Through domestication and adaptation to local environments, domestic animals and plants often acquire new phenotypes via interspecific hybridization, primarily through repeated backcrossing and artificial selection to meet human needs or enhance their own fitness. However, determining the specific genes or genetic components of a breed species that are derived from introgression from closely related species is challenging due to frequent occurrences of hybridization in the same or neighboring regions. The shared genetic variants between closely related species may arise from ancestral polymorphisms, which hinders our ability to quickly identify genes derived from interspecific introgression, accurately validate functional differences of such alleles, and efficiently utilize these genetic resources.

Domestic yaks (Bos grunniens) are reared at altitudes between 3,000m and 5,000m on the Qinghai-Tibet Plateau (QTP) and exhibit diverse phenotypes (Figure 1). They were domesticated from the wild yak (B. mutus) in the northern QTP during the early Holocene, more than 7300 years ago. On the other hand, cattle (B. taurus and B. indicus), which were primarily domesticated in the Near East and Indus river basin, were introduced to the QTP around 3,500 years ago. Natural or artificial hybridization between cattle and yaks has resulted in F1 hybrids, with the female hybrids being fully fertile while the male hybrids are sterile. Repeated backcrossing between fertile female hybrids and either yaks or cattle has inevitably led to genetic introgression between the two species. These introgressions may complicate genetic studies related to high-altitude adaptation and the domestication process of wild yaks.

In addition to SNPs, there is growing evidence indicating that genomic structural variations (SVs > 50bp) play a crucial role in diverse phenotypic variations in both animals and plants. These SVs, which have a lower probability of convergence, are more accurate than SNPs in determining allelic changes between species and examining interspecific introgressions.

Figure 1. phenotypic traits of wild and domestic yaks

Based on our previous genomic studies of bison, wisent, and wild and domestic yaks, our sample set comprises 386 individuals from these bovine species, covering their entire distributional ranges (Figure 2). Notably, this includes 18 wild yak samples collected from naturally deceased individuals in the Kekexili, a high-altitude uninhabited area, over the past 20 years. By constructing the bovine graph-genome and combining it with population resequencing data, we generated a high-quality dataset of SVs. Based on this dataset, we conducted a detailed analysis of high-altitude adaptation, domestication, and hybrid introgression among wild yaks, domestic yaks, and cattle, with a particular focus on the diverse origins of SVs in domesticated yaks.

Figure 2. Geographical distribution of bovine species

We excluded introgressed SVs from cattle and identified 1,254 SVs overlapping with 334 candidate genes associated with domestication. We uncovered multiple genes related to the nervous system (e.g., DISC1, VWC2, and SNX3) and performed functional validation of allelic SVs through biochemical and expression experiments (Figure 3). These results suggest that the control of aggression, tameness, and sociability in yaks was likely a primary target of selection during the initial domestication process. However, all of these candidate genes could not be identified if the interspecific introgressions could not effectively exclude.

By conducting phylogenetic analyses of allopatric individuals and excluding sympatric introgressions, we successfully identified many structural variations (SVs) associated with high-altitude adaptation in both domestic and wild yaks. These SVs are involved in various biological processes, including oxygen transport, energy metabolism, and cardiovascular function. Notably, we identified several candidate genes (e.g., EPAS1, and MB) that have previously been associated with high-altitude adaptation in humans and other animals (Figure 3). The likely functional differentiation of the allelic SVs in these genes were also confirmed by both gene expression and biochemical results. Our results also suggest that humans and many animals adopt similar genetic variations at the same pathways to adapt the arid high-altitude habitats. 

Figure 3. Functional verification of SVs involved in high-altitude adaptation and domestication

Through comparing sympatric and allopatric groups, we discovered that nearly 90% of domestic yaks exhibited introgression from cattle, accounting for approximately 1-5% of their genomes. These introgressions from cattle likely serve as the genetic basis and molecular markers for the observed phenotypic variations in domestic yaks. For example, the color-sided domestic yaks (Figure 1) may have originated from genetic introgression from color-sided cattle. Moreover, this introgression may have subsequently triggered new genetic mutations that gave rise to the white yaks, although the precise molecular mechanism underlying this process requires further investigation (Fig. 4). To validate these introgressions and their contributions to phenotypic variations, we conducted chromatin interaction, gene expression, and biochemical experiments and all results confirmed the likely phenotypic changes triggered by allelic genetic changes (Fig. 4). Therefore, humans might intentionally introduce color-sided cattle to the QTP that resulted in origin of color-sided and white yaks because of unintentional and/or intentional interspecific hybridization between yaks and cattle and further crossing and mutations within yaks.

Figure 4. The serial translocation SV with KIT introgressed from cattle contributed to white coat colors of yaks. a Scans of chromosomes 6 and 25 of the association with coat color 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). (e) Micrographs of immunohistochemistry of the KIT protein in sections of ear tissue (top) and croup skin (bottom) samples of white (left) and black (right) yaks. (f) Hematoxylin and eosin (Top) and DAB Detection Kit (bottom) staining of ear tissue (top) and croup skin (bottom) of black (right) and white (left) yaks. Images are representative of three experiments.

In addition, numerous allelic SVs derived from cattle play a role in aiding domestic yaks' adaptation to low-altitude environments (Figure 5). Concurrently, we also made the discovery that most QTP cattle have acquired allelic SV introgressions from yaks during their adaptation to high-altitude conditions. For instance, EPS1-SV from yaks was identified in many QTP cattle, potentially enabling their survival in low-oxygen environments (see Figure 5). These bidirectional introgressions contribute to the phenotypic diversities observed in both yaks and cattle (Figure 4). These findings underscore the significance of interspecific introgressions in driving phenotypic variations in domestic yaks and cattle.

Figure 5. Widespread interspecific SV introgressions between domestic yaks and cattle in the QTP. (a) The distributions of all detected introgressed SVs in domestic yaks from cattle on the total genome. These introgressions were determined based on tree topologies established based on SNPs of the phased SV-nearby ± 50 kb haplotypes. (b) Four tree topologies and genic locations of SVs: no introgression between domestic yaks and cattle (tree1), introgression from domestic yaks to cattle (tree2), introgression from cattle to domestic yaks (tree3), and introgression between domestic yaks and cattle to each other (tree4). (c) introgression between domestic yaks and cattle with the tree based on SNPs of the 50 kb flanking regions of the phased EPAS1 SV haplotype of using water buffalo as outgroup showing genetic exchange between cattle and yak. Hap-0 indicates the cattle type and Hap-1 indicates the yak type. (d) Geographical distribution of SV-haplotypes of EPAS1 with yellow dashed line suggested with introgression in domestic yaks from cattle while the green dashed line for cattle with introgression from domestic yaks.

Conclusion, our findings highlight that yaks and cattle comprise an excellent yet intricate system for studying evolution in the QTP, encompassing three major drivers, adaptative selection by arid habitat, artificial selection through domestication, and interspecific introgression (Figure 6).

Figure 6. Evolutionary processes of yaks and cattle and the key underlying genes in the QTP