Whole genome investigation of the genetic structure of South African sheep breeds

Abstract

In this study whole-genome SNP data from 1977 animals, originating from thirteen sheep breeds in South Africa was analysed. The animals were classified into three production types: dual-purpose (Afrino, Dohne Merino, Dormer, Merino, SA Mutton Merino), meat (Black Headed Persian, Damara, Dorper, Meatmaster, White Dorper) and indigenous (Fat-tailed, Namaqua Afrikaner, Pedi). All animals were genotyped using the Illumina Ovine 50K SNP BeadChip. The aim was to investigate the genetic diversity and inbreeding levels of both indigenous and commercial sheep breeds in South Africa. The populations were first investigated individually to analyse within population diversity, whereafter they were merged in order to perform between population diversity analysis. During sample-based quality control, a total of 207 animals were removed due to low call rates. The number of SNPs remaining after marker-based quality control ranged from 32 422 to 44 778. Average observed heterozygosity values of 0.360, 0.355 and 0.340 were observed for dual-purpose, meat and indigenous populations respectively. Average linkage disequilibrium (LD) estimates were highest for dual-purpose populations at 0.277, followed by meat type populations at 0.259, and the lowest LD was observed for indigenous populations at 0.255. Dual-purpose, meat and indigenous populations showed average minor allele frequencies (MAF) of 0.393, 0.427 and 0.444 respectively. Principal component analysis (PCA) and ADMIXTURE results showed a clear differentiation between the dual-purpose populations, and the meat and indigenous populations. Average inbreeding coefficient (Fis) estimates of 0.008, -0.029 and 0.003 was observed for dual-purpose, meat and indigenous populations respectively. All populations showed a decline in effective population size (Ne) across generations. For all populations, the highest percentage of ROH was found in the shortest length category (1-3.99 Mb), and the largest number of ROH were observed on the first ten chromosomes. The average FROH estimates was highest for dual-purpose populations at 0.015, followed by the meat type and indigenous populations at 0.011 and 0.005 respectively. An unrooted phylogenetic tree based on pairwise Fst estimates showed a divergence of the dual-purpose populations from the meat and indigenous populations. Results from this study indicated lower variation within dual-purpose populations, and higher variation within indigenous populations. High population differentiation was observed between the various production types.