Modeling missing pedigree in single-step genomic BLUP

摘要

The objective was to compare methods of modelingmissing pedigree in single-step genomic BLUP (ssGBLUP).Options for modeling missing pedigree includedignoring the missing pedigree, unknown parent groups(UPG) based on A (the numerator relationship matrix)or H (the unified pedigree and genomic relationshipmatrix), and metafounders. The assumptions for thedistribution of estimated breeding values changed withthe different models. We simulated data with heritabilitiesof 0.3 and 0.1 for dairy cattle populations thathad more missing pedigrees for animals of lesser geneticmerit. Predictions for the youngest generation and UPGsolutions were compared with the true values for validation.For both traits, ssGBLUP with metafoundersprovided accurate and unbiased predictions for younganimals while also appropriately accounting for genetictrend. Accuracy was least and bias was greatest for ss-GBLUP with UPG for H for the trait with heritabilityof 0.3 and with UPG for A for the trait with heritabilityof 0.1. For the trait with heritability of 0.1 and UPGfor H, the UPG accuracy (SD) was −0.49 (0.12), suggestingpoor estimates of genetic trend despite havinglittle bias for validations on young, genotyped animals.Problems with UPG estimates were likely caused bythe lesser amount of information available for the lowerheritability trait. Hence, UPG need to be defined differentlybased on the trait and amount of information.More research is needed to investigate accounting forUPG in A_(22) to better account for missing pedigrees forgenotyped animals.