Review: How to improve genomic predictions in small dairy cattle populations

Abstract

This paper reviews strategies and methods to improve accuracies of genomic predictions from the perspective of a numerically small population. Improvements are realized by influencing one or both of the main factors: (1) improve or increase genomic connections to phenotypic records in training data. (2) Models and strategies to focus genomic predictions on markers closer to the causative variants. Combining populations into a joint reference population results in high improvements when combining populations of the same breed and diminishes as the genetic distance between populations increases. For distantly related breeds sophisticated Bayesian variable selection models in combination with denser markers sets or functional subsets of markers is needed. This is expected to be further improved by the efficient use of sequence information. In addition predictions can be improved by the use of phenotypes of genotyped and non-genotyped cows directly. For a small population the optimal approach will combine the above components.

Genomic prediction using preselected DNA variants from a GWAS with whole-genome sequence data in Holstein–Friesian cattle

“…This confirms the expectation that, within the Holstein population, it is probably difficult to increase the accuracy of prediction, due to the small size of the effective population and the long-range LD in the population, as was previously demonstrated in a simulation study by MacLeod [30]. Hence, using ISQ and selected variants might be especially beneficial for across-breed prediction in small populations [31] or when traits are used with some QTL having large effects, as for fat percentage [32]. …”

The effectiveness of genomic selection for milk production traits of Holstein dairy cattle

“…Wiggans et al [26] reported that during the genomic selection of cattle conducted in 2011 in the USA, the reliability of the selection of milk yield increased by 34.0% over the parent average, and that of fat and protein yields increased by 33.8% and 24.9%, respectively, indicating that reliabilities can be increased even more than those we obtained in our study. The smaller improvements we found might have been due to the relatively very small reference population we used [29,30]. When genomic selection is applied in the selection of dairy cattle in the domestic population, the size of the reference population will increase continuously and potentially result in greater improvements, but this will take time.…”

Multi-breed genomic prediction using Bayes R with sequence data and dropping variants with a small effect

“…This is likely because the Red Holstein is much more closely related to the Holstein individuals in the reference populations. Because LD is conserved over much shorter distances across breeds than within breeds, sequence data is thought to be especially beneficial for multi-breed and across-breed prediction [ 2 ].…”

“…In dairy cattle, LD is extensive within a breed but the phase of LD varies between breeds [ 1 ], which is expected to decrease across-breed prediction. Use of sequence data is expected to increase the accuracy of multi-breed and across-breed prediction, which would be beneficial for breeds with small reference population sizes [ 2 ].…”