JP6208622B2 - Analysis device, database creation method, and system - Google Patents

Description

  The present invention relates to an analysis apparatus, a database creation method, and a system.

  Next generation sequencing (NGS) using next-generation sequencers has made it possible to determine genome sequences at a dramatically lower cost than conventional Sanger methods. Genomic sequencing costs, which were about US $ 100 million in 2001, have been reduced to US $ 5,826 in 2013 by next-generation sequencing (DNA Sequencing Costs, http://www.genome.gov/sequencingcosts/) . Next-generation sequencer that is not only low in cost, but can obtain a huge amount of sequence data in a short period of time, and can generate a huge amount of sequence data exceeding 1 trillion bases by operating the device once. Is also present. With this technology, a large-scale project that determines the genome sequence of a large number of subjects can be executed, and a number of projects that determine the genome sequence of a large-scale cohort (a group of subjects) have been put into practice both at home and abroad.

  On the other hand, as the genome sequence of many people is determined by the next-generation sequencer, not only single nucleotide polymorphism (Single Nucleotide Polymorphism, SNP), which has been the main target of analysis, but also several hundred bases or more. It has been found that the phenomenon of relatively long sequence changes frequently is observed. These mutations are called genomic structural variations (SVs), and the main ones are insertions where sequences that are not found in the standard genome sequence are found, or deletions where the sequence that should be is not found (Examples of insertion and deletion are shown in FIG. 5). Structural polymorphisms are thought to play an important role in differences in individual phenotypes and in diseases related to the genome because of the large amount of bases that change at one time. In order to analyze the generation mechanism of structural polymorphisms and the effects of structural polymorphisms on in vivo functions, it is necessary to comprehensively analyze structural polymorphisms.

  Regarding such analysis, for example, Patent Document 1 discloses a technique of mapping a DNA sequence to a genome and comparing the mapped sequences. In addition, Non-Patent Document 1 analyzes the sequence data of a large number of subjects, thereby detecting a mutation that appears commonly in a plurality of subjects even when the amount of sequence data per subject is small. A technique enabling analysis is disclosed.

JP 2005-135053 A

Handsaker et al. 2011, Discovery and genotyping of genome structural polymorphism by sequencing on a population scale, Nature Genetics 2011 Mar; 43 (3): 269-76.

  Since the length that the next-generation sequencer can sequence at a time is limited to about 100 to 400 bases, it is difficult to accurately capture large genomic changes such as structural polymorphisms. On the other hand, although genome sequencing costs have been reduced, genome sequencing in large cohorts is costly, so limited sequencing of gene regions that produce proteins, etc. It is widely practiced to reduce the amount of data per subject when performing sequencing alone or when sequencing the entire genomic sequence. When determining a genome sequence using a next-generation sequencer, it is desirable to obtain a data amount that is 30 times or more that of the sequence to be analyzed. However, there may be a case where only 10 times or less of the data amount can be obtained. Technology to make up for the shortage is needed.

  On the other hand, the technique disclosed in Patent Document 1 is a process for an array in a public database (DB), and does not handle an array of a specific subject group. And it is not a technology that efficiently compensates for the shortage of data. In addition, the technique disclosed in Non-Patent Document 1 detects a mutation that appears commonly in a plurality of subjects even when the amount of sequence data per subject is small by simultaneously analyzing the sequence data of many subjects. Enables structural polymorphism analysis. However, since it is necessary to analyze a large number of subjects, we handle a lot of data from subjects who do not share the same mutation, so we cannot efficiently compensate for the lack of data volume, and false positives are detected when structural polymorphisms are detected. It tends to occur.

  Therefore, a main object of the present invention is to efficiently compensate for a shortage of data amount and increase detection sensitivity even when the amount of data per subject is small.

  A representative analysis apparatus according to the present invention is a DNA sequence analysis apparatus that selects a first subject from a plurality of subjects and outputs a mutation of the first subject, and is obtained from genomes of the plurality of subjects. Storage means for storing mapping data obtained by mapping sequence data to a reference genome sequence in a public database, and pedigree information representing the blood relationship of the plurality of subjects, and a second subject under a predetermined condition based on the pedigree information Family information analysis means for sequentially selecting, and detecting the mutation by simultaneously analyzing the mapping data of the first subject and the second subject, the detected mutation is common to the first subject and the second subject Pair analysis means for performing pair analysis to record as a mutation, mutation determination means for removing duplication from the detected mutation, and outputting as a mutation of the first subject, Characterized by comprising.

  For example, the predetermined condition is the parents of the first subject, and the family information analyzing means selects the parents of the first subject as the second subject.

  Further, the present invention is grasped as a database creation method and system.

  According to the present invention, even when the amount of data per subject is small, the shortage of the amount of data can be efficiently compensated, and the detection sensitivity can be increased.

