation of whole-genome sequencing into clinical practice. Int J Clin Pract. 10: 832-834 doi: 10.1111/ijcp.12871. Epub 2016 Sep (2016) [PMID: 27586018]
25. Sanger F., et al. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci US A. 74:5463-5467. (1977) [PMID: 271968]
26. Venter, JC et al. The Sequence of the Human Genome. Science 291: 1304-1351. (2001) [PMID: 11181995]
27. Petrovski S & Goldstein DB Unequal representation of genetic variation across ancestry groups creates healthcare inequality in the application of precision medicine. Genome Biol 2016;17:157.doi: 10.1186/s13059-016-1016-y. (2016) [PMID : 27418169]
28. Koboldt DC, Ding L, Mardis ER, Wilson RK. Challenges of sequencing human genomes. Brief Bioinform. 11:484-498. (2010) [PMID: 20519329]
29. Dewey FE, Chen R., Cordero SP, Ormond KE, Caleshu C., Karczewski KJ et al. Phased whole-genome genetic risk in a family quartet using a major allele reference sequence. PLoS Genet. 2011 Sep;7(9 ):e1002280. doi: 10.1371/journal.pgen.1002280. Epub 2011 Sep 15. (2011) [PMID: 21935354]
30. Cao H, Wu H, Luo R, Huang S, Sun Y, Tong X et al. De novo assembly of a haplotype-resolved human genome. Nat Biotechnol 33: 617-622. (2015) [PMID: 26006006]
31. Wu L., Yavas G., Hong H., et al. Direct comparison of performance of single nucleotide variant calling in human genome with alignment-based and assembly-based approaches. Sci Rep. 2017 Sep 8;7(1) :10963. doi: 10.1038/s41598-017-10826-9. (2017) [PMID: 28887485]
32. Meyer, LR et al. The UCSC Genome Browser database: extensions and updates 2013. Nucleic acids research41: D64-D69 (2013). [PMID: 23155063]
33. Sudmant, PH et al. An integrated map of structural variation in 2,504 human genomes. Nature 526: 75-81 (2015). [PMID: 26432246]
34. Iqbal, Z., Caccamo, M., Turner, I., Flicek, P. & McVean, G. De novo assembly and genotyping of variants using colored de Bruijn graphs. Nature genetics 44: 226-232 (2012). [PMID: 22231483]
35. Cornish-Bowden A. (1985). Nomenclature for incompletely specified bases in nucleic acid sequences: recommendations 1984. Nucleic Acids Res. 13: 3021-3030.(1985) [PMID: 2582368]
36. Mersha TB, & Abebe T. Self-reported race/ethnicity in the age of genomic research: its potential impact on understanding health disparities. Hum. Genomics 9:1. (2015) [PMID: 25563503]
37. Baye TM Inter-chromosomal variation in the pattern of human population genetic structure. Hum Genomics 5:220-240. (2011) [PMID: 21712187]
38. Fondevila M. et al. Revision of the SNPforID 34-plex forensic ancestry test: Assay enhancements, standard reference sample genotypes and extended population studies. Forensic Sci Int Genet 7: 63-74. (2013) [PMID: 22749789]
39. Ainsworth C. Sex redefined. Nature518: 288-291. doi: 10.1038/518288a. (2015) [PMID: 25693544]
40. Gall JG, Pardue ML Formation and detection of RNA-DNA hybrid molecules in cytological preparations. Proc. Natl. Acad. Sci.
41. Kallioniemi A. et al. Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science Band 258, Nr. 5083, 1992, S. 818-821.
42. Goodwin S., McPherson JD, McCombie WR. Coming of age: ten years of next-generation sequencing technologies. Nat. Rev. Genet. 2016 May 17;17(6):333351
43. Al-Ali M, Osman W., Tay GK, AlSafar HS A 1000 Arab genome project to study the Emirati population. J. Hum. Genet. 63(4): 533-536 (2018). [PMID: 29410509]
44. Cancer Genome Atlas Research Network et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat. Genet., 45(10), 1113-1120 (2013).
45. Rand, KA et al. Whole-exome sequencing of over 4100 men of African ancestry and prostate cancer risk. Hum. Mol. Genet., 25(2), 371-381 (2016).
46. Wu,C. et al. A Comparison of Association Methods Correcting for Population Stratification in Case-Control Studies. Ann. Hum. Genet., 75(3), 418-427 (2011).
47. Romanel, A. et al. EthSEQ: ethnicity annotation from whole exome sequencing data. Bioinformatics, 33(15), 2402-2404 (2017).

