TW201619456A - Method for detecting chromosomal structural abnormalities and device therefor - Google Patents

Abstract

Disclosed are a method for detecting chromosomal structural abnormalities and a device therefor, the method comprising: obtaining the whole genome sequencing result of a target individual, i.e. multiple pairs of read pairs positioned at both ends of the chromosome fragment to be measured; aligning the sequencing result with the reference sequence to obtain an abnormal matching set, including two read sequences respectively matched to read pairs of different chromosomes of the reference sequence; clustering the read sequences in the abnormal matching set into clusters according to the locations matched; and filtering the clusters obtained by clustering using pre-set requirements, such as one related to the degree of compactness, to obtain a filtered result cluster for judging the occurrence of chromosome translocation structural abnormalities.

Description

Method and device for detecting abnormality of chromosome structure

The invention relates to the technical field of genomics and bioinformatics, in particular to a method and a device for detecting abnormalities of chromosome structure.

The current common methods of chromosome examination are: karyotype analysis: for example, G-band karyotype analysis, because the distribution of 400 to 600 BAND is used to judge the abnormality of chromosome structure, it is usually only possible to detect abnormalities at the chromosome level. Deletions and repeats above 5 Mbp were detected and were incapable of detection of smaller fragments (<5 M). Moreover, this method requires the cultivation of living cells, requiring the cells to remain active.

Fluorescence in situ hybridization (FISH) method: deletion, repetitive and balanced translocation of smaller fragments can be detected, but the detected chromosome fragments need to be determined in advance to prepare corresponding probes, thus being probed Design restrictions. Because FISH cannot detect unknown areas, it is often used to verify test results.

Microarray method: It includes two probe methods, one based on single nucleotide polymorphism (SNP) design, one based on CNV design, and thus has similar limitations to FISH.

With the continuous development of genome-wide sequencing technology, the cost of sequencing is decreasing, making the popularization of whole genome sequencing possible. It is necessary to study the results based on whole genome sequencing. A means of discovering abnormalities in chromosome structure.

According to an aspect of the present invention, a method for detecting abnormality of a chromosome structure includes the steps of: obtaining a whole genome sequencing result of a target individual, wherein a plurality of pairs of read length pairs are included, and each pair of read length pairs is composed of two read length sequences, respectively Located at both ends of the measured chromosome fragment, each pair of read pairs are from the positive and negative strands of the corresponding chromosome fragment, respectively, or each pair of read pairs are from the positive or negative strand of the corresponding chromosome fragment; sequencing results and references The sequences are aligned to obtain an abnormal matching set, and the abnormal matching set includes a first type of read length pair that meets the following description, and the two read length sequences of the first type of read long pair are respectively matched to different chromosomes of the reference sequence; The obtained position clusters the read long sequences in the abnormal matching set into clusters, each cluster contains a single-ended read length sequence from a set of read long pairs, and the corresponding read length sequence at the other end is located in another cluster; The cluster obtained by the class is filtered, which includes, calculating the degree of compactness of each cluster, filtering out clusters whose degree of compaction does not satisfy the preset requirement R-va and clusters paired therewith, The filtered result cluster containing the first type of read length pair is used to determine the occurrence of chromosomal translocation structural abnormalities.

According to another aspect of the present invention, an apparatus for detecting an abnormality of a chromosome structure includes: a data input unit for inputting data; a data output unit for outputting data; and a storage unit for storing data including an executable program The processor is coupled to the data input unit, the data output unit, and the storage unit for executing an executable program stored in the storage unit, and the executing of the program includes performing the foregoing method for detecting an abnormality of the chromosome structure.

According to still another aspect of the present invention, a computer readable storage medium for storing a program for execution by a computer is provided, and those skilled in the art can understand that when the program is executed, the above-mentioned detection of abnormality of a chromosome structure can be completed by instructing related hardware. All or part of the method Step by step. The storage medium may include: a read only memory, a random access memory, a magnetic disk or an optical disk, and the like.

The method according to the present invention obtains a pair of read lengths matched to different chromosomes by alignment of the whole genome sequencing result with the reference sequence, thereby enabling screening of chromosomal translocation structural abnormalities, and further improving the obtained results by clustering and filtering. Sex and reliability make it possible to obtain analytically meaningful results.

The above and/or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments of the accompanying drawings in which: FIG. FIG. 2 is a schematic diagram of a first type of Reads for abnormal matching according to an embodiment of the present invention; FIG. 3 is a schematic diagram of a second type of Reads for abnormal matching according to an embodiment of the present invention; FIG. 4 is a schematic diagram of a second type of Reads according to an embodiment of the present invention; FIG. 5 is a schematic diagram of a pair of clusters of abnormal chromosomes according to an embodiment of the present invention; FIG. 6 is a schematic diagram of Experimental Example 1 according to an embodiment of the present invention. RPK schematic diagram of FA"; FIG. 7 is a "SON" of Experimental Example 1 according to an embodiment of the present invention. RPK schematic.

According to an embodiment of the present invention, a method for detecting an abnormality in a chromosome structure is provided, comprising the steps of:

Step 1. Obtain the whole genome sequencing results of the target individual.

