CN113096734A

CN113096734A - Method for screening molecular marker combination for diploid population paternity test

Info

Publication number: CN113096734A
Application number: CN202110514979.4A
Authority: CN
Inventors: 夏晓勤; 夏雷; 石米娟; 张婉婷; 程莹寅
Original assignee: Institute of Hydrobiology of CAS
Current assignee: Institute of Hydrobiology of CAS
Priority date: 2021-05-11
Filing date: 2021-05-11
Publication date: 2021-07-09
Anticipated expiration: 2041-05-11
Also published as: CN113096734B

Abstract

The invention discloses a method for screening a marker combination for paternity test from molecular markers of a parent population of a given diploid species, which comprises the following steps: and taking all the parent pairs as root nodes of a tree structure, classifying the parent pairs according to the classification of all possible generated filial generations of each parent pair to construct child nodes, and marking the obtained child node with the minimum sum of the number and the variance of the parent pairs as a first molecular marker. The branch node obtained by the first marking (at least containing the child nodes of two parent pairs) is used for screening the second marking by the same method until the final child nodes are all leaf nodes (only containing the child nodes of one parent pair). The method can screen out the minimum marks required by completing paternity test from a large number of marks, the obtained marks have high typing efficiency, the paternity test result is reliable, and the workload and the cost required by paternity test can be reduced by minimizing the number of the marks.

Description

Method for screening molecular marker combination for diploid population paternity test

Technical Field

The invention belongs to the technical field of bioinformatics, and particularly relates to a method for screening a marker combination for paternity test from molecular markers of a given diploid population.

Background

Diploid species are widely found in nature, and the diploid species closely related to human beings include various mammals such as cows, sheep, pigs and dogs, and many common economic fishes. Most of these species are kept as food for humans or as pets. Through genetic improvement breeding of the diploid organisms, the yield can be improved, the development period can be shortened, domestication can be accelerated, the appearance characteristics can be changed, and the like, so that better economic benefit and social benefit can be generated. Molecular marker assisted breeding is a commonly used breeding scheme at present, namely, individuals with excellent genes are screened through specific molecular markers and are used as improved parents to breed offspring so as to achieve the aim of genetic breeding. One widely used method for screening specific genes or molecular markers is to search for molecular marker sites (e.g., single nucleotide polymorphism sites) associated with the trait from the whole genome by bioinformatics means, and then verify the associated genes by combining molecular means. This analysis process usually requires paternity testing and creation of families for individuals to eliminate the effects of genetic background. Besides, paternity testing is widely applied to legal procedures such as human forensic testing.

At present, paternity test is mainly carried out according to the typing of molecular markers in different parents and filial generations, the molecular markers mainly comprise microsatellite molecular markers (SNP), Single Nucleotide Polymorphism (SNP) molecular markers and haplotype (or micro-haplotype) molecular markers, and the molecular markers are widely distributed in a genome range. In order to minimize the paternity test workload, it is desirable to use as few molecular markers as possible. In order to find a marker for paternity test, a series of molecular markers are obtained from an existing marker database, parents and offspring are subjected to PCR amplification to complete the typing of the markers, and then paternity test marker screening can be performed by using the minimum allele frequency (MAF value), the parent exclusion rate (PE), the Polymorphic Information Content (PIC) and the like as parameters, and can also be performed by Principal Component Analysis (PCA), a greedy algorithm, a Monte Carlo algorithm and the like. However, both methods either fail to obtain a genome-wide paternity-specific molecular marker with high discrimination between parents, or fail to ensure that the number of markers obtained is "as small as possible".

The invention uses greedy algorithm and tree data structure to screen paternity test mark, and can obtain minimum combination of paternity test mark, thereby reducing cost of paternity test and improving efficiency.

Disclosure of Invention

The invention aims to provide a method for screening out a marker combination for paternity test from molecular markers of a given diploid parental group, which has the advantages of screening out the minimum marker combination capable of completing paternity test and improving paternity test efficiency.

