Machine learning-based pig litter size character genome phenotypic value prediction method
Through the prediction method of genomic phenotype value of pig birth traits based on machine learning, the problem of small local pig breeding population and inaccurate early genetic assessment was solved, and a prediction model of birth litter weight, birth number and living litter number was constructed. Improved prediction accuracy and is suitable for breeding selection of small groups of local pigs.
Patent Information
- Application Number
- CN202411116021.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-08-14
AI Technical Summary
The problem of large differences in early estimated breeding values and later actual phenotypes caused by the small size of local pig breeding population and the low accuracy of early genetic assessment.
Using machine learning-based genomic phenotype value prediction method for pig birth traits, we collected the reproductive trait data of sows, performed linear model correction, SNP genotyping, whole genome association analysis and machine learning model construction to construct birth litter weight, Prediction model of litter count and live litter count to improve prediction accuracy.
It improves the prediction accuracy of local pig breeding traits, reduces the difference between early estimates and later actual phenotypes, and is suitable for breeding selection of small groups of local pigs.
Smart Images

Figure CN119049549A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of local breed pig production trait prediction, and in particular relates to a method for predicting genomic phenotype values of pig litter traits based on machine learning. Background Art
[0002] With the development of animal husbandry, pig production trait prediction technology has become an important means to improve breeding efficiency and optimize breed selection. These technologies not only rely on traditional genetics knowledge, but also integrate the latest achievements in modern biotechnology, big data analysis and artificial intelligence.
[0003] The prediction of reproductive traits of local sows is a multidisciplinary research field based on genetics, nutrition, environmental science, information technology, etc. With the rapid development of modern biotechnology, especially the rise of molecular genetics, genomics and bioinformatics, it provides strong technical support for the accurate prediction of reproductive traits of local sows.
[0004] Genetic basis:
[0005] Genetic markers and linkage analysis: Molecular marker technology (such as SNP, SSR, etc.) is used to locate and conduct linkage analysis on the genetic genes of sows to reveal the relationship between reproductive traits and genetic genes.
[0006] Genome-wide association study (GWAS): Through large-scale genome sequencing, genetic variation sites that are significantly associated with reproductive traits are discovered, providing a genetic basis for trait prediction.
[0007] Nutrition and Environmental Sciences:
[0008] Nutritional regulation: Study the effects of different nutrients on the reproductive performance of sows, and improve the reproductive efficiency and production performance of sows by optimizing feed formula and feeding management.
[0009] Environmental stress management: Analyze the impact of environmental factors (such as temperature, humidity, light, etc.) on the reproductive performance of sows, propose corresponding environmental control measures, reduce stress responses, and improve reproductive success rates.
[0010] information Technology:
[0011] Big data analysis: Collect multi-source information such as sow reproductive performance data, genetic information, nutrition and environmental parameters, use big data technology for in-depth mining and analysis, and establish a reproductive trait prediction model.
[0012] Artificial Intelligence and Machine Learning: Use artificial intelligence algorithms (such as neural networks, decision trees, etc.) to predict reproductive traits and improve the accuracy and efficiency of predictions.
[0013] The application prospect of the method for predicting the reproductive traits of local sows is broad. On the one hand, it can provide a scientific basis for the selection and mating of sows for breeding farms, and improve the production performance and reproductive efficiency of breeding pigs; on the other hand, it can provide personalized feeding and management plans for pig farms, optimize feed formula and feeding environment, reduce production costs, and improve economic benefits.
[0014] With the continuous development of biotechnology and information technology, the prediction method of local sow reproductive traits will become more accurate and efficient. In the future, through the cross-integration and collaborative innovation of multiple disciplines, it is expected to promote the local sow breeding industry to a higher level. Summary of the invention
[0015] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for predicting genomic phenotypic values of pig litter traits based on machine learning, which solves the problem of large differences between early estimated breeding values and later actual phenotypes due to the small size of local pig breeding populations and low accuracy of early genetic evaluation.