The sequencing results include paired read length pairs (also called "reads") Reads, each pair of Reads consisting of two read length sequences located at the ends of the measured chromosome fragments, each pair of Reads from the positive strand of the corresponding chromosome fragment And the negative strand, or, each pair of Reads comes from both the positive or negative strand of the corresponding chromosome fragment.

The measured chromosome fragment is usually obtained by interrupting the chromosome sample from the target individual, and the corresponding library preparation is performed according to the selected sequencing method. The optional sequencing method is based on the sequencing platform from but not limited to CG. (Complete Genomics), Illumina/Solexa, ABI/SOLiD, and Roche 454, preparation of single-ended or double-ended sequencing libraries based on the selected sequencing platform. According to a specific embodiment of the present invention, double-end sequencing can be performed, and the two read-length sequences Read1 and Read2 in each pair of Reads obtained are respectively derived from the positive-strand Sp and the negative-strand Sm of the corresponding chromosome segment, as shown in FIG. 1; The length L-r1 and the length L-r2 of Read2 may be the same or different. Of course, if a single-read sequencing method is used to completely obtain the sequence of the entire chromosome segment, it is also feasible to intercept a sequence of an appropriate length from both ends of the completely obtained sequence to form a pair of Reads. The two read length sequences Read in each pair of Reads are simultaneously from the positive or negative strand of the corresponding chromosome fragment. This embodiment does not limit the specific sequencing method selected.

In the present invention, the size of the library used for sequencing is referred to as L-lib, and a library in which L-lib is 100 to 1000 bp is generally referred to as a small fragment library, and L-lib is 2K, 5K-6K, 10K, 20K, A 40 Kbp library is called a large fragment library. The present invention does not require the size of L-lib, but in general, a library of a larger length is beneficial for obtaining an effective result while ensuring the quality of library construction. Therefore, Llib is preferred. 300bp. Large fragments, such as 5 Kbp libraries, or small fragments, such as a 500 bp library, can generally be used. In order to make the sequencing results have better abundance, the sequencing depth of the large fragment library can be selected to be greater than 2 times, and the sequencing depth of the small fragment library can be selected to be greater than 5 times. To avoid data waste, the sequencing depth of the large fragment library is preferably 2 Multiplication, the sequencing depth of the small fragment library is preferably 5 times. It should be noted that since the specific data involved in the present invention is mostly statistically significant, any numerical value expressed in an accurate manner represents a range, that is, a range including plus or minus 10% of the numerical value, unless otherwise specified. Repeat the instructions.

L-r1 and L-r2 are preferably 25 bp or more, because if it is less than 25 bp, the unique aligning ratio is lowered, and the complexity of subsequent obtaining alignment results is increased. L-r1 and L-r2 also need not be too large to avoid wasting data, so it is preferably 50 bp. L-r1 and L-r2 have no maximum value and can be changed according to the development of sequencing technology. For example, according to current sequencing technology, L-r1 and L-r2 generally do not exceed 150 bp.

Step 2. Align the sequencing results with the reference sequence.

The reference sequence used is a known sequence and may be any reference template in the biological category to which the target individual belongs in advance. For example, if the target individual is a human, the reference sequence may select HG19 provided by the National Center for Biotechnology Information (NCBI). Further, the resource library including more reference sequences may be pre-configured, and the sex, race, and region of the target individual may be determined according to the target individual before the sequence comparison. Select a closer reference sequence to help obtain more accurate results. In the alignment process, according to the setting of the alignment parameter, a pair of Reads is allowed to have at most n base mismatches, n is preferably 1 or 2. If there are more than n bases in the Reads, then It is considered that the pair of Reads cannot be compared to the reference sequence, or if the mismatched n bases are all located in a Read in Reads, then the Read in the Reads cannot be compared to the reference sequence. For the specific comparison, various comparison softwares, such as SOAP (Short Oligonucleotide Analysis Package), bwa, samtools, etc., may be used, which is not limited in this embodiment.

According to the comparison of Reads, the following classifications can be obtained:

(1) Normal match set *.pair: This includes Reads that conform to the description below, that is, the two read length sequences Read1 and Read2 in Reads match the same chromosome of the reference sequence, and the positive and negative links of the matched positions Consistent with the positive and negative chain relationship in Reads, and the length of the chromosome segment L-pr calculated from the matched position is less than the preset threshold V-lib. V-lib is preferably 5% x L-lib~15% x L-lib, further preferably 10% x L-lib. The above thresholds are set empirically according to the standard deviation of the library size. According to experience, the standard deviation of small fragment libraries is about 15 bp, and the standard deviation of large fragment libraries is about 50 bp. It can be considered that the deviation of L-pr and L-lib is within the range of 3 standard deviations, for example, for 500 bp. The library can be considered to have a suitable range of 455 bp to 545 bp for L-pr.

Based on *.pair, the number of Reads can be obtained according to the number of positions matched. For example, the number of Reads included in the unit length can be counted. The unit length can be set according to L-lib, for example, it can be set to 1.5~4 times L. -lib. If L-lib is 500 bp, the unit length can be set to 1 Kbp, and the RPU can be recorded as RPK. According to the change of the RPU relative to the average value, for example, whether the change exceeds the preset threshold value V-rm, it can be used to assist in judging the occurrence of structural anomalies and increase The accuracy of the results analysis. Preferably, V-rm is from 10 to 30%, further preferably 20%. In addition, the average value of the RPU can be obtained by statistics or by estimation. For example, the average value of the RPU can be estimated by the following method: sequencing depth × (unit length / L-lib). If you do not need to use RPU, you do not need to get *.pair.