In order to achieve the purpose, the invention adopts the following technical scheme:

s1, constructing parent pairs by parents according to a male parent and a female parent, taking all possible combinations of the parent pairs as root nodes of a tree structure shown in figure 1 to represent all parent pair sets to be distinguished, and screening a first mark by taking the root nodes as branch nodes of a first level (at least comprising child nodes of two parent pairs);

s2, the method for screening a paternity test molecular marker from a certain level of the tree structure comprises the following steps: for each candidate molecular marker, estimating the subtyping of the filial generation in each branch node of the same level according to the subtyping of the parents, and then constructing a sub-node of the branch node at the next level of the tree structure by using the parent pair corresponding to the subtyping of the same filial generation according to the subtyping condition of the filial generation, so that a plurality of sub-nodes can be obtained by the filial generation of different subtyping; in the next level, after leaf nodes (child nodes containing only one parent pair) and repeated child nodes are removed, the number of parent pairs contained in each child node is counted, the mean value and the variance of the parent pairs are calculated, the typing effect of the marker is evaluated by the sum of the mean value and the variance (the square mean sum), and the molecular marker with the minimum square mean sum is selected from all candidate markers as the molecular marker for parent-child identification of the level;

the specific method for constructing the child node in the above steps is as follows:

obtaining all filial generation genotypes which can be generated by each parent pair according to Mendelian segregation law and free combination law, and then generating child nodes according to the filial generation genotypes: all parent pairs which can generate a certain offspring genotype form a child node, and the child nodes form the next layer of the tree structure together;

the variance and mean value calculation method in the steps has the following meanings:

in the child nodes constructed by the Mendelian genetic law, the number of parent pairs contained in all nodes of the same hierarchy is counted, and the variance and the average number of the parent pairs are calculated. The smaller the sum of the two is, the more evenly the parent pairs to be distinguished are distributed in the nodes contained in the layer as much as possible, thereby reducing the number of branch nodes containing more significant parent pairs and being beneficial to screening of subsequent parent pairs;

s3, acquiring child nodes generated by the screened molecular marker, removing leaf nodes and repeated child nodes in the child nodes, using the rest nodes as father nodes to construct child nodes of a next layer of tree structure, and screening the next marker according to the step S2 until no new branch node is generated;

and S4, summarizing the molecular markers screened in the S2 and the S3, wherein the marker combination is a set of paternity test markers which can be used for identifying parents corresponding to all filial generations.

Compared with the prior art, the invention has the following advantages:

for the common SSR markers, at present, parent exclusion rate (PE) and Polymorphic Information Content (PIC) parameters which are marked in a database are mainly used for screening parent-child identification markers, and the requirement of 'as few as possible' cannot be met. For SNP markers, candidate molecular markers are optimized through principal component analysis and a Monte Carlo algorithm, then improper markers are eliminated one by one mainly according to an estimated value of paternity test effectiveness, and finally the number of obtained effective markers is large. The invention mainly predicts the possible offspring genotypes of all the parents based on Mendelian genetic law, and screens the markers layer by layer according to a tree structure and a greedy algorithm, so that all the parent pairs can be identified singly. According to the principle, the method used by the invention has high accuracy and strong operability, and has the advantages that the method is not limited by the type of the marker, only the typing of each parent is needed to be known, the obtained paternity test marker theoretically can ensure that the paternity test success rate is higher, and the minimum marker set for paternity test can be screened out.

Drawings

FIG. 1 is a tree structure used for screening paternity test markers, wherein P1 and P2 are male parents, P3 and P4 are female parents, each circle in the figure represents a child node, open circles represent leaf nodes, filled circles represent branch nodes, and characters in the circles represent offspring as a pair of parents of a certain type.

The specific implementation mode is as follows:

example 1

The sample used in this example is derived from the grass carp parent (9 female fish and 9 male fish) for breeding in 2017 of the laboratory, the molecular marker used is a micro-haplotype molecular marker, and the typing result is obtained by whole genome re-sequencing.