It is a figure which shows the structural example of a DNA sequence analyzer. It is a figure which shows the outline | summary of the process of DNA sequence analysis. It is a figure which shows the example of a processing sequence. It is a figure which shows the example of family tree information. It is a figure which shows the example of insertion and deletion. It is a figure which shows the example which detects insertion and deletion by a split read system. It is a figure which shows the example of the data amount amplification by a some test subject. It is a figure which shows the example of the inclusion relationship of the variation | mutation which a test subject shares with a parent and child. It is a figure which shows the example of the process flowchart of a pair analysis means. It is a figure which shows the example of the process flowchart of a family tree information analysis means. It is a figure which shows the example of the process flowchart which determines a 2nd test subject from a set. It is a figure which shows the example of the process flowchart of a variation | mutation determination means. It is a figure which shows the example of a pair end. It is a figure which shows the example of a mutation analysis tree. It is a figure which shows the example of the process flowchart which updates the variation | mutation analysis tree and calculates the probability that an undiscovered variation | mutation will be discovered. It is a figure which shows the example of the pair analysis of founders. It is a figure which shows the example of the processing flowchart of the family tree information analysis means using a variation | mutation analysis tree.

  Below, based on next-generation sequencer sequence data and pedigree information, a preferred embodiment for improving the sensitivity of detecting the genomic variation of each subject will be described. In describing the embodiment, first, terms used in the description will be described.

  The human genome has been sequenced by an international collaborative project and the resulting sequence is published as a standard human genome sequence. Many other model organisms have been subjected to genome sequencing. Such a standard genomic sequence is hereinafter referred to as a reference genomic sequence. When compared to this reference genome sequence, the subject's genome has a single nucleotide polymorphism that changes one base at a specific coordinate (Single Nucleotide Polymorphism). There may be a variety of mutations such as microsatellite, structural polymorphism in which long region changes simultaneously In the following, the description will focus on the detection of structural polymorphism, but other types of mutations may be detected.

  By using the next-generation sequencer, it is possible to obtain a large number of sequences at two locations existing at positions separated by several hundred bases from the DNA to be sequenced. These sequences are called paired ends. FIG. 13 is a diagram illustrating an example of a pair end. Two arrows 1301 and 1302 facing each other represent an array at two positions in a paired end. A curve 1303 connecting two arrows 1301 and 1302 indicates which arrow and which arrow are paired. It is calculated where each sequence obtained by the next generation sequencer originates from the genome sequence. This process is a technique commonly used in the art called mapping. For each subject, data in which the results of mapping all given next-generation sequencer sequences are recorded is hereinafter referred to as mapping data.

  Family information is a set of parent-child relationships between subjects, and can be expressed by a set of a plurality of family trees as shown in the example of FIG. The reason for the multiple is that not all subjects are related to blood. A subject whose family information does not have at least one of the two parents in the data is called the founder. The parent-child relationship here is a parent-child relationship in a biological relationship.

  Human cells contain two sets of chromosomes inherited from their parents, and there are two chromosomes in which the same genes are arranged in the same order except for sex chromosomes that differ according to gender. Such two chromosomes are called homologous chromosomes. Each gene on the homologous chromosome is called an allele. If there is a mutation in only one of the alleles, the mutation is said to be heterogeneous. If both alleles have the same mutation, the mutation is said to be homozygous. In addition, if the mutation is caused by a mutation that is not found in either of the parents, the mutation is said to be de novo, but it is not included here because the probability of the de novo mutation occurring is low.

  Hereinafter, typical embodiments will be described as Examples 1 to 3 with reference to the drawings.

  Example 1 will be described below. FIG. 1 is a diagram illustrating a configuration example of a DNA sequence analysis apparatus 100. The DNA sequence analysis apparatus 100 includes a CPU (Central Processing Unit) 101, a main storage device (memory) 102, an auxiliary storage device 103, and a user interface unit 106. The DNA sequence analyzer 100 is connected to an external network via a network 105 such as a LAN (Local Area Network). The main storage device 102 holds various programs executed by the CPU 101 and various data necessary for the CPU 101 to execute these programs. The main storage device 102 is caused to function as at least the mapping data (1) 107-1, the family information (1) 108-1, and the CPU 101 so as to function as the family information analysis means 109, the pair analysis means 110, and the mutation determination means 111. A memory such as a RAM (Random Access Memory) in which a program is stored. These programs collectively become a DNA sequence analysis program.

  The auxiliary storage device 103 is a storage device such as an HDD (Hard Disk Drive) capable of recording mapping data (2) 107-2, family information (2) 108-2, and the like. Further, the auxiliary storage device 103 stores a mutation information DB (1) 112-1 for recording the detected mutation. The removable medium 104 is a storage device that can be attached to and detached from the DNA sequence analyzer 100 and records mapping data (3) 107-3, family information (3) 108-3, mutation information DB (2) 112-2, and the like. Possible recording media such as CD and DVD. The network-connected storage device that stores the mapping data (4) 107-4, the family information (4) 108-4, and the mutation information DB (3) 112-3 may be accessible via the network 105.