(2) Abnormal matching set *.sin: This includes three types of Reads that meet the following description.

The two read length sequences in the first type of Reads are matched to different chromosomes of the reference sequence, respectively; such Reads are associated with translocation structural anomalies, such as balanced translocations and unbalanced translocations. As shown in Figure 2, it shows a case of balanced translocation. Read1 in a pair of Reads matches the chromosome chra, and Read2 matches the chromosome chrb, while the other pair of Reads is the opposite. The dotted line connecting Read1 and Read2 in the figure Indicates their positional relationship in the chromosome segment (the same below), pa and pb respectively indicate the position of the possible breakpoint, and the so-called "breakpoint" refers to the boundary point of the structural abnormality of the chromosome.

The two read length sequences in the second type of Reads match the same chromosome of the reference sequence, but L-pr is negative; such Reads are associated with repetitive structural anomalies in tandem. As shown in Figure 3, both Read1 and Read2 in a pair of Reads match the chromosome chra, but the head-to-tail position relationship of the matched position is opposite to the head-to-tail position relationship of Read1 and Read2 in the chromosome segment, respectively, pa1 and pa2 indicate possible existence. The starting and ending position of the repeated segment, L-sv indicates the length of the repeated segment, and the dotted line in the middle of the chra indicates the length of the omission (the same below)

The two read length sequences in the third type of Reads match the same chromosome of the reference sequence, but L-pr is greater than L-lib and the deviation exceeds the preset threshold V-lib; such Reads are associated with missing structural anomalies. As shown in Figure 4, both Read1 and Read2 in a pair of Reads match the chromosome chra. And the matching position of the head and tail position is the same as the head and tail position relationship of Read1 and Read2 in the chromosome segment, but the distance exceeds the suitable range, pa1 and pa2 respectively indicate the starting and ending positions of the missing fragments, and L-sv indicates the missing fragments. length.

Since different types of Reads in the abnormal matching set respectively represent different types of chromosome structural abnormalities that may occur, it is not necessary to obtain all kinds of abnormal matching Reads according to the detection requirements, for example, if only the translocation result abnormality needs to be detected, You can get the first type of Reads only from the comparison results. Similarly, the exception matching set is not limited to including the above three types of Reads, as long as it does not belong to the normal matching set, but can match the read sequence of Reads or Reads in the reference sequence, can be counted into the abnormal matching set. One of ordinary skill in the art can associate different types of abnormally matched expressions with corresponding chromosomal structural anomalies that may occur. In addition, considering the influence of interference such as noise that may exist, the case of distinguishing positive or negative chain matching or mismatch may not be considered in the abnormal matching set.

(3) Unable to match set *.unmap: This includes Read that cannot be matched to the reference sequence. These Reads can be paired (both cannot match) or single-ended (the other Read can match).

The single-ended Read that exists in *.unmap can be used to further breakpoint assembly after obtaining the result cluster to obtain a more accurate breakpoint range. If you do not need to breakpoint assembly, you don't have to get *.unmap.

Step 3. Cluster the read length sequences in *.sin into clusters according to the matched locations.

The clustering can adopt various clustering algorithms, which is not limited in this embodiment. For example, a simple approach is to divide the cluster according to the set minimum inter-cluster distance V-cl, ie, search Read length sequence Read sorted by position, starting from the first Read, if the second Read is less than V-cl, it is divided into the same cluster, and continues to search from the second Read until The distance between the nth read and the n-1th Read is greater than V-cl, and the second cluster is divided from the nth Read, and the foregoing process is performed cyclically until all Reads are traversed. When clustering, it is not necessary to consider the case of positive and negative chains separately, and clustering according to the position of Read matching on the chromosome.

Each cluster after clustering contains a single-ended read length sequence from a set of Reads, and the corresponding read length sequence at the other end is located in another cluster, so these two clusters can be referred to as a pair of clusters. As shown in FIG. 5, which are schematic diagrams of a pair of clusters cluster1 and cluster2 located on different chromosomes, of course, the paired clusters may also be located on the same chromosome. In order to make the analysis after clustering meaningful, each cluster preferably contains more than two Reads. If a single Read is more than V-cl from both before and after Read, the abnormal data can be discarded.

V-cl is not lower than L-lib. If it is set too low, it will make the candidate cluster too much, and the number of Reads in the cluster is too small, which is not convenient for later screening and filtering, and may also lead to an increase in false positive results. If it is too high, it may be inconvenient to determine the breakpoint, and the range of the breakpoint is increased. Therefore, it may preferably be 10 Kbp. Depending on the clustering algorithm used, V-cl can have different specific meanings, such as the distance between the centers of gravity of two adjacent clusters, or the closest two clusters. The distance between the two Reads and so on.

Step 4. Filter the clusters obtained by clustering.