[0016] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a method for predicting genomic phenotypic values of pig litter traits based on machine learning, comprising the following steps:
[0017] S1. Determine local pig breeds, and based on the local pig breeds, collect the reproductive traits of several sows to obtain an original phenotypic data set; the reproductive traits include birth litter weight, litter size, and live piglets;
[0018] S2. Use the linear model to correct the original phenotypic data set and remove abnormal phenotypes to obtain the target phenotypic data set;
[0019] S3, extracting the genomic DNA of the sow corresponding to each target phenotype data in the target phenotype data set;
[0020] S4, performing SNP genotyping on each genomic DNA to obtain a genotyping data set;
[0021] S5, performing genotype filling on the genotyping data set to obtain a SNP data set;
[0022] S6. Perform GWAS genome-wide association analysis on the SNP data set to obtain a SNP training data set;
[0023] S7. Based on the SNP training data set and the target phenotype data set, a phenotypic sample set is obtained, and a reproductive trait prediction model is constructed according to machine learning to determine the sows to be tested. According to the genotype data of the sows to be tested, the reproductive trait prediction model is used to obtain the genomic phenotypic prediction value of the littering trait.
[0024] Furthermore, the step S2 is specifically as follows:
[0025] S201, correcting the original phenotypic data in the original phenotypic data set using a linear model; the expression of the linear model is:
[0026] y=xb+e
[0027] Among them, y is the original phenotypic data; x is the association matrix of b; b is the fixed effect vector; e is the residual vector;
[0028] S202, normalizing the residual vector, and eliminating abnormal phenotypic data based on the normalized residual vector and the principle of three times the standard deviation to obtain a target phenotypic data set.
[0029] Furthermore, the step S4 specifically includes performing SNP genotyping on each genomic DNA to obtain an initial genotyping data set, and using Plink software to perform quality control on the initial genotyping data set to obtain a genotyping data set; the quality control specifically includes: eliminating sites without chromosomal position information, sites on sex chromosomes, sites with a minimum allele frequency of less than 0.01, sites with an individual SNP detection rate of less than 90%, and sites with a Hardy-Weinberg equilibrium test P<10 -6 The location of .
[0030] Furthermore, the step S5 is specifically as follows:
[0031] S501, obtaining a local pig collection to be sequenced;
[0032] S502. Perform whole genome resequencing on the local pigs to be sequenced, and obtain several groups with MAF ≥ 0.05 and missing rate < 10 -6 , sequencing depth ≥ 6 and resequencing data with SNPs located on autosomes;
[0033] S503, dividing each resequencing data into a reference genotype data set and a comparison genotype data set;
[0034] S504, using the resequencing data in the reference genotype data set as a reference template, performing genotype filling on the genotyping data in the genotyping data set to obtain an initial filling result;
[0035] S505, using the genotype data in the comparison genotype data set as a reference to compare with the sites of the initial filling result, calculate the filling accuracy of each SNP:
[0036]
[0037] Among them, Q c is the filling accuracy of SNP; R SNP is the number of individuals filled correctly; N SNP is the total number of individuals at the locus;
[0038] S506. Retain SNPs with a filling accuracy of 1 and a MAF ≥ 0.01 to obtain a SNP data set.
[0039] Furthermore, the step S6 is specifically as follows:
[0040] S601. Based on the SNP dataset and the target phenotype dataset, a number of principal components are obtained by principal component analysis, each principal component is used as a covariate for GWAS genome-wide association analysis, and GWAS genome-wide association analysis is performed based on the SNP dataset and the target phenotype dataset to obtain the GWAS genome-wide association analysis results;
[0041] S602. Based on the results of GWAS genome-wide association analysis, select -5 SNPs are used as the first weighted locus set;
[0042] S603, selecting loci related to target phenotypic data according to the pig QTL database as a second weighted locus set;
[0043] S604, using GEMMA software to calculate the SNP explanation percentage of each site in the first weighted site set and the second weighted site set, and determining the weight of each weighted site according to the explanation percentage of each weighted site;
[0044] S605 . Obtain a SNP training data set according to the first weighted site set, the second weighted site set and the weight of each weighted site.
[0045] Furthermore, the step S7 is specifically as follows:
[0046] S701, making one-to-one correspondence between the SNP training data set and the target phenotype data set, eliminating the target phenotype data without corresponding relationship, and obtaining a phenotype sample set;
[0047] S702. Based on the phenotypic sample set, a reproductive trait prediction model is constructed using machine learning based on the three reproductive traits; specifically, the reproductive trait prediction model includes a birth litter weight prediction model, a litter size prediction model, and a live piglet number prediction model;
[0048] S703, obtaining the first weighted locus set, the second weighted locus set and the weight of each weighted locus of the sow to be tested, and obtaining the genomic phenotypic prediction value of the litter weight prediction model, the litter size prediction model and the live piglet number prediction model.