  Data stored in the main storage device 102, auxiliary storage device 103, removable media 104, and network-connected storage device, for example, mapping data (1) 107-1, mapping data (2) 107-2, and mapping data (3) 107-3 and mapping data (4) 107-4 may be the same data. These data may be stored as mapping data (1) 107-1 in the main storage device 102 as necessary when reading / writing with the CPU 101, for example, or when the DNA sequence analysis storage device 100 is turned off or in the main memory. When there is no more free space in the device 102, the mapping data (1) 107-1 may be copied from the main storage device 102 to another. Further, before starting the processing, the data is stored in the mapping data (3) 107-3 of the removable medium 104 or the external mapping data (4) 107-4 from another device (not shown), and the DNA sequence analyzer 100 is activated. It may be copied to the main storage device 102 or the auxiliary storage device 103 in the DNA sequence analyzer 100 at the time or at the start of processing. In the description that does not relate to where these data are stored, the mapping data 107, the family information 108, and the mutation information DB 112 are representative.

  The contents of the mapping data 107 are as already described with reference to FIG. 13, and the contents of the family information 108 are as already described with reference to FIG. The contents of the mutation information DB 112 may be contents including information such as chromosome number, mutation start position, mutation end position, mutation type, etc., as shown in FIG. 1, or The 1000 Genomes Project (http: / /www.1000genomes.org) VCF (Variant Call Format) content may be included, or may include content including, for example, the number of detections as statistical information in the analysis. Here, expressions for explanation are used in FIGS. 13, 4, 1, etc., but the mapping data 107, family information 108, and mutation information DB 112 can be processed by the CPU 101 without being limited to such expressions. Any data format may be used. The user interface unit 106 is an input / output device that provides a user interface, such as a keyboard, a mouse, and a display.

  FIG. 2 is a diagram showing an outline of mutation analysis processing by DNA sequence analysis, and FIG. 3 is a diagram showing an example of a processing sequence in which each processing means cooperates. Each process will be described with reference to FIGS. 2 and 3.

(1) Loop for sequentially processing subjects The processing of the DNA sequence analyzer 100 is applied to each subject, and detects a mutation in the genome sequence of those subjects (hereinafter referred to as the first subject). Therefore, as shown in FIG. 2, the entire process is a loop for each subject, and in one process of the loop, the unprocessed first subject is selected, the process described below is executed, and the process result is output. To do.

(2) Family information analysis means 109
As shown in FIG. 3, the family information 108 is stored in advance in the main storage device 102, the auxiliary storage device 103, or the like. In the example of FIG. 3, the CPU 101 is input from the user interface unit 106 and stored in the main storage device 102 or the auxiliary storage device 103, but may be input via the removable medium 104 or the network 105. Based on this pedigree information 108, the pair analysis means 110 selects another one subject (hereinafter referred to as a second subject) as a pair analysis target together with the first subject being processed. Here, if there is no family information (1) 102-1 in the main storage device 102 due to deletion or the like, the family information (2) 108-2 may be read from the auxiliary storage device 103 into the main storage device 102.

(3) Pair analysis means 110
The pair analysis unit 110 is executed for the selected first and second subjects, and a pair analysis described later is executed. That is, the mapping data 107 is read, a mutation is detected, and a mutation common to the selected first and second subjects is output. Here, when there is no mapping data (1) 107-1 in the main storage device 102 due to deletion or the like, the mapping data (2) 107-2 may be read from the auxiliary storage device 103 into the main storage device 102.

(4) Mutation determination means 111
Based on the result of the pair analysis means 110, the variation | mutation of the genome sequence of a 1st test subject is determined. That is, the pair analysis result is read from the auxiliary storage device 103, and duplication of the pair analysis result is removed.

(5) Writing to mutation information DB The variation | mutation of the genome sequence obtained by the above process is stored in mutation information DB112. The contents of the mutation information DB 112 are as already described with reference to FIG. 1, and a relational database, an XML database, or the like may be used as the database format of the mutation information DB 112.

  FIG. 5 is a diagram showing an example of deletion and insertion in the structural variation of the genome sequence. First, the deletion will be described. There are genomic sequences 501a and 502a of the subject, and genomic sequences 501b, 503, and 502b of the reference genome. Here, the array 501a and the array 501b have the same contents, and the array 502a and the array 502b have the same contents. The reference genome sequence 503 drawn in a black rectangle is not in the subject genome sequence. Such a sequence 503 is called a deletion. On the other hand, there are the genome sequences 504a, 506, and 505a of the suspect, the genome sequences 504b and 505b of the reference genome, and the subject genome sequence includes a sequence 506 that does not exist in the reference genome sequence. Such an array 506 is called an insertion.