Filtration is to remove as much as possible of possible interferences, such as sample contamination, sequencing errors, alignment errors, noise, etc., so that the results can reflect real chromosome structural anomalies as much as possible, so it can be set according to actual needs and possible types of interference. Filtering conditions, the present embodiment preferably provides the following filtering mode, and in practical applications, it can be used jointly or separately. One or several filtering methods:

(1) According to the degree of compactness of the cluster: Calculate the degree of compactness of each cluster, and filter out the clusters whose degree of compaction does not satisfy the preset requirement R-va and the clusters paired with them. Various available mathematical methods can be used to calculate the degree of compactness of each cluster. For example, the degree of compaction can be expressed by the variance, and the variance of the position of each Read in the cluster and the center or center of gravity of the cluster can be calculated. The smaller the variance, the more compact the degree high. Preferably, when calculating the degree of compactness of each cluster, the length of the read length sequence in the range of 5% to 25% of the length of both ends of the cluster may be discarded, preferably 20%, to reduce the influence of the peripheral data on the calculation result. Preferably, R-va may be set to a fixed threshold, for example, the required variance is lower than a fixed threshold, or set to a elimination ratio, for example, the ranking of the required variance in all clusters is within a preset minimum interval, for example, R-va is set to The ranking of the variance in all clusters is in the lowest range of 2% to 10%, preferably 5%.

The degree of compactness of the cluster reflects the stability of the Read distribution, indicating whether Read is concentrated in a small interval. In general, the actual structural variation will be submerged in numerous "environmental noise", but "environmental noise." "The effect on the whole genome is basically uniform, so there is a tendency to show a basic average distribution in the whole sequence (of course, it may also be affected by, for example, GC (guanine Guanine and cytosine Cytosine) content), but in reality Where the structural variation occurs, the Read in the cluster usually exhibits a trend similar to a normal distribution, so the degree of compactness, such as variance, can well reflect the differences between clusters.

(B) According to the linear correlation of the paired clusters: Calculate the linear correlation of the two pairs of pairs, and filter out the paired clusters whose linear correlation does not satisfy the preset requirement R-li. Various available mathematical methods can be used to calculate the linear correlation of a pair of clusters, such as calculating the correlation coefficients of two clusters, and the higher the correlation coefficient, the higher the linear correlation. Preferably, R-li can be set to a fixed threshold, for example If the correlation coefficient is required to be higher than the fixed threshold, or set to the elimination ratio, for example, the ranking of the correlation coefficient in all clusters is required to be within the preset maximum interval. For example, R-li is set to rank the correlation coefficient in all clusters at 2 Within the highest range of %~10%, preferably 5%.

The linear correlation pays more attention to the consistency of the Reads distribution in the paired clusters, that is, whether the distribution trends at both ends of Reads are basically the same, so the linear correlation can better reflect the distribution inside the paired clusters.

As a preferred embodiment, the degree of compaction of the clusters, such as the variance, and the linear correlation of the clusters to filter the candidate clusters can achieve good results.

(3) A comparison set according to a normal sample: the paired cluster is compared with a preset comparison set containing a plurality of normal samples, and the paired clusters whose number of hit normal samples reaches a preset threshold V-con are filtered out. . A normal sample refers to a collection of result clusters obtained by an analysis process such as "alignment-cluster-filtering" with other normal individuals of the same biological species as the target individual. For ease of alignment, all of the Reads in the cluster can be merged into one, and the paired clusters produce a pair of fused value pairs (similar to a pair of Reads), using the fused value pairs for comparison. By collecting a control set containing a large number of normal samples, the frequency of occurrence of the result cluster in a normal individual can be obtained. If the frequency of occurrence of a result cluster is high, it may indicate that the result cluster may be due to sample nature, experimental process, sequencing process or environment. What is caused by noise or the like does not mean that such structural variation has occurred in the sample itself. Such a result cluster is a common false positive result obtained by the same method analysis of different samples and should be removed. Therefore, filtering the clusters using the control set can further reduce the probability of false positives and help to obtain real structural variation analysis results. V-con can be determined according to the establishment manner and characteristics of the normal sample, for example, the ratio of the V-con to the normal sample number in the control set may be 3%-10%, preferably 5%-6%, for example, if the control set contains 90% Normal sample In this case, 5 hits can be considered as reaching the threshold.

(iv) Based on other ancillary parameters: The so-called auxiliary parameters include various parameters that help to further confirm, distinguish the type of structural anomaly, or help to understand the details of the structural anomaly. For example, the number of mismatch generated in the comparison process, the number of Reads of the support cluster, the RPU value of the relevant region obtained based on *.pair, whether the cluster is located in the N region, or the like. The use of auxiliary parameters may include two methods. One is as a filtering condition, the filtering requirements related to the auxiliary parameters are set, the clusters that do not meet the requirements are directly filtered out, and the other is used as a reference for the auxiliary judgment, and the auxiliary parameters are accompanied by the results. The clusters are provided together and judged by means of manual analysis. Therefore, the content of this section can be applied to step 4 (for filtering), and can also be applied to the next step after step 5 (for assisting manual analysis). The specific use of the parameters is not limited. The following sections list some auxiliary parameters and their relationship with the result analysis. In actual use, they can be set as filtering conditions according to the following description, or as auxiliary judgment basis for manual analysis. Different auxiliary parameters can be used in combination or separately. Use alone.