By using PHREG as a basis for NGS read mapping, we increase the coverage of biomarkers suitable for clinical use . ) were used. This dataset consists of 155 samples of African (AFR) ancestry, 33 samples of Latino/Mixed American (AMR) ancestry, 354 samples of European (EUR) ancestry, and South Asian (SAS) had 20 samples of ancestry. Each sample was cloned from the standard human reference genome (HRG) GRCh37 using Novoalign 4.00.1.
(see reference 3), to the PHREG defined by our ancestral classifier, and also to the HSA PHREG. The HSA PHREG was created by collecting mutation data for all of the Gnom v2.1 ancestors (see reference 4), including AFR, AMR, EAS, EUR and SAS.

To perform these read-mapping strategies, we used 1548 three pathogenic ClinVar biomarkers in the Gencode v31 CDS exon (see ref. 6) targeting 1288 genes.・We compared the coverage of version 2019-12 (see reference 6). Increased Climber biomarker coverage when aligned to PHREG but not HRG: 211 for AFR, 147 for AMR, 121 for EAS, 173 for EUR, 105 for SAS, and more 162 in HSA. Most of the mutations with increased coverage were near sites where population-specific nucleotides were embedded in PHREG. By mapping the reads of a sample to the PHREG closest to that sample, fewer mismatches occur during alignment , resulting in increased coverage and no loss of coverage when aligned against HRG. .

In summary, our analysis shows that correct PHREG can increase coverage and thus improve detection of biomarkers suitable for clinical use.

Brief description of Table 3 (ClinVar_PHREG_coverage_diff_relative.xlsx)
List of Climber biomarkers in exons of the Gencode CDS (gene name|contig| beginning | end ) showing differences in coverage when aligned to PHREG compared to HRG.
Coverage of each PHREG (AFR, AMR, EAS, EUR, SAS, HSA) as median for all cases by ancestry and all 741 cases (HSA) for coverage calculated based on alignment to HRG give a difference. A positive number means an increase in coverage, a negative number means a decrease in coverage.

References for Example 2
[1] https://portal.gdc.cancer.gov
[2] http://www.novocraft.com/products/novoalign
[3] https://www.ncbi.nlm.nih.gov/grc/human
[4] https://gnomad.broadinstitute.org/faq
[5] https://www.ncbi.nlm.nih.gov/clinvar
[6] https://www.gencodegenes.org/human/release_31lift37.html

[Table 1]

Appendix 2
chr1 36768200 rs1573020
chr1 159174683 rs2814778
chr1 204790977 rs2065160
chr2 7149155 rs896788
chr2 109513601 rs3827760
chr2 136616754 rs182549
chr3 168645035 rs1498444
chr4 38803255 rs4540055
chr4 159181963 rs2026721
chr5 33951693 rs16891982
chr7 4457003 rs917118
chr10 17064992 rs7897550
chr10 34755348 rs1978806
chr11 32424389 rs5030240
chr12 29369871 rs10843344
chr12 56603834 rs773658
chr13 20901724 rs1335873
chr13 22374700 rs1886510
chr13 34864240 rs2065982
chr14 36170607 rs10141763
chr14 101142890 rs730570
chr15 28365618 rs12913832
chr15 48426484 rs1426654
chr16 31079371 rs881929
chr16 90105333 rs3785181
chr17 75551667 rs2304925
chr18 75432386 rs1024116
chr19 42410331 rs2303798
chr20 38849642 rs1321333
chr21 16685598 rs722098
chr21 17710424 rs239031
chr21 25672460 rs2572307
chr22 26350103 rs5997008
chr22 47836412 rs2040411

