Whole-genome resequencing reveals genomic variation and dynamics in Ethiopian indigenous goats

Abstract

Ethiopia has about 52 million indigenous goats with marked phenotypic variability, which is the outcome of natural and artificial selection. Here, we obtained whole-genome sequence data of three Ethiopian indigenous goat populations (Arab, Fellata, and Oromo) from northwestern Ethiopia and analyzed their genome-wide genetic diversity, population structure, and signatures of selection. We included genotype data from four other Ethiopian goat populations (Abergelle, Keffa, Gumuz, and Woyto-Guji) and goats from Asia; Europe; and eastern, southern, western, and northern Africa to investigate the genetic predisposition of the three Ethiopian populations and performed comparative genomic analysis. Genetic diversity analysis showed that Fellata goats exhibited the lowest heterozygosity values (Ho = 0.288 ± 0.005 and He = 0.334 ± 0.0001). The highest values were observed in Arab goats (Ho = 0.310 ± 0.010 and He = 0.347 ± 4.35e-05). A higher inbreeding coefficient (F_ROH = 0.137 ± 0.016) was recorded for Fellata goats than the 0.105 ± 0.030 recorded for Arab and the 0.112 ± 0.034 recorded for Oromo goats. This indicates that the Fellata goat population should be prioritized in future conservation activities. The three goat populations showed the majority (∼63%) of runs of homozygosity in the shorter (100-150 Kb) length category, illustrating ancient inbreeding and/or small founder effects. Population relationship and structure analysis separated the Ethiopian indigenous goats into two distinct genetic clusters lacking phylogeographic structure. Arab, Fellata, Oromo, Abergelle, and Keffa represented one genetic cluster. Gumuz and Woyto-Guji formed a separate cluster and shared a common genetic background with the Kenyan Boran goat. Genome-wide selection signature analysis identified nine strongest regions spanning 163 genes influencing adaptation to arid and semi-arid environments (HOXC12, HOXC13, HOXC4, HOXC6, and HOXC9, MAPK8IP2), immune response (IL18, TYK2, ICAM3, ADGRG1, and ADGRG3), and production and reproduction (RARG and DNMT1). Our results provide insights into a thorough understanding of genetic architecture underlying selection signatures in Ethiopian indigenous goats in a semi-arid tropical environment and deliver valuable information for goat genetic improvement, conservation strategy, genome-wide association study, and marker-assisted breeding.

Keywords: Africa; Capra hircus; genome dynamics; pooled heterozygosity; population differentiation; whole genome.

FIGURE 1

Number of ROHs for the four genome length categories (ROH_100–150Kb, ROH_150–250Kb, ROH_250–400Kb, and ROH_>400Kb) of the study goat populations.

FIGURE 2

PCA plots for the first two components (PC1 and PC2, the respective variations explained in brackets) for the global goat populations.

FIGURE 3

PCA plots for the first two components (PC1 and PC2, the respective variations explained in brackets) for African goats.

FIGURE 4

PCA plots for the first two components (PC1 and PC2, the respective variations explained in brackets) for the East African goats.

FIGURE 5

TreeMix maximum likelihood phylogenetic tree showing the relationships among the 13 goat populations. Horizontal branch lengths are proportional to the amount of genetic drift that has occurred along that branch. The scale bar on the left shows 10 times the average standard error (s.e.) of the entries in the sample covariance matrix. Two migration edges between populations are shown with arrows pointing in the direction of the recipient group and colored according to the ancestry percentage received from the donor.

FIGURE 6

ADMIXTURE plot of the studied goat populations in a global context for 2 ≤ K ≤ 13 (EA-G1: East African Group 1; EA-G2: East African Group 2; NWSA: North, West and South African; AS: Asian; EU: European).

FIGURE 7

Manhattan plots for selection sweep analysis (A) between Arab vs. Tibetan, (B) between Fellata vs. Tibetan, and (C) between Oromo vs. Tibetan goat populations performed using the standardized population differentiation (ZF ST) approach. The horizontal line represents the arbitrary threshold for ZFst.

FIGURE 8

Manhattan plots performed using the standardized pool heterozygosity (ZHp) approach for each 100-kb sliding window with a 25-kb step size across all autosomes in the (A) Arab, (B) Fellata, and (C) Oromo goat populations. The horizontal line represents the arbitrary threshold for ZHp.