  Figure 6 shows a method called split read for deletion and insertion (K. Ye, MH Schulz, Q. Long, R. Apweiler, and Z. Ning, Pindel: A pattern growth approach to detect break points of large deletions and medium sized. insertions from paired-end short reads, Bioinformatics, 25 (21): 2865-2871, 2009.) is a diagram showing an example of detection based on the next-generation sequencer sequence.

  First, detection of deletion will be described. When there is a deletion, the next-generation sequencer sequence 602 derived from the position of the boundary of the deletion in the subject's genome is 603 and 604 with the corresponding sequence divided into two locations in the reference genome. These are mapped in the vicinity of 601b corresponding to the next-generation sequencer sequence 601a which is the other paired end of the next-generation sequencer sequence 602. If a plurality of such sequences are found at the same position, it is determined that there is a deletion at that position.

  Next, detection of insertion will be described. When there is an insertion, the next-generation sequencer sequences 606 and 607 derived from the position of the insertion boundary in the subject's genome partially lose the portion corresponding to the inserted sequence in the reference genome. Then, the next-generation sequencer arrays 605a and 608a correspond to the arrays 605b and 608b, and the remaining portions 609 and 610 are mapped in the vicinity of the other arrays 605b and 608b of the respective pair ends. If a plurality of such sequences are found on both sides of a specific position on the genome, it is determined that there is an insertion at that position.

  The split read method is implemented in a well-known software called Pindel (http://gmt.genome.wustl.edu/pindel/current). Compared to other structural polymorphism detection methods, the structural polymorphism that is output has relatively few errors, and the structure polymorphism position can be calculated accurately. Since a plurality of appearing arrays are necessary, the detection sensitivity is insufficient when the data amount is insufficient. When a computer experiment using simulation data was actually performed, in the case of a homomorphic polymorphism held by all homologous chromosomes, the sensitivity was insufficient when the sequence amount was 5 times the genome sequence. There has been a report that sufficient detection sensitivity was obtained at times above (Yasuda T, Suzuki S, Nagasaki M, Miyano S .: ChopSticks: High-resolution analysis of homozygous deletions by exploiting concordant read pairs, BMC Bioinformatics 2012 Oct 30; 13: 279).

  Even if the amount of sequence per subject is 10 times, if the structural polymorphism is heterogeneous (only one of the two homologous chromosomes has the structural polymorphism), for example, as shown in FIG. The reflected array data amount remains equivalent to five times. In genome sequencing of a large group of subjects, the amount of sequence per subject may not reach 10 times in order to secure the number of subjects while reducing costs. In such cases, subjects are processed individually. However, sufficient sensitivity cannot be obtained with the split read method.

  Utilizing the advantage of split read method that the structural polymorphism to be output has few errors and the position is accurate, but to compensate for the detection sensitivity when the amount of data is insufficient, the relative sequence data is used and the same Amplify the sequence derived from the mutation. Parents and children share half of the genome sequence and are likely to have the same mutation. FIG. 7 is a diagram illustrating an example in which subjects having the same mutation are simultaneously analyzed to amplify the data amount. In the example of this figure, both the subject 1 and the subject 2 have the same insertion hetero. When there is a 10-fold sequence amount per subject, the sequence amount derived from this insertion is equivalent to 5 times, so that sufficient detection sensitivity cannot be obtained with the split read method.

  However, if the data of these two subjects are mixed and analyzed simultaneously, the sequence amount derived from this insertion will be equivalent to 10 times, so that sufficient detection sensitivity can be obtained by the split read method. In this way, the process of analyzing the mapping data of the two persons is referred to as pair analysis here. It is also assumed that the sensitivity of pair analysis is very high, while there are few errors. Hereinafter, the subject to be analyzed is referred to as the first subject, and the subject who performs pair analysis with the first subject in order to detect the mutation of the first subject is referred to as the second subject. Note that if simultaneous analysis is performed with a larger number of people such as three or four instead of a pair, the number of false positive mutations increases due to an increase in sequences containing errors.

  FIG. 8 is a diagram schematically showing an example of the inclusion relationship of mutations shared by the first subject with the parent and the child. The variation in the first subject is inherited from the father or mother except for a rare variation called de novo that appears in the subject as a mutation. Therefore, all the mutations of the first subject can be covered by performing pair analysis with the father and the mother, respectively. Even if the mapping data of the parents is not available, there is a possibility that it can be detected by performing pair analysis with the child. If there are multiple children, most of the mutations in the first subject can be covered by pair analysis with each of those children.

  However, there are always mutations 801 that are not shared with children. Such mutations may be shared with unrelated subjects. In FIG. 8, this is the reason why there is an overlap in parental mutation. In a genome sequencing project for a large group of subjects, there are a large number of subjects who are not related, so the number of pairs of subjects who are not related is enormous. Therefore, it is necessary to select only pairs that are necessary and sufficient for mutation detection and perform pair analysis. For this reason, as shown in an example in FIG. 16, a pair of subjects having no blood relationship is limited to a pair analysis 1601 between founders. If the first subject is other than the founder, all mutations are shared with the parents, so pair analysis with a second subject who is not related is unnecessary. When the first subject is the founder, the mutation of the second subject who is not the founder is a mutation inherited from any of the founders of the family of the second subject. Therefore, the pair analysis 1602 with the second subject who is not the founder is unnecessary.

