SNP (Single Nucleotide Polymorphism) chip for distinguishing five kinds of Hubei native cattle

Abstract

The invention belongs to the technical field of animal genetic breeding, and particularly discloses an SNP (Single Nucleotide Polymorphism) chip for distinguishing five kinds of Hubei native cattle (yi-ling cattle, Dabie mountain cattle, troilin cattle, Enshi cattle and delta mountain cattle). Candidate SNP loci are obtained by combining whole genome sequencing and three selection signals (Tajima's D, FST and pi ratio), then the candidate loci are subjected to random forest scoring, the number of the needed SNP loci is further reduced, finally, an SNP locus combination with variety specificity is obtained, probes are designed according to the SNP loci and 60 bp sequences before and after the SNP loci, and the variety specificity is obtained. The SNP chip for distinguishing five kinds of Hubei native cattle can be developed, technical support is provided for variety breeding of Hubei native cattle, and meanwhile, compared with a traditional pedigree and microsatellite variety identification method, the SNP chip has the advantages of being high in accuracy and easy to operate.

Description

Technical Field

[0001] The invention belongs to the technical field of animal genetic breeding, and in particular relates to a SNP chip for distinguishing five kinds of Hubei native yellow cattle. Background Art

[0002] Cattle are an important livestock in my country, and there are many local cattle breeds in Hubei Province. Hubei cattle are tolerant to roughage, have strong fertility, and gain weight quickly in summer and autumn. Enshi cattle, Yiling cattle, and Yunba cattle are included in the national variety list as livestock and poultry genetic resources, and "Enshi beef" has obtained national geographical indication product protection. Yiling marbled beef sells for 1,200 yuan per kilogram, becoming the main beef cattle breed developed in Hubei Province. Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle have relatively similar appearance characteristics, including similar coat color and size, making it difficult to identify the varieties of these Hubei cattle from their appearance. In recent years, Hubei Province has continuously encouraged and developed beef cattle breeding, introduced large foreign beef cattle breeds as breeding breeds, and formed a certain scale in Hubei; however, since Hubei Province is located in the south of my country, the climate characteristics are mainly hot and humid, and foreign beef cattle breeds are poorly adaptable to the hot and humid climate in the south, which is not conducive to production; at the same time, the large-scale introduction of foreign breeds has threatened the protection of my country's genetic resources. The breeding scale and breeding of Hubei cattle have been impacted to a certain extent, and there are currently few studies on the genetic resources of Hubei cattle. In order to fully explore the genetic resources of Hubei cattle, it is necessary to crossbreed Hubei cattle with foreign breeds, and crossbreeding urgently requires a method for variety identification. Although the traditional pedigree identification method can identify varieties through statistical pedigree data, it still has certain difficulties in accuracy and complexity of data collection. In addition, variety identification can also be performed by identifying microsatellite sequences in sequences. This method belongs to the field of molecular biology. Although it has advantages in accuracy, the reagents required for the experiment are toxic to a certain extent, and the experimental cycle is long and the operation is difficult.

[0003] In recent years, high-throughput sequencing has developed rapidly, and second-generation sequencing and third-generation sequencing have been launched one after another. The cost of sequencing has further decreased, and the accuracy and speed of sequencing have further increased. Selection signal analysis plays a key role in population genetics, aiming to identify the effects of natural selection and artificial selection on variation in the genome, compare genetic variation in different populations, and reveal the relationship between adaptive traits and selection pressure. In specific applications, Chen et al. used FST, XP-EHH, and πratio selection signals to study the heat tolerance of zebu cattle (Chen, N., Xia, X., Hanif, Q. et al. Global genetic diversity, introgression, and evolutionary adaptation of indicine cattle revealed by whole genome sequencing. Nat Commun 14, 7803 (2023). https: / / doi.org / 10.1038 / s41467-023-43626-z). Random Forest (RF) is a machine learning method for classification and regression. Gao et al. used this method to identify breeds in pigs in China (Gao, J.; Sun, L.; Zhang, S.; Xu, J.; He, M.; Zhang, D.; Wu, C.; Dai, J. Screening Discriminating SNPs for Chinese Indigenous Pig Breeds Identification Using a Random Forests Algorithm. Genes 2022, 13, 2207. https: / / doi.org / 10.3390 / genes13122207), but it has not yet been carried out on cattle. The present invention combines whole genome sequencing and selection signal analysis to obtain SNP (single nucleotide polymorphism) sites, and then further scores the SNP sites by finding the classification branch with the least error rate, obtains breed-specific SNP sites, and further develops them into SNP chips for distinguishing five types of Hubei native yellow cattle. Compared with the traditional selection signal method, the combination of random forest and selection signal can further reduce the required loci and achieve more accurate breed identification with fewer loci, providing options for the breed selection and breeding of Hubei native yellow cattle.

