Evolutionary origin of genomic structural variations in domestic yaks

Abstract

Yak has been subject to natural selection, human domestication and interspecific introgression during its evolution. However, genetic variants favored by each of these processes have not been distinguished previously. We constructed a graph-genome for 47 genomes of 7 cross-fertile bovine species. This allowed detection of 57,432 high-resolution structural variants (SVs) within and across the species, which were genotyped in 386 individuals. We distinguished the evolutionary origins of diverse SVs in domestic yaks by phylogenetic analyses. We further identified 334 genes overlapping with SVs in domestic yaks that bore potential signals of selection from wild yaks, plus an additional 686 genes introgressed from cattle. Nearly 90% of the domestic yaks were introgressed by cattle. Introgression of an SV spanning the KIT gene triggered the breeding of white domestic yaks. We validated a significant association of the selected stratified SVs with gene expression, which contributes to phenotypic variations. Our results highlight that SVs of different origins contribute to the phenotypic diversity of domestic yaks.

Fig. 1. Phylogeny of bovine species and pan-genome of wild and domestic yaks.

a Phylogenetic tree of 47 bovine individuals with de novo genomes-based SNPs with buffalo as outgroup. CDMC Chaidamu cattle, TC Tibetan cattle, ZMC Zhangmu cattle, WNC Wannan cattle, BS Brown Swiss, OBV Original Braunvieh, SA Sanga Ankole, ARS-UCD1.2, Btau_5.0.1 and UMD_3.1.1: Hereford cattle. b A cladogram of Bovini species based on single-copy genes in the individuals marked with a red star in Fig. 1a. c Variation of gene families in the pan-genome and core genome along with an additional yak genome. d Numbers and frequencies of core, near-core, and variable gene families. e, f The gene expression levels and Ka/Ks values of core (16,900; 12,532), near-core (4186; 2875), and variable genes (2934; 774) of yaks. Lowercase letters indicate significant differences (p < 0.05); two-sided Student’s t-test. Box edges indicate the upper and lower quartiles, the centerlines indicate the median value, the horizontal bars indicate the maximum and minimum values, and the dots indicate outliers. g numbers of genes without homologues in the wild assemblies (“new genes”) and genes originating from cattle as inferred from their phylogenetic position (“introgression genes”). h Heat map representing the number of genes present in the assembly on the X-axis but not in the assembly on the Y-axis.

Fig. 2. Characterization of the graph-genome of 47 de novo bovine genomes and distributions of structural variants (SVs) across these genomes and 386 individuals.

a Decrease and increase of the core nodes (red) and pan nodes (red + blue) respectively with the increase in number of yak, cattle and bison/wisent/gaur (BWG) genome assemblies added to the graph-genome. b The number of nodes in each discovery class (core, near-core, and variable) is shown per assembly. c Diagrams of diverse SV types from bovine graph pangenome based on the yak reference genome BosMut3.0. The bar chart shows the average number and length of each type of SV separately. The pie chart shows the number of genes affected by SV as a proportion of the overall number of genes. d Genomic features and variation landscape across the reference genome assembly. SNPs, INDELs, and SVs from resequencing data. e Pan-SV and core-SV length with 386 individuals of 7 bovine species. The light pink and green lines at the right indicate the core and pan SVs, respectively of the cattle pangenome.

Fig. 3. SVs contributed to high-altitude adaptation and domestication of yaks.

a F_ST and Φ_ST between SV haplotypes in wild yaks and other bovine species (wisent, bison, European taurine). There is a significant linear relationship between the F_ST and Φ_ST (Pearson’s r = 0.98, p < 2.2e-16). b GO and KEGG enrichment analysis of the candidate genes of the identified SVs. c, d The two SV-haplotypes of EPAS1 (facultative adaptation) and MB (obligate adaptation) genes are determined by the presence or absence of 254 bp and 155 bp in the intron, respectively. Hap-0: SV haplotypes in non-QTP cattle, guar, bison, wisent and outgroup buffalo. Hap-1: SV haplotype in wild yaks. A retrotransposons LINE1 that overlaps with SV of EPAS1 is indicated with brown bar. VISTA plot of conserved shows sequence similarity of non-coding elements (CNEs) around and in the SVs of EPAS1 and MB (indicated by triangles) to the BosMut3.0 reference. e, f Luciferase signals and expression of two SV-haplotypes of EPAS1 and MB. Values are shown as means ± SD from three biological replicates; Exact p-values are shown; two-sided Student’s t-test. g, h Frequencies of SV-haplotypes of EPAS1 and MB in domestic yak, wild yak and cattle from graph genomes. i Frequency distribution of SVs haplotypes (dots) in wild and domestic yaks. j–l SV-haplotypes (Hap-0, mainly in wild yaks; Hap-1, mainly in domestic yaks) of SV-DISC1, VWC2 and SNX3 and corresponding luciferase signals. Values are shown as means ± SD from three biological replicates; Exact p-values are shown; two-sided Student’s t-test. Source data are provided as a Source Data file.

Fig. 4. The serial translocation SV with KIT introgressed from cattle contributed to white coat colours of yaks.

a Scans of chromosomes 6 and 25 of the association with coat colour and of F_ST and XP-CLR values for black vs white yaks (red: SNPs; blue: SVs). b Chromatin interaction heatmap based on Hi-C sequencing data of white yak and BosMut3.0 (black phenotype), respectively, using the BosMut3.0 genome as a reference. Black triangles indicate topologically associating domain (TAD). The BC fragment shows the strongest reciprocal signal within TAD. The bar graph block indicates the A/B compartment. Positive eigen values indicate A-compartment, corresponding to transcriptional activation regions. Negative eigen values indicate B-compartment, which correspond to closed chromatin regions. c Relative mRNA expression level and (d) FPKM of KIT in the ear tissue of black (n = 6) and white yaks (n = 6). Values are shown as means ± SD; Exact p-values are shown; two-sided Student’s t-test. e–g The introgression between domestic yaks and colour-sided cattle with the NJ trees on SNPs of the phased KIT nearby ±10 kb, BC and βγ haplotype, respectively. h The most likely histories of the white and colour-sided yaks. Cs6: the allele on chromosome 6 of colour-sided yaks and cattle. SCs6: the allele on chromosome 6 of white yaks. Wt6/25: Wild-type allele on chromosome 6 or 25 of black and colour-sided yaks. Wt6/29: Wild-type allele on chromosome 6 or 29 of cattle. Chromosomes 6 and 25 of the yaks are homologous to chromosomes 6 and 29 of the cattle, respectively. Source data are provided as a Source Data file.