The prediction of genetic structure of East African smallholder dairy cattle

Abstract

Identification of genetic structure and estimation of individual breed proportions in livestock species based on molecular data have become important tools in improvement of breeding programs in the developed world. In this study, we have applied high density SNP assays to understand genetic structure and breed composition of a developing world smallholder dairy system, which does not have pedigree records. Approximately 2000 East African smallholder crossbred dairy cattle were genotyped for 770k SNP. Principal component analysis (PCA) was used to identify the genetic structure and the ADMIXTURE program was used to estimate the proportional breed composition of individual animals. Genomic prediction and farmer prediction for breed proportions were compared. PCA revealed that the range of breed composition of smallholder dairy cattle is much larger than commonly believed. The correlation between breed proportion estimated using ADMIXTURE and the farmer's assessment of breed proportion was only 0.4, revealing that in this system farmers do not have a good understanding of the breed composition of their animals. This will be a problem to be overcome if farmers are to make optimum breeding choices to produce replacement heifers.