[0004] In summary, we developed a SNP chip for distinguishing five native Hubei yellow cattle (Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle) by combining second-generation high-throughput sequencing technology, selection signals, and random forest methods, providing important technical support for the breed identification and improvement of Hubei yellow cattle. Summary of the invention

[0005] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a SNP chip for distinguishing five types of Hubei native yellow cattle (Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle).

[0006] The present invention adopts the following technical solutions:

[0007] The present invention screens out SNP site combinations for identification of five Hubei yellow cattle breeds based on random forest and selection signal methods. The specific screening method includes:

[0008] S1. Collect the original second-generation whole genome sequencing data of cattle, align it to the reference genome ARS-UCD1.2 of cattle, and identify and screen SNP variant sites;

[0009] S2. SNP variant sites were screened according to S1, and the union of three selection signals, Tajima's D, FST, and πratio, was used, and then LD pruning was used to obtain the screened sites;

[0010] S3. Use the random forest model to input the corresponding site and variety information to score the site, and select the minimum number of sites required to make the accuracy of the five varieties reach 100%; specifically, input the site into the random forest model, score the site, obtain the site gini and accuracy scores, and then divide the site data into a test set and a training set. At the same time, input the variety information into the random forest model, use the sites with gini scores, and gradually increase the number of sites in descending order of scores to find the minimum number of 285 sites that can make the identification accuracy of the five varieties reach 100% and be stable.

[0011] S4. Use the random forest model to input the corresponding site and variety information to score the site, and select the minimum sites required to make the accuracy of the five varieties reach 100%; specifically, input the site into the random forest model, score the site, obtain the site gini and accuracy scores, and then divide the site data into a test set and a training set. At the same time, input the variety information into the random forest model, use the sites with gini scores, and gradually increase the number of sites in descending order of scores to find the minimum 285 sites that can make the identification accuracy of the five varieties reach 100% and are stable. The specific site information is shown in Table 1.

[0012] The molecular probe combination is used to detect the SNP sites described in Table 1, and the probes are designed according to the SNP sites described in Table 1 and their preceding and following sequences (60 bp).

[0013] A SNP chip for distinguishing five types of Hubei native yellow cattle (Yiling cattle, Dabieshan cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle), wherein the gene chip is loaded with a molecular probe combination for detecting the SNP sites described in Table 1.

[0014] A method for distinguishing Yiling cattle, Dabieshan cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle, comprising the following steps:

[0015] (1) Obtaining the SNP site genotype of the tested cattle, wherein the SNP site is shown in Table 1;

[0016] (2) Perform cluster analysis or principal component analysis on the SNP locus genotypes of the cattle to be tested and the known genotypes of Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle, and determine the breed of the cattle to be tested based on the analysis results.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] 1. The five Hubei cattle breeds involved in the present invention, namely Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle and Sanjiaoshan cattle, include the main Hubei cattle breeds, are of great representativeness for the selection and breeding of Hubei cattle, and have greater application value in the molecular breeding of Hubei cattle.

[0019] 2. When selecting SNP variant sites, the present invention uses three selection signals, Tajima's D, FST, and πratio, in addition to the conventional single selection signal, to further increase the accuracy of the selected sites. At the same time, the selected sites are scored by random forest, which further reduces the number of sites required. At the same time, the classification effect is better, lower sequencing costs are achieved, and the most representative sites are found.

[0020] 3. The present invention can provide a good research foundation and data support for the germplasm identification, selective breeding and other fields of Hubei Yellow Cattle, and further reduce the cost of cattle genome selection, accelerate the genetic progress of my country's high-quality cattle breed improvement, and has good social value and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 : Flowchart of genomic SNP combination screening for distinguishing five native Hubei yellow cattle.

[0022] Figure 2 : Line graph of the number of SNP sites and the accuracy of variety differentiation in the random forest model.

[0023] Figure 3:Three-dimensional PCA analysis results of different SNP loci combinations on five Hubei native yellow cattle. Figure A shows the PCA analysis results of 17,579,873 loci obtained based on sequencing data alignment, and Figure B shows the PCA analysis results of 285 SNP loci screened based on selection signal analysis and random forest method.