Claims (14)

A sequence data analysis method for genomic/genetic analysis of a human nucleic acid sample, comprising:
(a) providing a group of a plurality of human reference genome sequences including at least one gender-specific reference genome sequence and at least one ancestry-specific reference genome sequence ;
(b) testing the sequence of the human nucleic acid sample for gender and ancestry ;
(c) one sex-specific reference genome sequence and one or more ancestry-specific reference genome sequences from the group of said plurality of human reference genome sequences based on the results of the tests for gender and ancestry of step (b); and
(d) aligning the sequence of the human nucleic acid sample to the sex-specific reference genomic sequence selected in step (c) and the ancestry-specific reference genomic sequence selected in step (c) ;
(e) identifying variations in the sequence of the human nucleic acid sample aligned in step (d) against the sex-specific reference genomic sequence selected in step (c) and the ancestry-specific reference genomic sequence selected in step (c); a step of
method including. 2. The method of claim 1, wherein said alignment is performed at the major allele level or the non-rare allele level. 4. The method of claim 3, wherein the identification of said mutations is performed at the major allele level or the non-rare allele level. 4. The method of any of claims 1-3 , wherein the plurality of human reference genome sequences provided in step (a) are obtained from published human reference genome sequences . The testing for gender in step (b) comprises testing at least one position in a sex -specific gene on the X or Y chromosome; alignment of a plurality of human genomic samples on the X or Y chromosome; 5. The method of any of claims 1-4 , comprising one or more of utilizing differences, cytogenetic testing, FISH analysis and CGH analysis. 6. Any of claims 1-5 , wherein the test for ancestry in step (c) is based on a machine learning algorithm used on sequences of human nucleic acid samples or another classification scheme that utilizes ancestry-specific mutations. how to. The testing for the ancestry in step (c) includes using genotypes for at least one genomic location, testing multiple SNP arrays or multiple SNP chips, using markers from Sanger sequencing or mass spectrometry. 7. A method according to any preceding claim, comprising one or more of: testing. The progenitor tests include ABL2, ATP1A3, CIC, CYP2C8, CYP2C9, EPHA3, EPHA7, ERBB3, ERG, ETV1, F2, FAS, HFE, IL11RA, IL2RA, ITGB6, KIF11, KIT, KLK3, LRP6, MDM4, NAT2, NTRK2, PDGFB, PIK3R1, PLA2G3, PLAU, PRKCB, RICTOR, SLC7A11, STAT3, T, TSC1, VCAM1, VDR, VEGFB, ACVRL1, AXL, CA9, CALCR, CASP9, ENG, EPHB1, ERBB4, ESR1, FGFR2, HPSE, HSP90AA1, ITK, MRE11A, PLK1, PTPRC, SERPINE1, SMC4, TERT, TLR3, WISP3, WT1, XRCC1, ANGPT2, ARID2, BARD1, CBR3, CDH2, CYP1B1, DDR2, DNMT3A, EPCAM, ERCC2, FANCG, FANCL, GSTP1, IRS2, ITGB1, JAK3, LHCGR, MSH6, NCF2, RNF43, SLC5A5, TMPRSS2, TNFRSF8, AKT1, CD248, CD4, ESR2, EZH2, IGF1R, ITGAV, ITGB2, KLHL6, MAP3K1, MET, MLL, MTHFR, NFKB1, NUP93, PARP8, RB1, RPE65, TSHR, ABL1, BLM, CYP19A1, DPP4, EPHA6, ERBB2, EWSR1, FOXP4, ITGAM, KDM5A, LPA, LTK, MLH1, PBRM1, PHLPP2, SF3B1, TNFRSF10A, ABCG2, ACPP, ADAM15, DPYD, EPHA5, EPHB6, FOLH1, KDR, MSH3, MST1R, NTRK1, ROCK2, SLC6A2, TET2, TGM2, TH, ABCB1, CD22, CD40, CD44, CDH20, CYP11B2, ERCC5, GPR124, IL7R, ITGB3, ITGB5, NCL, NOD2, NR4A1, PGR, PLCG1, PPP2R1A, PRAME, PTCH2, RET, SETD2, XPC, ASXL1, EPHB4, PLA2G6, SYK, TET1, EP300, FLT1, ITGA1, LOXL2, PDGFRB, PIK3CD, SSTR5, TEC, APC, ATR, CLU, CREBBP, CYP2D6, EML4, MMP2, PARP2, PDGFRA, TRPM8, CSF1R, DOT1L, FGFR3, FGFR4, GLP2R, IKBKE, JAK1, NOTCH2, SPEN, SPG7, BRCA1, CYP11B1, GNAS, ITGA5, LTF, NRP2, PTK2B, TNKS, ABCC1, CEACAM5, CYP4B