期刊
TROPICAL ANIMAL HEALTH AND PRODUCTION
卷 54, 期 6, 页码 -出版社
SPRINGER
DOI: 10.1007/s11250-022-03340-2
关键词
Genomic selection; Prediction bias; Selective genotyping; Shallow pedigree; Simulation; Single-step GBLUP
This study aimed to propose a genomic prediction strategy for predicting sex-limited traits GEBV in shallow pedigree scenarios, and the results showed that ssGBLUPF can achieve more accurate and unbiased GEBV estimates. For populations with shallow pedigree, using ssGBLUPF can lead to desirable outcomes when resources are limited.
In unstructured dairy programs, pedigree is usually shallow, which leads to biased prediction of breeding values using best linear unbiased prediction (BLUP). The objective of this study was to come out with a genomic prediction strategy that can utilize shallow pedigree information and predict unbiased and more accurate GEBV for sex-limited traits in a small population using single-step GBLUP (ssGBLUP). The data and models for a population under selection were simulated. Out of current 10 generations, 10th generation with 1000 candidates served as validation population. For the complete pedigree scenario, pedigree (P)BLUP estimated breeding values (EBV) were unbiased with accuracy (r) of 0.35 +/- 0.02 and 0.26 +/- 0.01 for 0.3 and 0.1 h(2) scenario, respectively. For the shallow pedigree, biased prediction of breeding values and low accuracies were obtained with linear decline in the accuracy of EBV for removal of information on more distant pedigree. Accuracy and bias (rho) for scenario with removing 4 distant generations from pedigree were 0.30 +/- 0.02 and 0.55 +/- 0.03, respectively, in moderate h(2) scenario. Use of Genomic (G)BLUP, especially with extreme phenotypic contrast selective genotyping, (TB) resulted in higher accuracy for a small reference of females; however, GEBV were highly biased. We observed that ssGBLUPF, where the numerator relationship matrix is corrected for inbreeding, resulted in more accurate and unbiased estimates of GEBV across shallow pedigree scenario, with TB all female reference (missing 4 distant generations: r = 0.50 +/- 0.02; rho = 0.96 +/- 0.02). We recommend use of ssGBLUPF with two tailed selectively genotyped all female reference in shallow pedigree scenarios, to obtain unbiased and accurate GEBV for sex-limited traits, when resources are limited.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据