(1) Mismatch number: The average mismatch number of Reads in a paired cluster is generally no more than one or two, that is, one or two mismatches are allowed for each pair of Reads, preferably no more than one. If the matching requirement is set according to this setting, it is not necessary to consider the parameter again. If the setting is relatively loose, for example, if two mismatches are allowed, the result cluster can be filtered again according to the parameter. Or judge, for example, to set an average of only one mismatch.

(2) The number of Reads supporting the cluster: that is, the number of Reads included in the paired cluster. The larger the parameter, the better, in general, the judgment may be set to be substantially consistent with the normalized value of the sequencing depth, or slightly smaller than the value ( For example, rounding), the normalized value of the so-called sequencing depth is: sequencing depth × (L-lib impact range of breakpoints / L-lib) × (average of the span of the paired clusters / L-lib) . its The "L-lib impact range of breakpoints" is usually larger than the sum of the spans of the paired clusters. The range of influence of L-lib on breakpoints is generally fluctuated by 2 times L-lib. For example, between 1 and 4 times L-lib, in the specific setting, it can be appropriately relaxed or tightened according to the actual situation.

(3) The RPU value of the relevant area obtained based on *.pair: different types of structural anomalies usually have different effects on the RPU. For example, in the case of balanced translocation, the RPU on both sides of the breakpoint does not change significantly. In the case of missing or repetitive structural abnormalities, the RPU of the region between the breakpoints is significantly reduced or increased, so the RPU value of the relevant region can be used to further verify or assist in determining the occurrence of chromosomal structural abnormalities. For example, for a cluster containing the first type of Reads, if it is judged as a balanced translocation according to the relationship between Reads in the cluster (see step 5, section 1 below), the RPUs on both sides of the breakpoint are relative to the average value. The change should not exceed V-rm. If it is judged as an unbalanced translocation according to the relationship between Reads in the cluster (see step 5, section 1 below), the RPU of the side of the breakpoint away from the result cluster should be lower than Average, and the range of variation exceeds V-rm; for clusters containing the second type of Reads, the RPU of the region between the breakpoints should be higher than the average and the range of variation exceeds V-rm; for the third class of Reads Clusters, the RPU of the region between breakpoints should be below the average and vary beyond V-rm.

When the RPU is used as the auxiliary judgment basis for manual analysis, the RPU of the relevant area may be provided in a graphical, tabular or other easy-to-read manner, or the entire range of RPU changes may be provided in a graphical form, a table, etc., so that The operator understands the overall situation.

(4) Whether the cluster is located in the N zone: According to experience, the Reads alignment near the N zone (which includes the centromere and telomere regions) is relatively more complex than other regions, if obtained The cluster is not located in the N zone and can generally be considered to be judged based on the information obtained. If the obtained cluster is located in the N zone, more careful verification may be required, such as joint use of filtering conditions and auxiliary parameters, or may be combined with other external data. The final judgment is made, for example, by the phenotype of the target individual, and/or by further sequencing the breakpoints (eg, Sanger sequencing).

Step 5. Perform data analysis on the filtered result cluster.

The presence of the resulting cluster after filtering reflects the possible occurrence of a corresponding type of chromosome structural abnormality, so this step is not necessary if only structural anomalies that may exist are found. In order to obtain more detailed information about structural anomalies, the obtained result clusters can be further analyzed. According to different types of result clusters, the following analysis methods can be used:

(1) chromosomal translocation structural abnormalities (first class Reads)

Searching for a result cluster containing the first type of Reads, if the two adjacent read length sequences are in opposite positions in their respective Reads, the range between the positions where the two read length sequences match is obtained as the range of the breakpoint. This situation is usually related to balanced translocations, where Read in the same cluster is distributed on both sides of the breakpoint.

If there is no Read in the above case, the position of the innermost Read is obtained, and the preset length is extended inward from the position as the range of the breakpoint, and the innermost Read refers to if the cluster contains all Read on the left end, the Read on the far right is the innermost Read. If the cluster contains the right Read, the leftmost Read is the innermost Read. This situation is usually related to unbalanced translocations, where Read in the same cluster is distributed on one side of the breakpoint. The span of the breakpoint range extending from the innermost Read can be determined according to L-lib, L-r1/L-r2, sequencing depth, etc., for example, 0.5 to 2 times L-lib, generally not more than 2 times L -lib.

Referring to Figure 2, there is shown a case of balanced translocation if a pair of result clusters are obtained (only two read length sequences are drawn in each cluster, and the rest are treated as omitted) as shown in Fig. 2. One result cluster is located near the position pa of the chromosome chra, and the paired result cluster is located near the position pb of the chromosome chrb. . Because of the cluster on chra, Read1 is the Read end of the left end of the chromosome fragment, and its adjacent Read2 is the Read end of the right end of the chromosome fragment. Therefore, it can be considered that the breakpoint pa of chra is located between Read1 and Read2, and the analysis on chrb is Similar

According to the above data analysis, for possible translocation structural abnormalities, the following results can be output: the number of two chromosomes in which a translocation structural abnormality may occur (the chromosome in which the result cluster is located), and the two ends of the paired result cluster The range of positions (the range of positions of the ends of the cluster on the two chromosomes, the span of the two ends of the cluster can be obtained), the range of the breakpoint obtained by the analysis, and the like. The relevant parameters and other auxiliary parameters generated during the filtering process can also be output together, for example, the compactness of each pair of result clusters, the linear correlation between each other, the number of Reads supporting the pair of result clusters, and the performance breakpoints. Graphs, tables, etc. of the side RPU changes.