[0024] Figure 4 :The cluster analysis results of different SNP loci combinations on five Hubei native yellow cattle. Figure A shows the cluster analysis results of 17,579,873 loci obtained based on sequencing data alignment, and Figure B shows the cluster analysis results of 285 SNP loci screened based on selection signal analysis and random forest method. DETAILED DESCRIPTION

[0025] In order to better illustrate the purpose, technical solutions and advantages of the present invention, the present invention will be described in detail below in conjunction with specific embodiments.

[0026] The present invention screened out SNP loci combinations for identification of five Hubei native yellow cattle breeds based on random forest and selection signal methods. The specific screening and application are as follows:

[0027] S1. Collection of cattle second-generation whole genome sequencing data

[0028] Liver tissues from 63 Hubei cattle of five breeds (including Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle) were collected and sent to BGI for second-generation whole-genome sequencing to obtain double-end sequencing fastq data.

[0029] S2. Alignment of sequencing data

[0030] The sequencing data fastq format files were aligned to the reference genome ARS-UCD1.2 of cattle (https: / / www.ncbi.nlm.nih.gov / genome / ?term=txid9913) using the mem algorithm of the bwa software and default parameters. The bam files obtained by alignment were sorted according to chromosome position using SAMtools, and the default parameters of the SAMtools software were used to remove redundancy and create an index.

[0031] S3. Construct a variation map and perform quality control on it

[0032] S31. Using the bam file preprocessed by S2, the sentieon driver parameter "--algoHaplotyper" was used to identify SNP variant sites. A gvcf file was obtained for each individual, and a total of 63 gvcf files and corresponding indexes were obtained.

[0033] S32. Use the sentieon driver parameter "--algo GVCFtyper" to merge all the gvcf files obtained in S31 to obtain a vcf file containing the SNPs and InDels of 63 Hubei Yellow Cattle.

[0034] S33. Use GATK to perform hard filtering on the obtained vcf file with the filtering parameters of "QUAL<30.0||QD<2.0||MQ<40.0||FS>60.0||SOR>3.0||MQRankSum<-12.5||ReadPosRankSum<-8.0\" to screen the corresponding sites. This parameter mainly screens the quality value of site variation, the quality-to-depth ratio of variation, mapping quality, and mapping quality ranking. The main function of these parameters is to filter false positive variations and improve the credibility of variation detection.

[0035] S34. Perform quality control on the vcf file and use GATK's SelectVariants to remove the InDels. The main object of our analysis is the SNP site.

[0036] S35. Use SelectVariants in GATK with the parameters set to “--exclude-filtered --restrict-alleles-to-BIALLELIC” to retain only biallelic loci, and finally obtain 133,388,567 loci.

[0037] S36. Use Plink to filter the data by site missing rate and minimum allele frequency. The specific parameters are "--geno 0.05--maf 0.05". This step reduces invalid sites to ensure the validity of the analysis data, and 17,579,873 sites were obtained.

[0038] S4. Use three selection signals and take their intersection and then perform LD pruning to select sites

[0039] S41. Use VCFtools to input the vcf file obtained in S36, and then input the grouping data of the five Hubei cattle breeds that we need to distinguish, and calculate the three selection signals: Tajima's D (as a selection signal within a group, the variation of the group genome is evaluated by nucleotide diversity and haplotype frequency of the sliding window), FST (group divergence index, an indicator to measure genetic differentiation between different groups. Its main principle is to compare the genetic variation within a group with the genetic variation between groups), and πratio (nucleotide diversity ratio, which is used to measure the genetic diversity within a group. It is calculated as the average nucleotide difference between each pair of alleles. This ratio can be used to compare the diversity levels of different groups or different genomic regions).

[0040] S42. The results of the three selection signals were sorted. FST, Tajima's D and πratio were arranged in descending order and the first 200 sites were taken respectively. The sites obtained by the three selection signals were taken as a union, and the duplicate sites were removed to obtain 3978 SNPs. Taking the union of the three selection signals can retain the results of the three selection signals to the greatest extent.

[0041] S43. The data obtained in S42 were trimmed by LD using Plink with the parameter "--indep-pairwise 500500.5", and 763 loci were finally obtained. LD trimming can remove loci with LD linkage, ensure the simplicity of loci, and further improve the accuracy of loci.

