Tracking the emergence of a new breed using 49,034 SNP in sheep

Abstract

Domestic animals are unique in that they have been organised into managed populations called breeds. The strength of genetic divergence between breeds may vary dependent on the age of the breed, the scenario under which it emerged and the strength of reproductive isolation it has from other breeds. In this study, we investigated the Gulf Coast Native breed of sheep to determine if it contains lines of animals that are sufficiently divergent to be considered separate breeds. Allele sharing and principal component analysis (PCA) using nearly 50,000 SNP loci revealed a clear genetic division that corresponded with membership of either the Florida or Louisiana Native lines. Subsequent analysis aimed to determine if the strength of the divergence exceeded that found between recognised breed pairs. Genotypes from 14 breeds sampled from Europe and Asia were used to obtain estimates of pair-wise population divergence measured as F(ST). The divergence separating the Florida and Louisiana Native (F(ST) = 6.2%) was approximately 50% higher than the average divergence separating breeds developed within the same region of Europe (F(ST) = 4.2%). This strongly indicated that the two Gulf Coast Native lines are sufficiently different to be considered separate breeds. PCA using small SNP sets successfully distinguished between the Florida and Louisiana Native animals, suggesting that allele frequency differences have accumulated across the genome. This is consistent with a population history involving geographic separation and genetic drift. Suggestive evidence was detected for divergence at the poll locus on sheep chromosome 10; however drift at neutral markers has been the largest contributor to the genetic separation observed. These results document the emergence of populations that can be considered separate breeds, with practical consequences for bio-conservation priorities, animal registration and the establishment of separate breed societies.

Figure 1. Allele sharing heat map of the Gulf Coast Native.

The proportion of alleles shared (A _s) between individuals was used to construct an ordered heat-map and dendogram. Cell colour represents the strength of allele sharing, where darker color indicates increasing allele sharing and relatedness between animals. As a guide, the lightest colors represent A _s values <0.65 and the darkest colors A _s values >0.80. Samples compared against themselves appear on the diagonal with the maximum A _s value of 1 and darkest colour. Two blocks were revealed that correspond to membership of either the Florida Native (indicated at right in green) or Louisiana Native (red).

Figure 2. Clustering of animals based on principal component analysis of allele sharing.

Individual animals were plotted for the first (PCA1) and second (PCA2) principal components using colours to distinguish the Louisiana Natives (red), Florida Natives (green) and Cracker (blue) lines. Four SNP sets were used to explore the effect of marker number and divergence on clustering: (A) all 49,034 SNP; (B) a random sample of 491 (1%) of SNP; (C) the top 1% of SNP ranked using F _ST; (D) the top 5% of SNP ranked using F _ST. Note that the scale for PCA1 differs between panels and that a larger proportion of the variation is captured by PCA1 using F _ST ranked SNP (C) compared with a random selection of markers (B).

Figure 3. Distribution of divergent SNP across the sheep chromosome.

The degree of divergence between the Florida and Louisiana Native populations was estimated for each SNP as F _ST. The genomic distribution of SNP with extreme values (either the top 1% or 5% of ranked values) is shown as a function of their chromosomal location. This revealed an over representation of high F _ST SNP on chromosome 10 when compared to the proportion expected.

Figure 4. Genetic relationship between Gulf Coast Native and 12 other breeds.

Individuals were clustered using PCA of allele sharing and plotted for PCA1 and 2 (A) and PCA1 and 6 (B). Individuals from different breeds are given using different colored symbols as follows: Gulf Coast Native (GCN gold), Merino (MER dark blue), Poll Dorset (APD dark green), Poll Merino (APM red), Castellana (CAS brown), Churra (CHU pink), Meat Lacaune (MEL light blue), Milk Lacaune (MIL black), Ojalada (OJA orange), Rambouillet (RAM purple), Rasa Aragonesa (RAS pale green), Sumatra (SUN violet) and Tibetan (TIB yellow). The two largest principal components (A) separated Asian (SUM, TIB) and Northern European animals (APD) from a cluster containing breeds developed in southern Europe (including RAM, MER and APM). Plotting PCA1 and PCA6 revealed the genetic division within GCN (B).