In the example of FIG. 16, all the pairs of subjects are ₁₃ C ₂ = 78, but if limited to the parent-child and founder pairs, pair analysis may be performed for ₇ C ₂ + 2 × 6 = 33 pairs. In general, if the number of founders is f and the number of subjects with parents is m, the number of pairs to be subjected to pair analysis is _f C ₂ + 2m.

FIG. 9 is a diagram illustrating an example of a processing flowchart of the pair analysis unit 110. Details of the pair analysis processing will be described with reference to FIG.
S901: It is determined whether a pair analysis has already been performed for a pair of a first subject and a second subject given as input (hereinafter referred to as an input pair).
S902: If the pair analysis of the input pair has already been executed in the determination of S901, the previous analysis result is copied and the process ends.
S903: The mapping data 107 of two input pairs is acquired.
S904: Merging is performed by concatenating all the mapping data 107 acquired in S903.
S905: The merged mapping result is used as an input of Pindel as if it were one person data, and a mutation is detected.
S906: The obtained mutation is expected to contain a mutation common to the first subject and the second subject, but the sequence amount is specified on the genome even though only the second subject possesses it. It may also be detected when many are obtained in the region or when the structural variation of the second subject is homozygous. In order to remove such a structural variation that does not exist in the first subject, a DNA sequence that supports the obtained structural variation as shown in FIG. 6 is obtained. Of these sequences, the mutation is output as the mutation of the first subject only when the number of sequences derived from the first subject exceeds a preset threshold value k. If the threshold is not exceeded, the found mutation is discarded.
S907: The mutation output in S906 is made available to the mutation determination unit 111, and when pair analysis is requested from the next time, it is saved so that it can be copied as the previous analysis result in S902.

FIG. 10 is a diagram illustrating an example of a processing flowchart of the family information analyzing unit 109. By this process, the second subject to be paired with the first subject is determined.
S1001: It is determined whether the parent of the first subject is included in the mapping data 107 and the family information 108. If it is included, the process proceeds to S1002, and if not, the process proceeds to S1005.
S1002: It is determined that pair analysis is performed between the first subject and the first parent of the first subject determined to be included in S1001. That is, the first parent is determined as the second subject.
S1003: It is determined whether another one of the parents of the first subject is included in the mapping data 107 and the family information 108. If it is included, the process proceeds to S1004. Otherwise, the process proceeds to S1005.
S1004: It is determined that pair analysis is performed between the first subject and the second parent of the first subject determined to be included in S1003. That is, this second parent is determined as the second subject.
S1005: It is determined that the first subject is x, the set of children of the first subject is S, and the processing shown in FIG. 11 is performed to perform pair analysis with all the children. That is, all children are determined as second subjects. The process of FIG. 11 will be described later.
S1006: It is determined that the first subject is x, the set of all the founders is S, the processing shown in FIG. 11 is performed, and the pair analysis with all the founders is performed. That is, all the founders are determined as second subjects.

FIG. 11 is a diagram illustrating an example of a processing flowchart for determining a second subject from the set S. A set of a plurality of subjects is set in the set S in advance. The variable x is preset with the first subject. Since this process is used in S1005 and S1006, all the children and all the founders are determined as second subjects, and is a part of the process of the family information analyzing unit 109.
S1101: When the first subject x is included in the set S, the first subject x is excluded from the set S.
S1102: If the set S is an empty set, the process ends.
S1103: Only one new second subject is selected from set S, set to variable y, and excluded from set S.
S1104: It is determined that the pair analysis of the first subject and the second subject set in the variables x and y is performed.

  When there are a plurality of second subjects determined by the processing described with reference to FIGS. 10 and 11, the plurality of second subjects are output as a set of second subjects from the family information analyzing unit 109 to the pair analyzing unit 110. The analysis unit 110 may extract a pair from the second subject group to form a pair with the first subject. Although different from the description using FIGS. 2 and 3, the pair analysis described using FIG. 9 may be immediately performed in each of S1002 and S1004 of FIG. 10 and S1104 of FIG. 11.

FIG. 12 shows the mutation determination for determining the mutation of the genome possessed by the first subject based on the mutation data obtained by the pair analysis between the first subject and each of the second subjects determined by the family information analysis means 109. It is a figure which shows the example of the process flowchart of the means.
S1201: All structural mutation data obtained by pair analysis are linked. At this stage, there are many identical structural mutations that are included in duplicate results of pair analysis with different second subjects.
S1202: Sort the linked structural variation data based on the position on the genome.
S1203: The sorted structural variation data is scanned, the difference in position on the genome is less than the integer w which is a parameter given in advance (within w bases), and the mutation such as insertion or deletion Merge (merge) by considering the same type as the same mutation.
S1204: The result without the duplication of mutation obtained by the above processing is output to the mutation information DB 112.