[0049] Further, the step S702 is specifically as follows:
[0050] S7021, using litter weight, litter size and live piglets as labels for the phenotypic sample sets as the first training sample set, the second training sample set and the third training sample set based on the three reproductive traits;
[0051] S7022, obtaining a current training sample set, taking the reproductive traits in the current training sample set as prediction targets, and the values of the reproductive traits in the current training sample set as actual target values;
[0052] S7023. According to the current training sample set, a gradient boosting decision tree is constructed to obtain a reproductive trait prediction model;
[0053] S7024, executing steps S7022 to S7023 for the first training sample set, the second training sample set, and the third training sample set, respectively, to obtain a birth litter weight prediction model, a litter size prediction model, and a live piglet number prediction model accordingly.
[0054] Furthermore, the expression of the reproductive trait prediction model in step S7023 is:
[0055]
[0056] Among them, F(x,w) is the predicted value of reproductive traits; x is the input sample, w is the model parameter; M is the total number of decision trees; α m is the weight of the mth decision tree; h m is the mth decision tree; w m is the parameter of the mth decision tree; f m (x,w m ) is the result of the mth decision tree.
[0057] The beneficial effects of the present invention are as follows: the present invention solves the problem of large differences between early estimated breeding values and later actual phenotypes caused by small scale of local pig breeding population and low accuracy of early genetic evaluation, thereby improving the prediction accuracy of the reproductive traits of local pigs. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] Figure 1 The present invention is a flow chart of the method.
[0059] Figure 2 This is a schematic diagram of the accuracy of phenotype prediction under different SNP densities in an embodiment of the present invention.
[0060] Figure 3 This is a correlation test diagram between the phenotypic prediction results of the GBDT_PP method in Rongchang pigs and the true phenotypic values in an embodiment of the present invention. DETAILED DESCRIPTION
[0061] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.
[0062] like Figure 1 As shown, in one embodiment of the present invention, a method for predicting genomic phenotypic values of pig litter traits based on machine learning comprises the following steps:
[0063] S1. Determine local pig breeds, and based on the local pig breeds, collect the reproductive traits of several sows to obtain an original phenotypic data set; the reproductive traits include birth litter weight, litter size, and live piglets;
[0064] S2. Use the linear model to correct the original phenotypic data set and remove abnormal phenotypes to obtain the target phenotypic data set;
[0065] S3, extracting the genomic DNA of the sow corresponding to each target phenotype data in the target phenotype data set;
[0066] S4, performing SNP genotyping on each genomic DNA to obtain a genotyping data set;
[0067] S5, performing genotype filling on the genotyping data set to obtain a SNP data set;
[0068] S6. Perform GWAS genome-wide association analysis on the SNP data set to obtain a SNP training data set;
[0069] S7. Based on the SNP training data set and the target phenotype data set, a phenotypic sample set is obtained, and a reproductive trait prediction model is constructed according to machine learning to determine the sows to be tested. According to the genotype data of the sows to be tested, the reproductive trait prediction model is used to obtain the genomic phenotypic prediction value of the littering trait.
[0070] In this embodiment, in order to solve the problem that the early estimated breeding value and the actual phenotype in the later stage are greatly different due to the small scale of local pig breeding population and the low accuracy of early genetic evaluation, Rongchang pig is used as the research object, and a machine learning model is used to carry out the early prediction study of Rongchang pig reproductive traits based on whole genome resequencing, explore the influence of different SNP densities and different prediction models on genetic evaluation, and construct the optimal model for early prediction of Rongchang pig reproductive traits based on machine learning, so as to provide data support for the genome collaborative selection technology with estimated breeding value and phenotypic prediction value as the selection criteria. For small groups of local pigs, an optimal model for phenotypic prediction of pig reproductive traits based on whole genome information is constructed, which is a method for genomic phenotypic prediction of reproductive traits based on gradient boosting decision tree (Gradient Boosting Decision Tree for Phenotypic Prediction, GBDT_PP), which has the advantages of high prediction accuracy and fast calculation speed, and has excellent prediction effect in small groups of local pigs.
