Whole-genome sequencing reveals selection signatures associated with important traits in six goat breeds

Abstract

Comparative population genomics analysis is an effective approach to identify selection signatures in farm animals. In this study, we systematically investigated the selection signatures in six phenotypically diverse goat breeds using SNPs obtained from pooled whole-genome resequencing data. More than 95.5% of 446-642 million clean reads were mapped to the latest reference goat genome, which generated a sequencing depth ranging from 22.30 to 31.75-fold for each breed. A total of 5,802,307, 6,794,020, 7,562,312, 5,325,119, 8,764,136, and 9,488,057 putative SNPs were detected in Boer, Meigu, Jintang Black, Nanjiang Yellow, Tibetan, and Tibetan cashmere goats, respectively. Based on the genome-wide F_ST and expected heterozygosity scores along 100-kb sliding windows, 68, 89, 44, 44, 19, and 35 outlier windows were deemed as the selection signatures in the six goat breeds. After genome annotation, several genes within the selection signals were found to be possibly associated with important traits in goats, such as coat color (IRF4, EXOC2, RALY, EIF2S2, and KITLG), high-altitude adaptation (EPAS1), growth (LDB2), and reproduction traits (KHDRBS2). In summary, we provide an improved understanding of the genetic diversity and the genomic footprints under positive selection or the adaptations to the local environments in the domestic goat genome.

Figure 1

Summary of all six domestic goat breeds included in this study. (A) Pools of genomic DNA from the following populations were sequenced: Boer (n = 20; sex = F/M), Meigu (n = 20; sex = F), Jintang Black (n = 20; sex = F/M), Nanjiang Yellow (n = 20; sex = F), Tibetan (n = 20; sex = F/M), and Tibetan cashmere goat (n = 20; sex = M). Photographs were taken by Hongping Zhang and Tianzeng Song. (B) Plot of the first three principal components (PC1, PC2, and PC3) of six goat breeds based on the detected SNPs.

Figure 2

Genome-wide histograms of SNPs detected in six goat breeds. The outermost circle represented GC content of the goat reference genome sequence. The circles from outer to inner show the genome-wide distribution of SNPs (counted in 1-Mb non-overlapping windows) in Boer, Meigu, Jintang Black, Nanjiang Yellow, Tibetan, and Tibetan cashmere goats, respectively.

Figure 3

Genome-wide distributions of selection signals in the Meigu (A), Nanjiang Yellow (B), and Tibetan cashmere (C) goat. Manhattan plots of ZF_ST (>0) and absolute values of ZH_p (<0) across all autosomes were plotted with different colors. The ZF_ST and ZH_p values were calculated for each sliding 100-kb window with steps of 25 kb across all autosomes. The horizontal dashed line indicates the 99.5 percentile of all the ZF_ST or ZH_p values.

Figure 4

Summary of the overlapped outlier windows detected via ZF_ST or ZH_p scores. (A) A Venn diagram of the outlier windows detected via ZF_ST or ZH_p scores in six goat breeds, based on their genomic positions. Green circles represent the numbers of the outlier windows detected via ZF_ST values, while the light blue circles represent the windows identified via ZH_p scores. (B) A Venn diagram of the overlapped outlier windows across all the six goat breeds, based on their genomic positions. (C) Summary of the biological processes into which the candidate genes can be classified.

Figure 5

Summary of F_ST and ZH_p values of each SNP at the RALY-EIF2S2 and IRF4-EXOC2 loci. (A) The RALY-EIF2S2 locus; (B) the IRF4-EXOC2 locus.