[0042] S5. Scoring loci using the random forest model and calculating the minimum number of loci required to identify the five Hubei cattle breeds based on the loci scores

[0043] S51. The 763 loci obtained in S43 were input into the random forest model and scored to obtain the gini and accuracy scores of the 763 loci. The gini model had a higher locus score, and the gini score was used to sort the loci in subsequent analysis.

[0044] S52. The loci of the gini model were arranged in descending order according to the scores. The selected loci combinations were increased from 5 to 5 each time. The individuals of each variety were randomly divided into a test set and a training set according to the variety. The ratio of the number of individuals in the test set: the training set was 7:3. At the same time, the variety information corresponding to the individuals was input into the random forest model. As the number of loci increased, the identification accuracy of the five varieties gradually changed ( Figure 2 ), and finally found the minimum combination of loci that can make the identification accuracy of the five breeds reach 100% and stable. The minimum combination of loci obtained includes 285 loci. The specific loci information is shown in Table 1 (the physical location of the loci is determined based on the ARS-UCD1.2 version of the cattle reference genome).

[0045] Table 1 SNP loci combinations used to distinguish five Hubei native yellow cattle

[0046]

[0047]

[0048]

[0049]

[0050] S6. Verify the accuracy of SNP site combination

[0051] S61. The 285 loci obtained in S52 were extracted and used to test the identification effect of 63 samples of five Hubei native cattle breeds using 3D PCA. Plink "--pca 10" was used to calculate the PCA value, and the R package scatterplot3d was used to draw a 3D PCA graph based on PC1, PC2, and PC3. At the same time, in order to compare with the results, the data obtained in S3 (17579873 loci) were used to draw a 3D PCA graph using the same method.

[0052] Figure 3 A is the PCA analysis result of 17579873 sites, Figure 3 B is the PCA analysis result of the 285 sites screened. Figure 3 Compared with A, Figure 3 The clustering effect of the five breeds in B is more significant. The 285 loci selected by three selection signals and random forest scoring can effectively distinguish the five Hubei Yellow Cattle breeds.

[0053] S62. The 285 loci obtained in S52 were extracted and tested for identification of 63 samples of five Hubei cattle breeds using the clustering method. The distance matrix was calculated using VCF2Dis, and the fastme website (http: / / www.atgc-montpellier.fr / fastme / ) was used to convert the data into a tre format file, which was then input into the ITOL website (https: / / itol.embl.de / tree) for visualization. At the same time, in order to compare with the results, the data obtained in S3 (17,579,873 loci) were used to draw a clustering diagram using the same method.

[0054] Figure 4 A is the cluster analysis result of 17,579,873 sites. Figure 4 B is the cluster analysis result of 285 sites screened. Figure 4 Compared with A, Figure 4The genetic distance between the five breeds in B is farther, and the differentiation is more obvious. The 285 loci selected by three selection signals and random forest scoring can increase the differentiation of the five Hubei Yellow Cattle breeds.

[0055] In summary, according to the results of three-dimensional PCA analysis and cluster analysis, the 285 loci obtained by three selection signals and random forest can effectively distinguish the five Hubei native yellow cattle breeds, namely Yiling cattle, Dabieshan cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle. According to the SNP loci and their preceding and following sequences (60 bp) described in Table 1, probes were designed, and SNP chips for distinguishing the five Hubei native yellow cattle could be further developed.

Claims

1. The application of SNP locus combination in distinguishing five types of Hubei native yellow cattle is characterized by: The information of the SNP loci is shown in Table 1. The five native Hubei cattle include Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle.

2. A molecular probe combination, characterized in that: The molecular probe combination detects the SNP sites described in Table 1.

3. Use of the molecular probe combination according to claim 2 in distinguishing Yiling cattle, Dabieshan cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle.

4. A SNP chip for distinguishing five types of Hubei native yellow cattle, characterized by: The gene chip is loaded with the molecular probe combination according to claim 2.

5. A method for distinguishing Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle, characterized in that: The following steps are involved: (1) Obtaining the SNP locus genotype of the tested cattle, wherein the SNP locus is shown in Table 1; (2) Perform cluster analysis or principal component analysis on the SNP genotypes of the cattle to be tested and the genotypes of the known Yiling cattle, Dabie Mountain cattle, Yunba cattle, Enshi cattle, and Sanjiaoshan cattle, and determine the breed of the cattle to be tested based on the analysis results.

Images