[0071] The step S2 is specifically as follows:
[0072] S201, correcting the original phenotypic data in the original phenotypic data set using a linear model; the expression of the linear model is:
[0073] y=xb+e
[0074] Among them, y is the original phenotypic data; x is the association matrix of b; b is the fixed effect vector; e is the residual vector;
[0075] S202, normalizing the residual vector, and eliminating abnormal phenotypic data based on the normalized residual vector and the principle of three times the standard deviation to obtain a target phenotypic data set.
[0076] In this embodiment, the Rongchang pig breeding core group of the Rongchang pig national core breeding farm was selected for this experiment. The reproductive traits (birth litter weight, litter size, live piglets) of 550 sows were collected, and ear tissue samples were collected and frozen in 75% alcohol for storage. To avoid the influence of other factors, a linear model was used to correct the recorded phenotype. The correction model is as follows:
[0077] y=xb+e
[0078] Among them, b is the fixed effect vector, including factors to be eliminated such as production year, month, and parity, and e is the residual vector. After calculating e, it was normalized, and after eliminating 35 abnormal phenotypic data outside 3 times the standard deviation according to the 3 times standard deviation principle, 515 Rongchang pigs were retained for the label data used for subsequent machine learning model training.
[0079] The step S4 specifically includes performing SNP genotyping on each genomic DNA to obtain an initial genotyping data set, and using Plink software to perform quality control on the initial genotyping data set to obtain a genotyping data set; the quality control specifically includes: eliminating sites without chromosomal location information, sites on sex chromosomes, sites with a minimum allele frequency of less than 0.01, sites with an individual SNP detection rate of less than 90%, and sites with a Hardy-Weinberg equilibrium test P<10 -6 The location of .
[0080] In this example, genomic DNA from pig ear tissue samples was extracted using chloroform extraction, and DNA quality was tested using an ultraviolet spectrophotometer (Nanodrop2000) and 1% agarose gel electrophoresis. -1 , 1.8≤OD 260nm / OD 280nm If the value is ≤2.0, the bands are clear and there is no tailing or degradation, the DNA quality inspection is qualified. Otherwise, re-prepare the DNA until it passes the quality inspection.
[0081] The genomic DNA samples that passed the quality inspection were subjected to SNP genotyping using the "SMIC-1" chip. The genotyping data obtained were quality controlled using Plink (V1.90) software. Sites without chromosome location information and sites on the Y chromosome, sites with a minor allele frequency (MAF) less than 0.01, sites with an individual SNP detection rate less than 90%, and sites with a Hardy-Weinberg equilibrium test P < 10 -6 Finally, 31756 SNPs were retained.
[0082] The step S5 is specifically as follows:
[0083] S501, obtaining a local pig collection to be sequenced;
[0084] S502. Perform whole genome resequencing on the local pigs to be sequenced, and obtain several groups with MAF ≥ 0.05 and missing rate < 10 -6 , sequencing depth ≥ 6 and resequencing data with SNPs located on autosomes;
[0085] S503, dividing each resequencing data into a reference genotype data set and a comparison genotype data set;
[0086] S504, using the resequencing data in the reference genotype data set as a reference template, performing genotype filling on the genotyping data in the genotyping data set to obtain an initial filling result;
[0087] S505, using the genotype data in the comparison genotype data set as a reference to compare with the sites of the initial filling result, calculate the filling accuracy of each SNP:
[0088]
[0089] Among them, Q c is the filling accuracy of SNP; R SNP is the number of individuals filled correctly; N SNP is the total number of individuals at the locus;
[0090] S506. Retain SNPs with a filling accuracy of 1 and a MAF ≥ 0.01 to obtain a SNP data set.