  As explained above, even if the amount of data per subject is small, by analyzing with subjects who are likely to share the same mutations as parents and children, the lack of data amount is compensated for, and the detection accuracy Can be increased.

  Next, a description will be given of a second embodiment that can cope with a case where detection sensitivity of pair analysis is low or an error occurs with a certain probability, with reference to the drawings. In the description of Example 1, first, pair analysis was performed with a parent, pair analysis was performed with a child as necessary, and pair analysis was further performed with a pair between founders. This is because, assuming that the sensitivity of pair analysis (hereinafter referred to as r) is sufficiently high, it can be assumed that mutations that include subjects who share mutations in the data can be detected using these combinations. It is. However, when the detection sensitivity r is low, there may be a mutation that may be overlooked if only the pair analysis of Example 1 is performed. In order to detect such mutations, not only the parents and children but also the relatives such as grandparents, grandchildren, and siblings can expand the pair analysis target, increase the number of times each mutation is paired, and increase the probability of detection. It is necessary to raise. Here, when selecting a relative, it is important to select a subject who shares the most undiscovered mutation as the second subject.

  On the other hand, let p be the probability that the detected mutation actually exists in the subject's genome. In Example 1, p was sufficiently high and close to 1 by setting the second degree as a second subject. However, when p is low, a mutation detected only once has a high possibility of error. It is desirable to repeat the pair analysis until it is detected multiple times. Specifically, in the following processing, it is preferable that N is an integer given as a parameter, and is processed as a mutation of the first subject only when N is detected.

To select subjects who share the most undiscovered mutations, the probability of having an undiscovered mutation is calculated for each subject. This probability relates to the closeness of the subject and the first subject, the degree of overlap with the already analyzed mutation, and the number of times the pair analysis has been performed. Specifically, when a subject y is used as a second subject, and the first subject is x, the probability P (y) that a mutation inherited from a common ancestor by the first subject and the second subject is newly found. Is expressed by the following equation.
P (y) = Σ _v∈V 2F (x, y) s (v) r (1-r) ^{n (v)}
Where F (x, y) is the relative coefficient of the first subject x and the second subject y (when alleles of x and y are selected, the probability that they were inherited from the same ancestor. Sequence of the same gene in each of the homologous chromosomes). The closeness coefficient may be included in the family information 108. V is the set of subsets of mutations in x, s (v) is the percentage of mutations in v in the total mutations in x, and n (v) is the number of pair analyzes performed so far on the mutations in v It is. An arbitrary subset v belonging to V is a set of mutations to which pair analysis with other second subjects is applied in the same way before the pair analysis of x and y is performed. In order to calculate P (y), it is necessary to find V and to know how many times the pair analysis has been applied to any v belonging to V. If pair analysis is performed many times, V becomes complicated and it becomes difficult to calculate V appropriately. Therefore, a process for calculating V will be described with reference to the drawings.

  FIG. 14 is a diagram showing an example of a mutation analysis tree used for calculating V. The mutation analysis tree is a tree structure rooted at a node 1401 representing the entire mutation of the first subject. Each time the pair analysis with the second subject is performed, each node is divided into a node 1402 included in the second subject and a node 1403 not included. A node with no child nodes is called a leaf. Each node L has a label h (whether or not possessed if there is no negative sign “¬”) that indicates whether or not the mutation is possessed by the second subject who has performed the pair analysis. A triplet is recorded, consisting of the number n of pair analyzes performed for the number, and the ratio s of the variation of the node in the total variation of the first subject. When L has a triplet (h, n, s), it is expressed as L = (h, n, s) below. The set of mutations of the first subject corresponding to each leaf in the mutation analysis tree is each element of V.

FIG. 15 is a diagram showing an example of a processing flowchart for updating the mutation analysis tree executed every time pair analysis is performed.
S1501: Variable X is initialized to zero. This variable X is used to store the progress of calculating P (x).
S1502: The closeness coefficient of the first subject and the second subject is set in the variable F.
S1503: The processing of S1504 to S1507 is repeated for all the leaves existing in the mutation analysis tree before update. This iteration is an iteration for performing processing on all subsets of mutations in the first subject.
S1504: Let L be one of the unprocessed leaves. Let the triple recorded in this leaf be (h, n, s).
S1505: Two new leaves L1 and L2 are added as children of L. The triples recorded in them are as shown in FIG.
S1507: The probability that a mutation corresponding to L is detected is added to X.
S1508: The value of X is output as P (y).

