Improving genomic prediction accuracy of pig reproductive traits based on genotype imputation using pre-selected markers with different imputation platforms

Abstract

Genomic prediction has been widely applied to the pig industry and has greatly accelerated the progress of genetic improvement in pigs. With the development of sequencing technology and price reduction, more and more genotype imputation panels of pig have been investigated, providing an effective and economical method to further study the genetic variation of pig economic traits. In this study, the imputation from 80k Single Nucleotide Polymorphism (SNP) chip data of 832 Large White pigs to WGS genotypes was performed by Swine Imputation Server (SWIM), Pig Haplotypes Reference Panel (PHARP), Animal Genotype Imputation Database (AGIDB) and 1k-pig-genomes (1k_pig) four thousand-pig imputation panels. Then linkage disequilibrium (LD) pruning and genome-wide association study (GWAS) preselected markers strategies were utilized to compare the genomic prediction accuracy of the different imputation data for reproductive traits, respectively. Our results showed that the PHARP panel exhibited the best genomic prediction accuracy among the four imputation panels. Meanwhile, the genomic prediction accuracy of the imputation data can be further improved by utilizing the LD pruning and GWAS preselected marker strategies. In conclusion, our study provides insights into imputation data for pig genetic breeding.