(2) Chromosome tandem repetitive structural abnormalities (second type Reads)

Searching for a result cluster containing the second type of Reads, obtaining a range between the two positions that are the farthest distance in the paired cluster as the range in which the repetition occurs, and extending the preset length from the two positions respectively For example, 0.5 to 2 times L-lib, as a range of breakpoints (starting and ending points of repeated segments).

Referring to FIG. 3, a case of tandem repetition is shown, in which both ends of the paired result cluster (only one read length sequence is drawn in each cluster, and the rest are regarded as omitted) are located within the range between the start and end points of the repeated segments. Therefore, it can be considered that the start and end points of the repeated segments are located in a range extending outward from the most edge of the cluster (the two Reads do not necessarily belong to a pair of Reads).

Result data class of repetitive structural anomaly output compared with translocation structural anomaly The types are roughly the same, except that the chromosome numbers at both ends of the cluster are the same, and data indicating the length of the estimated repeat fragment can also be output.

(3) Chromosomal deletion structural abnormalities (third type Reads)

Searching for a result cluster containing the third type of Reads, obtaining a range between the two closest positions in the paired cluster as the range in which the deletion occurs, and extending the preset length from the two positions respectively. For example, 0.5 to 2 times L-lib, as a range of breakpoints (starting and ending points of missing fragments).

Referring to FIG. 4, a case where a segment is missing is shown, and both ends of the paired result cluster (only one read long sequence is drawn in each cluster, and the rest are regarded as omitted) are located outside the start and end points of the missing segment, and thus The starting and ending points of the missing segment are considered to be in the range extending inward from the closest Read at both ends of the cluster (the two Reads do not necessarily belong to a pair of Reads).

The result data type of the missing structural anomaly output is roughly the same as that of the repetitive structural anomaly, except that the output data representing the length of the segment between the estimated breakpoints represents the length of the missing segment.

Step 6. Breakpoint assembly.

To further narrow the range of breakpoints, you can also use the data in *.unmap for breakpoint assembly, for example, to obtain a single-ended Read within the set range around the determined breakpoint range (for example, 0.5 to 2 times L-lib). Read that can be matched to the reference sequence with a single end, can be divided into *.sin when compared, extract Read from the *.unmap as a patch sequence, and cut all patch sequences into N segments, N preferred 2, and the subsequence obtained after truncating the patch sequence is re-aligned with the reference sequence, and the breakpoint region is assembled according to the result of normal matching.

In actual use. The value of N can be reasonably set according to the length of Lr1/Lr2, due to the order A column length of less than 25 bp results in a large decrease in the unique alignment ratio. Therefore, when setting the N value, it can be considered that the length of the truncated subsequence is not lower than or less than 25 bp.

After the breakpoint assembly, the range of the breakpoint can be effectively reduced. On this basis, the probe can be further prepared according to the position range of the breakpoint, and other accurate sequencing methods, such as Sanger sequencing, can be used to obtain accurate Breakpoint position for further study of breakpoints. If you do not need to narrow the breakpoint range, this step can be omitted.

A person skilled in the art can understand that all or part of the steps of the various methods in the above embodiments may be completed by a program to instruct related hardware, and the program may be stored in a computer readable storage medium. The storage medium may include: a read only memory, a random Memory, disk or disc, etc.

According to another aspect of the present invention, there is provided an apparatus for detecting an abnormality in a chromosome structure, comprising: a data input unit for inputting data; a data output unit for outputting data; and a storage unit for storing data including executable a program, coupled to the data input unit, the data output unit, and the storage unit, for executing an executable program stored in the storage unit, the execution of the program comprising completing all or part of the steps of the various methods in the above embodiments. .

The results of the operation of the specific detection method according to the present invention will be described in detail below in conjunction with a specific target individual. The specific parameters used in the following detection procedure are set as follows: 1. L-lib is 500 bp, PE50 is sequenced (pair-end sequencing, L-r1 and L-r2 are substantially 50 bp); 2. NCBI HG19 is selected as the reference sequence, Sequencing the sequencing results using SOAP software; 3. V-lib is ±45 bp, RPK has V-rm of 20%, V-cl is 10Kbp (inter-cluster distance is defined as the distance between two nearest Reads), and the minimum number of Reads in the cluster is 2, R- Va and set the variance in the rank of all clusters in the lowest interval of 5% (when calculating the variance, ignore the read length sequence in the length range of 20% at both ends of the cluster), R-li is set to the correlation coefficient at all The ranking in the cluster is in the lowest interval of 5%, and the control set includes 90 normal samples with a V-con of 5.

Experimental example one

This example is a study of the family of meow syndrome. The two target individuals in this example belong to a family, where "FA" means father and "SON" means son.

1. Complete genome-wide low-multiplier sequencing of the two target individuals, with a "FA" sequencing depth of 2.2 and a "SON" sequencing depth of 3.1.

2. Then use the SOAP comparison software to compare the sequencing results of the two target individuals with the reference sequence HG19, respectively, and obtain two files FA.sin and SON.sin.

