Abstract

Young bulls and heifers are selected based on genomic predictions in the US dairy population. The genomic pre-selection on young animals may bias down the traditional predicted transmitting ability (PTA) with BLUP whereas single-step genomic BLUP (ssGBLUP) is expected to account for the pre-selection. The objectives of this study were to show the pre-selection bias in PTA compared with genomically-enhanced PTA from ssGBLUP (ssGPTA) in US Holstein and to discuss a possible impact of that pre-selection in a validation study. Three production traits, 305-d milk, fat, and protein yield were considered. The data included more than 50 million phenotypes per trait from 22 million cows, 30 million pedigree animals, and 764 thousand genotyped animals. Computing with ssGBLUP took 18 h as opposed to 6 h with BLUP. The genetic trends of the traditional PTA were less than the trends of ssGPTA in genotyped bulls and cows. Both genotyped bulls and cows are pre-selected by genomic predictions. Therefore, the traditional PTA biases down, and ssGBLUP likely provides more reasonable evaluations. In a validation study, the traditional PTA, yield deviation, and daughter yield deviation are no longer suitable as pseudo-phenotypes, and the benchmarks should be calculated based on genomically-enhanced PTA from ssGBLUP. Additional studies in ssGBLUP indicated possibilities of inflation and biases especially with incomplete pedigree and foreign animals. Remedies include pedigree truncation, accounting for nonzero inbreeding of unknown parents, and separate unknown parent groups for genotyped animals. After careful modelling, ssGBLUP seems capable of providing genomic evaluations without pre-selection bias and inflation in a reasonable computing time. Keywords: bias, computation, genetic trend, genomic prediction

Yutaka Masuda, Ignacy Misztal, Paul VanRaden, Thomas Lawlor

Proceedings of the World Congress on Genetics Applied to Livestock Production, Volume Biology & Species - Bovine (dairy) 1, , 540, 2018
Download Full PDF BibTEX Citation Endnote Citation Search the Proceedings



Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.