[0091] In this embodiment, 150bp paired-end reads were used on the Wuhan platform of BGI to perform 30× whole genome sequencing on 120 Rongchang pigs. After sequencing, FastQC software was used to evaluate the quality of the original reads by minimum Phred score 20 to filter out reads contaminated by adapters and reads containing too many "N" characters (where "N" accounts for more than 10% of single reads) to obtain Clean reads. Then, clean reads were aligned with the pig reference genome (Sscrofa11.1) using BWA software with the parameter "mem-t 10-k32-M". The generated SAM file was converted to a BAM file using SAMtools. Potential PCR (polymerase chain reaction) duplications were eliminated using the markduplduplicate (marker duplicate read elimination) program in Picard version 1.119. Subsequently, SNPs (single nucleotide polymorphisms) were detected using BAM files using the GATK software of the multi-sample method. The original SNPs generated by GATK were quality controlled according to the following criteria: QualByDepth < 2.0, FisherStrand < 60.0, RMSMappingQuality < 40.0, MappingQualityRankSumTest < -12.5, ReadPosRankSumTest < -8.0. Finally, a total of 21,104,245 SNPs were retained. In subsequent analyses, MAF ≥ 0.05 and missing rate < 10 -6 , sequencing depth ≥ 6, and retaining the resequencing data of SNPs located on autosomes as reference gene data for subsequent genotype filling.
[0092] Based on the whole genome sequencing data of 120 pigs, the whole genome sequencing data of 90 Rongchang pigs were randomly selected as reference templates, and the genotyping data set obtained by Beagle was used for genotype filling. In order to evaluate the accuracy of filling, the sequencing data of the other 30 Rongchang pigs were used as references to compare with the filled sites, and the genotypes of each locus were compared, and any incorrect loci were judged to be correctly filled through this comparison process. Subsequently, the filling accuracy of the locus was calculated as the ratio of the number of correctly filled individuals to the total number of individuals at the locus, and SNPs with a filling accuracy of 1 and MAF ≥ 0.01 were retained. Therefore, a total of 8,823,367 were retained, on the one hand for subsequent GWAS analysis, that is, to construct a priori information data set; on the other hand, for machine learning model training to build the optimal machine learning model for litter trait prediction.
[0093] The step S6 is specifically as follows:
[0094] S601. Based on the SNP dataset and the target phenotype dataset, a number of principal components are obtained by principal component analysis, each principal component is used as a covariate for GWAS genome-wide association analysis, and GWAS genome-wide association analysis is performed based on the SNP dataset and the target phenotype dataset to obtain the GWAS genome-wide association analysis results;
[0095] S602. Based on the results of GWAS genome-wide association analysis, select -5 SNPs are used as the first weighted locus set;
[0096] S603, selecting loci related to target phenotypic data according to the pig QTL database as a second weighted locus set;
[0097] S604, using GEMMA software to calculate the SNP explanation percentage of each site in the first weighted site set and the second weighted site set, and determining the weight of each weighted site according to the explanation percentage of each weighted site;
[0098] S605 . Obtain a SNP training data set according to the first weighted site set, the second weighted site set and the weight of each weighted site.
[0099] In this example, plink1.9 was used to perform genome-wide association analysis (GWAS) on 515 Rongchang pigs, GCTA software was used to perform principal component analysis (PCA) of 30 populations, and EIGENSTRAT software was used to calculate the P value of each PCA, and the PCA with a P value less than 0.01 was taken as a covariate.
[0100] In addition, based on the GWAS results, sites significantly associated with the target traits were screened out as weighted site 1; based on the pig QTL database, sites associated with the target traits were screened out as weighted site 2; the GEMMA software was used to calculate the SNP explanation percentage of each site, and the weight was determined according to the explanation percentage of the weighted site.
[0101] The step S7 is specifically as follows:
[0102] S701, making one-to-one correspondence between the SNP training data set and the target phenotype data set, eliminating the target phenotype data without corresponding relationship, and obtaining a phenotype sample set;
[0103] S702. Based on the phenotypic sample set, a reproductive trait prediction model is constructed using machine learning based on the three reproductive traits; specifically, the reproductive trait prediction model includes a birth litter weight prediction model, a litter size prediction model, and a live piglet number prediction model;
[0104] S703, obtaining the first weighted locus set, the second weighted locus set and the weight of each weighted locus of the sow to be tested, and obtaining the genomic phenotypic prediction value of the litter weight prediction model, the litter size prediction model and the live piglet number prediction model.