3. Cluster, filter and analyze the two files FA.sin and SON.sin, and obtain the result cluster and related parameters as follows:

"FA":

The number of the two chromosomes in which the paired result cluster is located: chr12, chr5

The position range of the two ends of the paired result cluster: 14797215-14780233, 23314785-23314205

Span of the two ends of the paired result cluster: 618, 580

Number of Reads supporting the pair of result clusters: 5

Tightness at the left and right ends (variance): 90.59, 87.01

Whether it is located in the N zone: No

The range of breakpoints: chr12:14779968~14780233, chr5:23314205~23314455

The change of RPK in the relevant region on the chromosome: as shown in Figure 6, the abscissa is the position on the chromosome, in units of 10Kbp, the ordinate is RPK, the curve is drawn according to the data of FA.pair, and pa and pb are broken. The position of the point, as can be seen from the figure, there is no significant change in the RPK of "FA".

"SON":

The number of the two chromosomes in which the paired result cluster is located: chr12, chr5

The position range of the two ends of the paired result cluster: 14797618-14779968, 23314455-23314830

Span of the two ends of the paired result cluster: 350,375

Number of Reads supporting the pair of result clusters: 6

Tightness (variance) at the left and right ends: 22.43, 18.44

Whether it is located in the N zone: No

The range of breakpoints: chr12: greater than 14779968, chr5: less than 23314455

The change of RPK in the relevant region on the chromosome: as shown in Figure 7, the abscissa is the position on the chromosome, in 10Kbp, the ordinate is RPK, the curve is drawn according to the data of SON.pair, and pa and pb are broken. Point position, it can be seen from the figure that the RPK of "SON" has obvious changes. Looking at the RPK calculation value, the RPK of the forearm of chromosome 5 of SON is only 0.5 times of the average value, and the forearm of chromosome 12 is more than the average value. 0.5 times.

Through the analysis results, it can be clearly judged that "FA" is a balanced translocation, "SON" is an unbalanced translocation, and the range of breakpoints analyzed by the result of "FA" is already within 300 bp. for Further research on the breakpoint position, then we took the corresponding sequence from the reference sequence HG19, designed the primers, verified the qPCR and Sanger sequencing, and finally got the exact breakpoint position: Chr12:14780019, Chr5: 23314435.

Experimental example 2

This case is a study of congenital heart disease. The target individual in this case is a patient with congenital heart disease, expressed as "XX".

1. Perform a genome-wide low-multiplier sequencing of the target individual with a sequencing depth of 2.7.

2. The sequencing results are then compared to the reference sequence HG19 using SOAP alignment software to obtain XX.sin.

3. Cluster, filter and analyze XX.sin, and obtain the result cluster and related parameters as follows:

"XX":

The number of the two chromosomes in which the paired result cluster is located: chr14, chr14

The position of the two ends of the paired result cluster is in the range of 7357040-73557288, 73670432-73670682

Estimated length of the repeating segment: 113392

Span of the two ends of the paired result cluster: 248,250

Number of Reads supporting the pair of result clusters: 4

Tightness (variance) at the left and right ends: 100.63, 100.59

Whether it is located in the N zone: No

The range of breakpoints: chr14:73556540-73557040, Chr14:73670682-73671182 (the range size is estimated by 1 times L-lib, ie 500bp)

From the analysis results, it can be clearly judged that a chromosome of "XX" has a repetition of about 113 Kbp, and the repetition occurs in tandem. To further study the position of the breakpoint, we took the corresponding sequence from the reference sequence HG19, designed the primers, verified the qPCR and Sanger sequencing. The amplification ratio of qPCR was >1, which was shown as repeat, and Sanger was finally sequenced. The exact breakpoint position is: Chr14:73557008, Chr14:73670820, confirming that a chromosome of chromosome 14 of "XX" has a 113812 bp repeat, and the repeat fragment is inserted in tandem to the end of the fragment.

The above is only the preferred embodiment of the present invention, and it should be understood that these embodiments are only used to explain the present invention and are not intended to limit the invention. Variations to the above-described embodiments may be made in accordance with the teachings of the present invention.

Claims (14)