S1) determination of the first flag:

s1.1) pairing all grass carp parents for propagation according to sex, wherein the sex relationship is shown in Table 1, so that 81 pairs of parent pairs are obtained, taking FIG. 1 as an example, P1 and P2 are male, P3 and P4 are female, then P1 and P3, P1 and P4, P2 and P3, P2 and P4 can respectively form a pair of parent pairs, and the parent pairs are used as root nodes of the tree structure shown in FIG. 1 for subsequent marking screening;

TABLE 1 parent population

S1.2) according to the patent application "a method for obtaining a mini-haplotype and its typing directly from whole genome re-sequencing data" (application no: 201811248346.8) obtaining 6192 marks from the 18 parents, and using 81 parent pairs formed by the 18 parents as the root nodes shown in figure 1 and as the parent nodes. For each candidate marker, respectively counting the offspring types of each parent pair in the root node. And according to the child typing condition, constructing child nodes of the second-layer tree structure by using parent pairs corresponding to children with the same typing, and counting the sum of the mean value and the variance of the number of the parent pairs of each child node in the second-layer tree structure after removing leaf nodes and repeated child nodes. The mark with the smallest sum is selected as the first mark and recorded. Table 2 shows the case of markers where the sum of the mean and variance appeared to decrease during the screening of the first marker:

TABLE 2 markers with reduced sum of mean and variance during the first marker screening

The process of screening the marks with smaller sum of variance and mean is shown in the table above, the first column is the current best mark obtained in the screening process, the second column is the number of parent pairs in child nodes constructed by each best mark, and the third column is the sum of mean and variance calculated after removing a single parent pair in the second column and child nodes with the same parent pair;

s1.3) according to the above table, the sum of the mean variances is minimum 35.51, and is therefore selected

CI01000040_900697_900879 marks the child nodes of the second-level tree structure, and the child nodes of the second-level tree structure are obtained by removing a single parent pair in the obtained child nodes and the child nodes with the same parent pair.

S2) determination of other flags:

and screening the second marker, and constructing child nodes of a third-level tree structure by taking the non-repetitive nodes with parent pairs more than 1 in the second-level tree structure generated by the CI01000040_900697_900879 marker as parent nodes. And (3) carrying out statistics on the sum of the average value and the variance of parent pairs contained in child nodes which can be generated by each candidate marker, and calculating the sum of the marker and the average value and the variance obtained in the running process as follows:

(CI01000009_12794233_12794303,30.65),(CI01000009_130841_131085,19.93),

(CI01000008_1950827_1950948,17.71),(CI01000009_11236184_11236253,14.91),

(CI01000009_14535665_14535913,9.15),(CI01000008_1804841_1805103,7.48),

(CI01000006_5364061_5364302,6.97),(CI01000011_222156_222320,6.08),

(CI01000027_989512_989749,4.96), (CI01117242_1048_1191,4.80), therefore, the second marker selects CI01117242_1048_ 1191;

when the third mark is screened, the non-repetitive nodes with the number of parent pairs larger than 1 in the third-layer tree structure generated by the mark CI01117242_1048_1191 are used as parent nodes to construct child nodes of the fourth-layer tree structure. The sum of the markers and the mean and variance obtained during the run were calculated as follows:

(CI01000009_12794233_12794303,4.74),(CI01000009_130841_131085,3.97),

(CI01000008_1950827_1950948,3.09),(CI01000006_5364061_5364302,2.73),

(CI01000013_12036913_12037141,2.57), (CI01000062_1652939_1653162,2.54), so the third marker selects CI01000062_1652939_ 1653162;

by analogy, the sum of the markers, the mean and the variance obtained in the screening process of the fourth marker is calculated as follows:

(CI01000009_12794233_12794303,2.53),(CI01000009_130841_131085,2.21),

(CI 01000008-1950827-1950948, 2.10), (CI 01000304-11510857-11511051, 2.0), wherein there are many markers where the sum of mean and variance is 2.0. We select the marker with the least number of children nodes that do not repeat, therefore, the fourth marker is selected CI01000304_11510857_ 11511051;

the sum of the markers and the mean and variance obtained during the screening of the fifth marker was calculated as follows:

(CI01000009_12794233_12794303,2.53),(CI01000009_130841_131085,2.21),

(CI01000008_1950827_1950948,2.10), (CI01000000_1804567_1804763,2.0), and so on, and having many tags that are 2.0, we still select the tag that contains the fewest number of children that do not repeat, so the fifth tag selects CI01000000_1804567_ 1804763;

the sixth marker selects CI01000006_5364061_5364302, the parent pairs can be completely distinguished.