As the family information analysis means 109, the second subject is selected using P (y) calculated by the processing example shown in FIG. 15 instead of the processing example shown in FIG. 10 described in the first embodiment. The pair analysis is performed on the selected second subject. FIG. 17 shows an example of a flowchart of this process. The contents of the processing will be described with reference to the drawings. Here, as input, a first subject, a set S of past second subjects that have been paired with the first subject, and a probability B having the same mutation between subjects who are not related to each other are given.
S1701: The variable P _max is initialized to zero.
S1702: A set of subjects who are parents and children of subjects in set S and do not belong to set S is set in S ′.
S1703: It is determined whether the set S ′ is an empty set. If it is an empty set, the process proceeds to S1708. If it is not an empty set, the processing of S1704 to S1707 is repeated for each element of the set S ′.
S1704: One element y is extracted from the set S ′, and the element y is deleted from the set S ′.
S1705: P (y) is calculated by the process described using FIG. 15, but the mutation analysis tree in S1505 is not changed.
S1706: If P _max is not less than P (y), the process returns to S1703 to process the next subject.
S1707: P _max If there is less than P (y), set P a (y) as a new value of P _max, is set to y variable y _max.

Returning to S1703, S1704 to S1707 are repeated.
S1708: If P _max is smaller than the probability B, that is, the probability that subjects who are not related have a common mutation, it is more likely that a new mutation will be found by performing pair analysis with a non-related person than by a related person. For this reason, when it is smaller than the probability B, the process proceeds to S1709, and when it is not smaller than the probability B, the process proceeds to S1711.
S1709: It is determined whether or not there is an unprocessed founder. If there is an unprocessed founder, the process proceeds to S1710, and if not, the process proceeds to S1711.
S1710: One unprocessed founder is arbitrarily selected as a second subject.
S1711: Let y _max be the second subject.
S1712: Pair analysis is performed on the first subject and the second subject. In preparation for selecting the next second subject, the mutation analysis tree is updated by the process shown in FIG.

  As explained above, even if the amount of data per subject is small and the detection sensitivity is not sufficiently obtained even when analyzing subjects such as parents and children, the probability that the shortage of data amount can be compensated by the minimum number of people. And detection accuracy can be increased.

  In the processing described in Examples 1 and 2, the threshold k for the number of next-generation sequencers in pair analysis, the threshold w for the position of the mutation considered to be the same used by the mutation determining unit 111, and sharing the mutation among unrelated subjects Use a parameter such as probability B An example of value adjustment and estimation processing when the values of these parameters cannot be obtained in advance will be described.

(1) Threshold number of next-generation sequencer sequences in pair analysis k
Using the data for which the next-generation sequencer sequence and the position information of the mutation are publicly disclosed, that is, the data whose detection result is known, the value of k that provides the highest detection accuracy is selected. Instead of public information, the value of k may be adjusted to match the mutation information determined by processing other than the next-generation sequencer such as the Sanger method.

(2) Threshold value w of the mutation position considered to be the same
Based on the result of analyzing the published data as in (1) above, the fluctuation range of the position of the detected mutation is obtained, and the correctly detected mutation position is included in the published data. Adjust the value of.

(3) Probability of sharing mutations among unrelated subjects B
Pair analysis is performed between a few founders. For the mutation detected by the pair analysis, the first number of times that the next-generation sequencer sequence supporting the mutation exceeds the threshold value k in the mapping data of each subject is calculated. On the other hand, a pair analysis of a small number of parents and children is also performed, and a second number of mutations obtained by this pair analysis is calculated. In the pair analysis of one set of parent and child, about half of the mutations in each subject are detected. Therefore, the value of B can be estimated by dividing the first number by the value twice the second number.

  As described above, each value can be adjusted and estimated even if each value of the parameter cannot be obtained in advance. And even if the amount of data per subject is small, the shortage of the amount of data can be compensated and the detection accuracy can be increased.

  Each example has been described above, but the above example shows a preferable example, and is not intended to limit the specific configuration of each example. Various modifications can be made without departing from the scope of the above description.

100 DNA sequence analyzer 101 CPU
102 Main storage device 103 Auxiliary storage device 104 Removable media 105 Network 106 Interface unit 107 Mapping data 108 Family information 109 Family information analysis means 110 Pair analysis means 111 Mutation determination means 112 Mutation information DB

Claims (15)