A method for detecting abnormality of a chromosome structure, comprising the steps of: obtaining a whole genome sequencing result of a target individual, wherein the sequencing result comprises a plurality of pairs of read length pairs, each pair of read length pairs consisting of two read length sequences, respectively Located at both ends of the measured chromosome fragment, each pair of read pairs are from the positive and negative strands of the corresponding chromosome fragment, respectively, or each pair of read pairs are simultaneously from the positive or negative strand of the corresponding chromosome fragment; Aligning with a reference sequence to obtain an abnormal matching set, the abnormal matching set including a first type of read length pair that meets the following description, and the two read length sequences of the first type of read long pair respectively match the difference of the reference sequence a chromosome; clustering the read length sequences in the abnormal matching set into clusters according to the matched positions, each cluster containing a single-ended read length sequence from a set of read length pairs, and the corresponding read length sequence at the other end is located at another In a cluster; filtering clusters obtained by clustering, including, calculating the degree of compactness of each cluster, filtering out clusters whose degree of compaction does not satisfy the preset requirement R-va and pairing with them The clusters obtained the filtered result clusters containing the first type of read length pairs for use in determining the occurrence of chromosomal translocation structural abnormalities. The method of claim 1, wherein when filtering the clusters obtained by clustering, the method further comprises: calculating a linear correlation between the two pairs of the clusters, and filtering out the linear correlation does not satisfy the pre- Paired clusters of R-li are required, and/or, the paired clusters are compared with a preset control set containing a plurality of normal samples, and the number of normal samples hitting reaches a preset threshold V-con Paired clusters. The method of claim 1, further comprising searching for a result cluster containing the first type of read length pair, if the two adjacent read length sequences are in respective read length pairs. Instead, the range between the positions where the two read length sequences are matched is obtained as the range of the breakpoint, and if the read length sequence of the above case does not exist, the position of the innermost read long sequence is acquired and outward from the position Extend the preset length as the range of breakpoints. The method of claim 1, wherein the abnormal matching set further comprises a second type of read length pair that conforms to the following description, and the two read length sequences of the second type of read long pair are matched to Reference to the same chromosome of the sequence, but the length L-pr of the chromosome fragment calculated from the matched position is negative; and the filtered result cluster containing the second type of read length pair is also obtained for judging the tandem repeat of the chromosome The occurrence of sexual structural abnormalities. The method of claim 4, further comprising: searching for a result cluster containing the second type of read length pair, and obtaining the closest distance between the two locations in the paired cluster The range is taken as the range in which the repetition occurs, and the preset length is extended outward from the two positions as the range of the breakpoint. The method of claim 1, wherein the abnormal matching set further comprises a third type of read length pair that conforms to the following description, and the two read length sequences of the third type of read length pair are matched to The same chromosome of the reference sequence, but the length L-pr of the chromosome fragment calculated from the matched position is larger than the library size L-lib and the deviation exceeds the preset threshold V-lib, and V-lib is preferably 5%×L- Lib~15%×L-lib, further preferably 10%×L-lib; also obtained a filtered result cluster containing the third type of read length pair for judging the chromosomal deletion knot The occurrence of an abnormality. The method of claim 6, further comprising: searching for a result cluster having a third type of read length pair, and obtaining a matching distance between the two positions in the paired cluster The range is the range in which the deletion occurs, and the preset length is extended inward from the two positions as the range of the breakpoint. The method according to any one of claims 1 to 7, wherein, when the sequencing result is compared with a reference sequence, the method further comprises: obtaining a normal matching set, the normal matching set Including a read length pair that meets the following description, the two read length sequences of the read long pair match the same chromosome of the reference sequence, and the positive and negative chain relationship of the matched position is consistent with the positive and negative chain relationship in the read length pair, And the deviation of the length L-pr of the chromosome fragment calculated from the matched position and the size L-lib of the library used for sequencing is less than a preset threshold V-lib, and V-lib is preferably 5%×L-lib ~15%×L-lib, further preferably 10%×L-lib, the number of Reads in the normal matching set included in the statistical unit length is RPU, and the change of the RPU relative to the average value is obtained to assist in judging the structural abnormality. The occurrence of the RPU with respect to the average value is preferably expressed by whether the change in RPU exceeds a preset threshold value V-rm. Preferably, V-rm is 10 to 30%, and more preferably 20%. The method according to any one of claims 1 to 7, wherein when the sequencing result is compared with a reference sequence, the method further comprises: obtaining an unmatchable set, the unmatchable set Including a read length sequence that cannot be matched to a reference sequence, including a pair of read length sequences that cannot be matched in pairs or a read length sequence that cannot be matched by a single end, and after obtaining the result cluster, includes Obtaining a single-ended read length sequence within a set range around the determined breakpoint range, extracting a paired read length sequence from the unmatchable set as a patch sequence, and cutting all the patch sequences into N segments, N is preferably 2, and The subsequence obtained by truncating the patch sequence is re-aligned with the reference sequence, and the breakpoint region is assembled according to the result of normal matching. The method according to any one of claims 1 to 7, characterized in that, in calculating the degree of compactness of each cluster, abandoning each of the 5% to 25% of the read length sequences located at both ends of the cluster is not Participate in the calculation, and/or, the degree of compaction is expressed by the variance, and R-va is set such that the ranking of the variance in all clusters is within the lowest interval of 2% to 10%, preferably 5%. The method of claim 1, wherein when calculating the linear correlation of the paired two clusters, the linear correlation is represented by a correlation coefficient, and R-li is set as the correlation coefficient in all clusters. The ranking is in the highest range of 2% to 10%, preferably 5%, and/or the ratio of V-con to the number of normal samples in the control set is 3%-10%, preferably 5%-6%. The method of claim 1, wherein the size of the library used for sequencing is L-lib 300 bp, preferably 500 bp or 5 Kbp, and/or the length of the read length sequence is greater than or equal to 25 bp, preferably 50 bp plus or minus 10%. A device for detecting abnormality of a chromosome structure, comprising: a data input unit for inputting data; a data output unit for outputting data; a storage unit for storing data, including an executable program; a processor, Data connection with the data input unit, the data output unit, and the storage unit for executing the executable program, and the execution of the program includes completing the first item of the patent application scope The method of any of item 12. A computer readable storage medium for storing a program for execution by a computer, the execution of the program comprising the method of any one of claims 1 to 12.