New Insight into the Genome-Wide Diversity and Admixture of Six Colombian Sheep Populations

Abstract

Creole sheep represent a strategic genetic resource for populations living in marginal areas under financial restrictions on the American continent. Six Colombian sheep breeds (two wool (BCL-Boyacá and NCL-Nariño, 12 and 14 samples) and four hair (OPCE-Ethiopian, 54 samples; OPCS-Sudan, 74 samples; OPCP-Pelibeuy, 59 samples; OPCW-Wayúu, 24 samples) were genotyped using the Illumina Ovine SNP50 BeadChip. Data was also included from international 44 breeds from International Sheep Genomics Consortium (ISGC) and from data published in previous a previous work on the Caribbean and African breeds. Although geographically separated, wool (NCL, BCL) and hair types (OPCE, OPCS, OPCW) presented little genetic differentiation (FST 0.05) at a global level but several groups of animals separated suggesting local clustering due to geographical isolation. The OPCP underwent a recent crossing with Mexican Pelibuey, explaining its differentiation. Findings in this work such as the proximity to West African Djallonké (WAD) and Barbados Black Belly (BBB), suggest different introductions of African type animals from the Caribbean region on a pre-existing genetic basis formed by animals deriving from the first importations coming from Europe in colonial times. As expected, Colombian wool breeds showed, in particular in Admixture software results, a greater genomic component in common with European breeds and in particular with Iberian ones (Churra). This study provides a basis for future research into the genetic diversity within and between the Colombian sheep breeds analysed, and scientific data for policy decisions on Farm Animal Genetic Resources (FAnGR).

Keywords: Africa; Ovis Aries; SNPs; South America; admixture; animal genetic resources; differentiation; genomics.

Figure 1

Distribution of runs of homozygosity inbreeding coefficients for each Colombian population.

Figure 2

Genetic differentiation pairwise FST Matrix values among 44 sheep populations/breed. the dashed red line highlights the Colombian breeds.

Figure 3

3D MDS plot showing the genetic relationships among all sheep breeds analysed. For a full definition of the breeds, see Table S1. The red circle defines the cluster of African breeds, green those of the Middle East including Egypt, the light blue that of the European breeds including Oceania, purple that of the Caribbean breeds and finally black that of the Colombian hair-type breeds.

Figure 4

Neighbor network constructed from Nei’ genetic distances. In black the Colombian sheep breeds. Origin of breeds are indicated by colours: light blue, Asian, Middle East including Egypt; red, Africa; green, South-American native; blue, Caribbean; violet, European including Oceania.

Figure 5

Clustering of 44 sheep breeds based on Admixture estimation for K = 2 to K = 10. Each bar indicated an individual and each colour the membership to a particular cluster; the number of clusters is determined by the K value. The sector inside the graphic indicated the geographical precedence of the breed: Africa, Asia and Middle-East including Egypt, the Caribbean, Europe including Oceania; South America and Colombia.

Figure 6

Circos plot of the number of haplotypes shared between breeds. Each coloured band. each coloured band represents the total number of haplotypes that are divided between a pair of races, where a greater thickness of the band represents a greater number of haplotypes. The flows within the same populations and those where the count is less than 50 are excluded from the graph. In bold are the Colombian sheep breeds. Origin of breeds are indicated by colours: light blue, Asian, Middle-East including Egypt; red, Africa; green, South-American native; blue, Caribbean; violet, European including Oceania.

Figure 7

Dendrogram of individual pairwise genetic distances.

Figure 8

MDS plot constructed using the first two dimensions, showing the genetic relationships among the six Colombian populations (OPCP, OPCE, OPCS, OPCW, NCL, BCL).

Figure 9

Model-based clustering using genotypic data from seven sheep breeds. K, Number of clusters, from 2 to 5, and 10. The best prediction model was K = 3. Each bar indicated an individual and each colour the membership to a particular cluster, the number of clusters is determined by the K value.

Figure 10

Visualisation of high-resolution population networks for 6 Creole sheep breeds. K-NN = 10 (left) investigated small-scale structures, K-NN = 30 (right) targeted large-scale structures and breeds are distinguished by node colour.