EGFR, FLT3, INSR, PTCH1, SMARCA4, ZNF217, BCR, EEF2, SELP, SLCO1B1, ABCC2, FLT4, MTR, IL4R, MTOR, RPTOR, TEK, ATM, CARD11, FANCD2, MEFV, NF1, TP73, BRCA2, CD109, Genes consisting of PTPRD, ABCC6, IGF2R, P2RX7, ROS1, ACE, PARP1, PRKDC, CENPE, TSC2, ALK, NOTCH1, TNC, NOTCH3, POLE, MLL2, MYH11, POLD1, GRIN3B, F5, FANCA, LRP1B, LRP2, and VWF 8. A method according to any one of claims 1 to 7 , comprising testing at least one gene selected from the group. The human nucleic acid sample has a set of reads derived from next generation sequencing or NGS, and the alignment aligns the reads with the sex-specific reference genomic sequence selected in step (c). 9. A method according to any preceding claim, comprising mapping against the ancestry-specific reference genomic sequence chosen in ) . A computer system for genomic or genetic analysis of sequences of human nucleic acid samples, comprising:
(a) a first module having computer instructions for providing a group of a plurality of human reference genomic sequences including at least one gender-specific reference genomic sequence and at least one ancestry-specific reference genomic sequence ;
(b) a second module that tests the sequence of the human nucleic acid sample for gender and ancestry;
(c) one gender-specific reference genome sequence and one or more ancestry-specific sequences from the group of the plurality of human reference genome sequences based on the results of the tests for gender and/or ancestry in step (b); a third module having computer instructions for selecting a reference genome sequence ;
(d) computer instructions to align the sequences of said human nucleic acid sample to said one sex-specific reference genomic sequence selected in said step (c) and to one or more ancestry-specific reference genomic sequences selected in said step (c) ; a fourth module comprising
(e) identifying variations in the sequence of the human nucleic acid sample aligned in step (d) against the sex-specific reference genomic sequence selected in step (c) and the ancestry-specific reference genomic sequence selected in step (c); a fifth module having computer instructions for performing
A computer system having A computer program product comprising instructions which, when executed by a computer, cause the computer to perform the steps (a) to (d) of any one of claims 1 to 9 . A computer readable storage medium having instructions that, when executed by a computer, cause the computer to perform steps (a) to (d) of any one of claims 1 to 9 . A method of detecting disease-associated mutations in the sequence of said human nucleic acid sample by performing steps (a) through (e) of the methods of claims 1 and 3. A method of detecting mutations associated with a therapeutic effect on a disease in the sequence of said human nucleic acid sample by performing steps (a) through (e) of the methods of claims 1 and 3.

Images