[0105] The step S702 is specifically as follows:
[0106] S7021, using litter weight, litter size and live piglets as labels for the phenotypic sample sets as the first training sample set, the second training sample set and the third training sample set based on the three reproductive traits;
[0107] S7022, obtaining a current training sample set, taking the reproductive traits in the current training sample set as prediction targets, and the values of the reproductive traits in the current training sample set as actual target values;
[0108] S7023. According to the current training sample set, a gradient boosting decision tree is constructed to obtain a reproductive trait prediction model;
[0109] S7024, executing steps S7022 to S7023 for the first training sample set, the second training sample set, and the third training sample set, respectively, to obtain a birth litter weight prediction model, a litter size prediction model, and a live piglet number prediction model accordingly.
[0110] The expression of the reproductive trait prediction model in step S7023 is:
[0111]
[0112] Among them, F(x,w) is the predicted value of reproductive traits; x is the input sample, w is the model parameter; M is the total number of decision trees; α m is the weight of the mth decision tree; hm is the mth decision tree; w m is the parameter of the mth decision tree; f m (x,w m ) is the result of the mth decision tree.
[0113] In this example, in order to avoid the reuse of group genotype information, 500 additional Rongchang pig 50K chip data (genotyping data set) were collected, and the chip data was genotyped. 400 of them were randomly selected for machine learning model training (the remaining 100 were used for subsequent independent verification), and the correlation coefficient was calculated using a 5-fold cross-validation method, and the average value was taken as the accuracy of the evaluation model.
[0114] In addition, the remaining 100 Rongchang pigs were selected for independent verification, which simulated the actual production and used the trained model to estimate the phenotype of the candidate group. The results of the two tests were combined to evaluate the differences between different models.
[0115] The results are as follows:
[0116] The present invention uses 50K SNP chips (genotyping data sets) and Rongchang pig whole genome sequencing data for genotype filling. The results show that after genotype filling, a total of 21,104,245 SNPs can be used for analysis, with an average accuracy of 0.94. The filling accuracy was evaluated by comparing the sites between the whole genome resequencing and filling data of 30 Rongchang pigs. At the three thresholds of 0.3, 0.6, and 0.9, the filling accuracy ranges were 0.930-0.953, 0.934-0.955, and 0.967-0.977, respectively. Subsequently, after screening for 100% filling accuracy and excluding sites with MAF<0.01, a total of 8,823,367 SNP sites were obtained for subsequent analysis.
[0117] Among the three reproductive traits of Rongchang pigs, the significant loci associated with LW (birth litter weight) were mainly located on SSC12, while the loci associated with TNB (litter size) and NBA (live piglets) were mainly located on SSC14 and SSC15. The significance level was selected as P < 10 -5 The sites were taken as the first weighted site set, retaining 639, 250 and 266 sites respectively.
[0118] In addition, through the Pig QTL database, 75,933 SNPs of litter weight were screened out, which were significantly associated with litter size and live piglets, and 69,591 SNPs were significantly associated with litter weight at birth, and these loci were used as the second weighted locus set.
[0119] Taking the uniform distribution of the whole genome as the standard, 10 gene data sets of 10, 20, 30, 40, 50, 60, 70, 80, 90, and 1 million SNPs were constructed based on the obtained filling data, and the accuracy of total litter size, live born number, and birth litter weight was evaluated using 10 repeated 5-fold cross-validation.
[0120] like Figure 2 As shown in the figure, under different SNP densities, the phenotypic prediction accuracy increases with the number of sites, and the prediction accuracy shows a trend of first increasing and then decreasing, suggesting that the density of genomic data is not as high as possible. The peak values of the phenotypic prediction accuracy of the three traits all appear in the range of 80w to 90w, and the prediction accuracies are 0.581±0.085, 0.642±0.044, and 0.669±0.029, respectively, indicating that the prediction effect of reproductive traits based on the GBDT model is good. Figure 2 Weighted indicates whether to use weight.
[0121] Two training sample sets were constructed, one with weights for the sites in the first weighted site set and the second weighted site set, and one without weights. The results of training sets with different weights showed that the weighting of prior information data of reproductive traits had different effects on machine learning models in different situations, and there were large differences. However, when the number of sites was large, the phenotypic prediction accuracy fluctuated less after using prior information weighting, indicating that site weighting is beneficial to improving the stability of the model in the case of ultra-high site numbers.