S3) summarizing the above markers, the final marker combination obtained is:

CI01000040_900697_900879，CI01117242_1048_1191，

CI01000062_1652939_1653162，CI01000304_11510857_11511051，

CI01000000_1804567_1804763, CI01000006_5364061_ 5364302. And constructing seven-layer tree structure.

Example 2

This example uses the 3-tailed female 2-tailed male (total 5-tailed parents), and its 171-tailed offspring (paternity to the 5-tailed parents has been acquired using SSR markers) in 2016 in this laboratory. For subsequent analysis, parents adopt a whole genome re-sequencing technology, the sequencing depth is 30 x, and offspring adopts a whole genome re-sequencing technology, the sequencing depth is 15 x. Based on each parental sequencing data, 9519 micro-haplotype molecular markers have been obtained from the genome-wide range.

Because progeny may not have an exact typing due to insufficient sequencing depth in some fragments, we used the above method to obtain three sets of paternity testing markers, the first set is CI01000059_6276399_6276582, the second set is CI01000167_440528_440807, and the third set contains two markers, CI01000372_42297_42492 and CI01000316_637447_ 637943.

Compared with SSR paternity test results, the paternity relationships of 110 filial generations can be tested by using the first set of markers, and the test consistency rate is 100%; 162 progeny can be identified by using the first two sets of markers, the identification consistency rate is 99.38%, and 1 false progeny exists; the 169 progeny can be identified by using the three sets of markers, the identification consistency rate reaches 98.22%, and 2 wrong progeny exist. Through analysis, the reason for the error in the identification of the wrong offspring is probably that the sequencing depth is insufficient, so that more accurate typing cannot be obtained, and the final typing result is influenced.

In conclusion, this example shows that the method provided by the present invention can screen molecular markers that can be used for paternity test, and the success rate of paternity test of the obtained markers can reach more than 98%, which illustrates the effectiveness and feasibility of the method of the present invention.

Claims

1. A method of screening a combination of molecular markers for paternity testing, comprising the steps of:

s1, constructing parent pairs by parents according to a male parent and a female parent, combining all candidate parent pairs into a root node of a tree structure to represent all parent pair sets to be distinguished, and starting to screen a first mark by taking the root node as a branch node of a first level;

s2, the method for screening a paternity test molecular marker from a certain level of the tree structure comprises the following steps: for each candidate molecular marker, estimating the subtyping of the filial generation in each branch node of the same level according to the subtyping of the parents, and then constructing a sub-node of the branch node at the next level of the tree structure by using the parent pair corresponding to the subtyping of the same filial generation according to the subtyping condition of the filial generation, so that a plurality of sub-nodes can be obtained by the filial generation of different subtyping; in the next level, after leaf nodes and repeated child nodes are removed, the number of parent pairs contained in each child node is counted, the mean value and the variance of the parent pairs are calculated, the typing effect of the marker is evaluated according to the sum of the mean value and the variance, and the molecular marker with the minimum sum of the mean value and the variance is selected from all candidate markers to serve as the paternity test molecular marker of the level;

s3, obtaining child nodes generated by the screened molecular marker, eliminating leaf nodes and repeated child nodes in the child nodes, using the rest nodes as father nodes to construct child nodes of a next layer of tree structure, and screening the next marker according to the step S2 until no new branch node is generated;

s4, summarizing the molecular markers screened in the S2 and the S3, and combining the markers to form a set of paternity test markers.