Ancient Homozygosity Segments in West African Djallonké Sheep Inform on the Genomic Impact of Livestock Adaptation to the Environment

Abstract

A sample of Burkina Faso Djallonké (West African Dwarf) sheep was analyzed to identify stretches of homozygous segments (runs of homozygosity; ROH) overlapping with ancient homozygosity-by-descent (HBD) segments. HBD segments were considered ancient if they were likely to be inherited from ancestors living from 1024 to 2048 generations ago, roughly coinciding with the time in which sheep entered into West Africa. It is hypothesized that such homozygous segments can inform on the effect of the sheep genome of human-mediated selection for adaptation to this harsh environment. PLINK analyses allowed to identify a total of 510 ROH segments in 127 different individuals that could be summarized into 124 different ROH. A total of 32,968 HBD segments were identified on 119 individuals using the software ZooRoH. HBD segments inherited from ancestors living 1024 and 2048 generations ago were identified on 61 individuals. The overlap between consensus ROH identified using PLINK and HBD fragments putatively assigned to generations 1024 and 2048 gave 108 genomic areas located on 17 different ovine chromosomes which were considered candidate regions for gene-annotation enrichment analyses. Functional annotation allowed to identify six statistically significant functional clusters involving 50 candidate genes. Cluster 1 was involved in homeostasis and coagulation; functional clusters 2, 3, and 6 were associated to innate immunity, defense against infections, and white blood cells proliferation and migration, respectively; cluster 4 was involved in parasite resistance; and functional cluster 5, formed by 20 genes, was involved in response to stress. The current analysis confirms the importance of genomic areas associated to immunity, disease resistance, and response to stress for adaptation of sheep to the challenging environment of humid Sub-Saharan West Africa.

Keywords: West African dwarf sheep; adaptation; domestication; identity-by-descent; runs of homozygosity.

Figure 1

Dispersion of the individual Djallonké sheep genotypes on the bidimensional space formed by the two first factors retained after computation of principal component analysis (PCA) on the between individuals distance matrix. Factor 1, on the X-axis explained 14.5% of the variance. Factor 2, on the Y-axis, explained 4.7% of the variance. Individuals sampled in the surroundings of Mangodara are in open squares while those sampled in Dédougou are in solid squares.

Figure 2

Summary of the identification runs of homozygosity (ROH) carried out using the program PLINK. (A) The ROH complement (total length and number) per individual in Djallonké sheep. (B) The frequency of ROH according to their length.

Figure 3

Runs of homozygosity distribution and coverage for each autosome in Djallonké sheep. The barplot shows the number of runs of homozygosity in different length classes: 1–5 Mb, 5–10 Mb, 10–15 Mb, 15–20 Mb, and >20 Mb.

Figure 4

Genomic inbreeding coefficients estimated with respect to different base populations (in 11 generation classes from G2 to G2048) estimated as the probability of a given genomic segment belonging to any of the 11 homozygosity-by-descent (HBD) classes fitted. (A) The variation in genomic inbreeding of individuals with no HBD segments assigned to G1024 or G2048. (B) The same variation for the individuals harboring HBD segments kept for enrichment analyses. Dark lines are the average values for each subpopulation. Note that (A) and (B) have different ranges on the Y-axis.

Figure 5

Circular map summarizing information on functional clusters identified in Djallonké sheep. Chromosomes are shown in the outermost circle. The innermost circles show the distribution of the candidate genomic areas identified within a chromosome. Differences in height among candidate areas inform on the number of SNPs defining them. Candidate genes forming the significantly enriched functional term clusters are indicated beside their genomic localization. At the center of the map, links among candidate genes belonging to the same functional cluster are illustrated using the same color (cluster 1 in brown on OAR23; cluster 2 in red on OAR3; cluster 3 in dark blue; cluster 4 in light blue; cluster 5 in green; cluster 6 in yellow on OAR5).