[0122] Based on the machine learning method, the best performing hyperparameters in the cross-validation were used to construct and optimize the prediction model for evaluation. In the independent validation results, the prediction results of litter weight and litter size were basically consistent with the cross-validation results, with the maximum correlation coefficient being 0.998 and the minimum being 0.935, indicating that the trained machine learning model has good versatility.
[0123] The CPU nodes in the Slurm high-performance cluster are used for calculation, with an operating frequency of 2.60GHz and a memory of 192G. In the stage of machine learning model training, it is necessary to invest a lot of energy and spend a certain amount of time. The investment cost is proportional to the SNP density. After the model training is completed, the GBDT_PP model can quickly predict individual phenotypes based on the input data, with the advantages of high throughput and fast speed.
[0124] Table 1 Time consumption of each step of machine learning model training for litter weight trait at different SNP densities
[0125]
[0126] Table 2 Optimal model hyperparameters for genomic phenotypic prediction of pig litter size
[0127]
[0128] like Figure 3 As shown in the figure, after the GBDT_PP method (the present invention) was optimized and improved, 30 pig samples were collected from the pig farm to evaluate the phenotypic prediction effect. The results showed that in the three reproductive traits of litter weight, litter size, and live piglets, the prediction results of the GBDT_PP model were not significantly different from the actual phenotypes of the individuals (P>0.05), and the prediction accuracy was 0.46, 0.65, and 0.51, respectively. Overall, GBDT_PP has a good effect on the phenotypic prediction of reproductive traits in small groups of local pigs and can be applied to actual production breeding. Figure 3 Litter Weight is litter weight, TNB is litter size, NBA is live piglets.
Claims
1. A method for predicting genomic phenotypic values of pig litter traits based on machine learning, characterized in that: The following steps are involved: S1. Determine local pig breeds, and based on the local pig breeds, collect the reproductive traits of several sows to obtain an original phenotypic data set; the reproductive traits include birth litter weight, litter size, and live piglets; S2. Use the linear model to correct the original phenotypic data set and remove abnormal phenotypes to obtain the target phenotypic data set; S3, extracting the genomic DNA of the sow corresponding to each target phenotype data in the target phenotype data set; S4, performing SNP genotyping on each genomic DNA to obtain a genotyping data set; S5, performing genotype filling on the genotyping data set to obtain a SNP data set; S6. Perform GWAS genome-wide association analysis on the SNP data set to obtain a SNP training data set; S7. Based on the SNP training data set and the target phenotype data set, a phenotypic sample set is obtained, and a reproductive trait prediction model is constructed according to machine learning to determine the sows to be tested. According to the genotype data of the sows to be tested, the reproductive trait prediction model is used to obtain the genomic phenotypic prediction value of the littering trait.
2. The method for predicting genomic phenotypic values of pig litter traits based on machine learning according to claim 1, characterized in that: The step S2 is specifically as follows: S201, correcting the original phenotypic data in the original phenotypic data set using a linear model; the expression of the linear model is: y=xb+e Among them, y is the original phenotypic data; x is the association matrix of b; b is the fixed effect vector; e is the residual vector; S202, normalizing the residual vector, and eliminating abnormal phenotypic data based on the normalized residual vector and the principle of three times the standard deviation to obtain a target phenotypic data set.
3. The method for predicting genomic phenotypic values of pig litter traits based on machine learning according to claim 1, characterized in that: The step S4 specifically includes performing SNP genotyping on each genomic DNA to obtain an initial genotyping data set, and using Plink software to perform quality control on the initial genotyping data set to obtain a genotyping data set; the quality control specifically includes: eliminating sites without chromosomal location information, sites on sex chromosomes, sites with a minimum allele frequency of less than 0.01, sites with an individual SNP detection rate of less than 90%, and sites with a Hardy-Weinberg equilibrium test P<10 -6 The location.