A DNA sequence analyzer for selecting a first subject from a plurality of subjects and outputting a mutation of the first subject,
Storage means for storing mapping data obtained by mapping sequence data obtained from a plurality of subject genomes to a reference genome sequence in a public database, and pedigree information representing the relative relationships of the plurality of subjects;
Based on the family information, family information analysis means for sequentially selecting a second subject of a predetermined condition,
A pair for performing pair analysis for simultaneously detecting mapping data of the first subject and the second subject to detect a mutation, and recording the detected mutation as a common mutation in the first subject and the second subject Analysis means;
Mutation determination means for removing duplication from the detected mutation and outputting as the mutation of the first subject,
An analysis device characterized by comprising: The predetermined condition is the parents of the first subject,
The analysis apparatus according to claim 1, wherein the family information analyzing unit selects the parents of the first subject as the second subject. The predetermined condition is the founder who is the child of the first subject and any subject who does not have at least one parent in the family information,
The analysis apparatus according to claim 1, wherein the family information analyzing unit selects the child of the first subject and the founder as the second subject. The predetermined condition is the founder who is one parent of the first subject, the child of the first subject, and any subject who does not have at least one parent in the family information,
The pedigree information analyzing means analyzing apparatus according to claim 1, characterized in that selecting the initiator and one of the parents and children of the first test subjects as the second subject. The predetermined condition is a subject having the largest probability P (y),
The pedigree information analysis means for the first subject x and each second subject candidate y,
F (x, y) is an affinity coefficient between the first subject x and the second subject candidate y,
V is a set of subsets obtained by dividing the set of mutations of the first subject, and each subset consists only of mutations inherited by the same subject.
Let n (v) be the number of times that the subset of mutations of the first subject with v∈V has been subject to pair analysis for the first subject,
Let s (v) be the proportion of the mutation of v in the total mutation of the first subject x,
When r is the mutation detection probability when mutation analysis is performed using mapping data of two subjects,
_{Evaluate the} probability that a new mutation will be detected by P (y) = Σ _v∈V 2F (x, y) s (v) r (1-r) ^{n (v)}
2. The analysis apparatus according to claim 1, wherein a subject having the largest probability P (y) is selected as a second subject. Storing mapping data obtained by mapping sequence data obtained from a plurality of subject genomes to a reference genome sequence in a public database;
Storing pedigree information representing the blood relationship of the plurality of subjects;
And selecting whether we first subject among the plurality of subjects,
A step of sequentially selecting a second subject under a predetermined condition based on the family tree information;
A step of performing pair analysis for simultaneously detecting mapping data of the first subject and the second subject to detect a mutation, and recording the detected mutation as a mutation common to the first subject and the second subject. When,
Removing duplicates from the detected mutation and creating a database as the mutation of the first subject;
A database creation method characterized by comprising: The predetermined condition is the parents of the first subject,
The database creation method according to claim 6, wherein in the step of sequentially selecting the second subject, the parents of the first subject are selected as the second subject. The predetermined condition is the founder who is the child of the first subject and any subject who does not have at least one parent in the family information,
The database creation method according to claim 6, wherein in the step of sequentially selecting the second subject, the child of the first subject and the founder are selected as the second subject. The predetermined condition is the founder who is one parent of the first subject, the child of the first subject, and any subject who does not have at least one parent in the family information,
Wherein the step of sequentially selecting the second subject, a database creation method according to claim 6, characterized in that selecting the initiator and one of the parents and children of the first test subjects as the second subject. The predetermined condition is a subject having the largest probability P (y),
The step of sequentially selecting the second subject includes
For the first subject x and the candidate y for each second subject,
F (x, y) is an affinity coefficient between the first subject x and the second subject candidate y,
V is a set of subsets obtained by dividing the set of mutations of the first subject, and each subset consists only of mutations inherited by the same subject.
Let n (v) be the number of times that the subset of mutations of the first subject with v∈V has been subject to pair analysis for the first subject,
Let s (v) be the proportion of the mutation of v in the total mutation of the first subject x,
When r is the mutation detection probability when mutation analysis is performed using mapping data of two subjects,
_{Evaluate the} probability that a new mutation will be detected by P (y) = Σ _v∈V 2F (x, y) s (v) r (1-r) ^{n (v)}
The database creation method according to claim 6, wherein a subject having the largest probability P (y) is selected as a second subject. Determining a threshold k to be a common mutation based on the position and DNA sequence of known mutations published;
7. The database creation method according to claim 6, wherein the step of executing the pair analysis records a common mutation when the number of the detected mutation sequences exceeds the threshold value k. Determining a threshold k to be a common mutation based on the Sanger process;
7. The database creation method according to claim 6, wherein the step of performing pair analysis records a common mutation when the number of detected mutation sequences exceeds the threshold value k. A system including a computer that selects a first subject from a plurality of subjects and outputs a variation of the first subject,
Storage device for storing mapping data obtained by mapping sequence data obtained from a plurality of subject genomes to a reference genome sequence in a public database, pedigree information representing the relative relationships of the plurality of subjects, and a database output by the computer Further including
The calculator is
Storage means for inputting and storing the mapping data and the family information from the storage device;
Based on the family information, family information analysis means for sequentially selecting a second subject of a predetermined condition,
A pair for performing pair analysis for simultaneously detecting mapping data of the first subject and the second subject to detect a mutation, and recording the detected mutation as a common mutation in the first subject and the second subject Analysis means;
Mutation determination means for removing duplication from the detected mutation and outputting the mutation of the first subject as the database;
A system characterized by comprising: The system according to claim 13, wherein the storage device and the computer are connected via a network. The system according to claim 13, wherein the storage device is a removable medium, and the storage device is attached to the computer.

Images