4. The method for predicting genomic phenotypic values of pig litter traits based on machine learning according to claim 1, characterized in that: The step S5 is specifically as follows: S501, obtaining a local pig collection to be sequenced; S502. Perform whole genome resequencing on the local pigs to be sequenced, and obtain several groups with MAF ≥ 0.05 and missing rate < 10 -6 , sequencing depth ≥ 6 and resequencing data with SNPs located on autosomes; S503, dividing each resequencing data into a reference genotype data set and a comparison genotype data set; S504, using the resequencing data in the reference genotype data set as a reference template, performing genotype filling on the genotyping data in the genotyping data set to obtain an initial filling result; S505, using the genotype data in the comparison genotype data set as a reference to compare with the sites of the initial filling result, calculate the filling accuracy of each SNP: Among them, Q c is the filling accuracy of SNP; R SNP is the number of individuals filled correctly; N SNP is the total number of individuals at the locus; S506. Retain SNPs with a filling accuracy of 1 and a MAF ≥ 0.01 to obtain a SNP data set.
5. The method for predicting genomic phenotypic values of pig litter traits based on machine learning according to claim 1, characterized in that: The step S6 is specifically as follows: S601. Based on the SNP dataset and the target phenotype dataset, a number of principal components are obtained by principal component analysis, each principal component is used as a covariate for GWAS genome-wide association analysis, and GWAS genome-wide association analysis is performed based on the SNP dataset and the target phenotype dataset to obtain the GWAS genome-wide association analysis results; S602. Based on the results of GWAS genome-wide association analysis, select -5 SNPs are used as the first weighted locus set; S603, selecting loci related to target phenotypic data according to the pig QTL database as a second weighted locus set; S604, using GEMMA software to calculate the SNP explanation percentage of each site in the first weighted site set and the second weighted site set, and determining the weight of each weighted site according to the explanation percentage of each weighted site; S605 . Obtain a SNP training data set according to the first weighted site set, the second weighted site set and the weight of each weighted site.
6. The method for predicting genomic phenotypic values of pig litter traits based on machine learning according to claim 1, characterized in that: The step S7 is specifically as follows: S701, making one-to-one correspondence between the SNP training data set and the target phenotype data set, eliminating the target phenotype data without corresponding relationship, and obtaining a phenotype sample set; S702. Based on the phenotypic sample set, a reproductive trait prediction model is constructed using machine learning based on the three reproductive traits; specifically, the reproductive trait prediction model includes a birth litter weight prediction model, a litter size prediction model, and a live piglet number prediction model; S703, obtaining the first weighted locus set, the second weighted locus set and the weight of each weighted locus of the sow to be tested, and obtaining the genomic phenotypic prediction value of the litter weight prediction model, the litter size prediction model and the live piglet number prediction model.
7. The method for predicting genomic phenotypic values of pig litter traits based on machine learning according to claim 6, characterized in that: The step S702 is specifically as follows: S7021, using litter weight, litter size and live piglets as labels for the phenotypic sample sets as the first training sample set, the second training sample set and the third training sample set based on the three reproductive traits; S7022, obtaining a current training sample set, taking the reproductive traits in the current training sample set as prediction targets, and the values of the reproductive traits in the current training sample set as actual target values; S7023. According to the current training sample set, a gradient boosting decision tree is constructed to obtain a reproductive trait prediction model; S7024, executing steps S7022 to S7023 for the first training sample set, the second training sample set, and the third training sample set, respectively, to obtain a birth litter weight prediction model, a litter size prediction model, and a live piglet number prediction model accordingly.
8. The method for predicting genomic phenotypic values of pig litter traits based on machine learning according to claim 7, characterized in that: The expression of the reproductive trait prediction model in step S7023 is: Among them, F(x,w) is the predicted value of reproductive traits; x is the input sample, w is the model parameter; M is the total number of decision trees; α m is the weight of the mth decision tree; h m is the mth decision tree; w m is the parameter of the mth decision tree; f m (x,w m ) is the result of the mth decision tree.
Citation Information
Patent Citations
Genome breeding value estimation method based on high-throughput sequencing platform and application thereof
CN110564832A
Machine learning method and system for egg laying character genome selection of laying hens
CN116072226A
Molecular genetic marker related to malformation rate of boar sperms as well as acquisition method and application of molecular genetic marker
CN116574816A
Machine learning driven plant gene discovery and gene editing
CN116982113A
SNP (Single Nucleotide Polymorphism) site combination for identifying Rongchang pig variety and application of SNP site combination
CN117431325A
Cited By
Target character prediction method suitable for low-depth sequencing and whole genome selection model suitable for low-depth sequencing
CN120932742A
