Discussion
One of the benefits of the current availability of WGS datasets is the access to an abundance of sequence variants, which allow a comparison of individuals or populations for several purposes. This is especially useful for the analysis of Y-chromosomal diversity, which used to be restricted by the availability of Y-chromosomal markers. The male-specific part of the Y-chromosome constitutes the longest haplotype in the mammalian genome and may serve as marker for mammalian paternal lineages. Here, we combined the dataset of the VarGoats project with published data and genotyped diagnostic SNPs in male goat samples from several sources.
The Y-chromosomal phylogeny of wild and domestic goats is in agreement with the Y-chromosomal tree on the basis of AMELY and ZFYgene fragments (Pidancier et al. 2006) and with a phylogeny of WGS sequences (Grossen et al., 2020; Cai et al., unpublished). MtDNA trees confirm the close relationship of markhor (Capra falconeri ) with bezoar and domestic goat, but do not show the separation of these species and the other wild goats. In addition, mitochondrial DNA (mtDNA) sequences of some, but not all East-Caucasian turs (Capra cylindricornis ) cluster with the mtDNAs of markhor, bezoar and domestic goat, illustrating a separate history of maternal and paternal lineages in cross-fertile species (Chen et al., 2018; Marín et al., 2017; O’Connell et al., 2014; Zhang et al., 2020; Zhang et al., 2016).
On the basis of WGS data, Zheng et al. (2020) and Xiao et al. (2021) reproduced the divergence of the domestic Y1 and Y2 haplogroup previously found on the basis of SNPs within or near Y-chromosomal genes (Lenstra et al., 2005). Here we report a further differentiation of haplogroups, resulting in a phylogeny supported by a largely independent preliminary dataset (Fig. S1) and two phylogenetic algorithms. We found that the major haplogroups corresponds with haplotypes defined by SNPs (Çinar Kul et al., 2015; Lenstra & Econogene Consortium, 2005; Pereira et al., 2009; Tabata et al., 2018, 2019; Vidal et al., 2017; Waki et al., 2015), but Y1A haplotypes belong to either haplogroup Y1AA or haplogroups Y1AB.
The phylogeny also indicates that these haplogroups diverged after the split of the markhor and the cluster of the wild bezoar and the domestic goats. The domestic goats and the two bezoar populations from Anatolia and Iran do not share haplotypes, whereas the bezoar haplotypes are attached to deep nodes in the tree of domestic haplotypes. This suggests an absence of male gene flow between the bezoar populations and between the bezoar and domestic goats from the same region. Thus, domestic goats, which possibly were derived from bezoar populations not sampled in this study, maintained their paternal lineages during the migration from the Fertile Crescent via Anatolia to Europe, this in spite of indications of management of wild goats in central Anatolia (Stiner et al., 2022).
Geographic plots of domestic haplogroup frequencies show a considerable spatial differentiation, which resonates with the strong phylogeography displayed by autosomal SNPs (Colli et al., 2018), but is in clear contrast with the weak phylogenetic structure displayed by the major mtDNA haplogroups (Luikart et al., 2001; Naderi et al., 2007, 2008; Zhao et al., 2014b, 2014a, Colli et al., 2015). Remarkably, Y-chromosomal haplotypes from all five haplogroups have been found in ancient DNA samples from Southwest Asia or Southeast Europe. The locations of ancient Y1AA, Y1B and Y2B samples are well outside the range of the corresponding domestic haplogroups (Fig. 3). This indicates that during the Neolithic and later worldwide migrations a series of bottlenecks and expansions in the domestic male lineage created a strong geographic differentiation of the haplogroup distribution (Fig. 4):
  1. The dominance of haplogroup Y1B in central and northern Europa may very well reflect population bottlenecks during the Neolithic introduction of agriculture via the Danube route (Cymbron et al., 2005; Rivollat et al., 2015; Tresset & Vigne, 2007)
  2. Y2A and Y1AA are almost the only haplogroups in Africa south of the Sahara. The two African Y1AA haplotypes are related to those of Asian, indicating that only Y2A expanded during the first introduction of domesticated goats in central and southern Africa
  3. Y2B has been found in two Neolithic Iranian samples whereas related CaYB2 haplotypes are present in Iranian and Anatolian bezoars. However, by population bottlenecks during the global spreading of domestic goats Y2b now only occurs in Asia east of the Indus River and in one goat from Madagascar (see below).
  4. Y1AA was found in Neolithic samples in southeast Europe, but now has a low frequency in Europe. In Asia it expanded together with Y2B and later came to South Africa when Asian goats were used to breed the Boer goat (see below).
  5. Y1AB is the most frequent haplogroup in north China. The distributions of Y1AB and Y1AA/ Y2B in East Asia correspond to ranges of the north-Chinese cashmere goats and the small Southeast Asian ‘katjang’ type, respectively (Porter et al., 2016). This obviously reflects the large difference of climate between northern and southern China, which determined a similar distribution of taurine and indicine cattle. These two types of cattle are supposed to have entered China via a northern and southern migration route, respectively (Chen et al., 2018; K. Zhang et al., 2020), which support the separate eastern expansions of the Y1AB and Y1AA/Y2B goats, respectively.
Exceptions to these geographic patterns follow from close relationships between haplotype from different continents, which are most likely explained by later major introgressions. Interestingly, in the phylogenetic trees and networks (Figs. 2 and S2) the Y2A haplotypes on Madagascar are closely related to Asian haplotypes and one Diana goat from northern Madagascar even has an eastern-Asian Y2B haplotype. However, autosomal DNA shows that the Malagasy goats are more related to the southern and eastern African continental goats (Colli et al., 2018; Denoyelle et al., 2021). This parallels a recent finding that Malagasy cattle combines Indian and admixed African zebu ancestry (Magnier et al., 2022). The Malagasy language has an Austronesian origin, which testifies the colonization of Madagascar by immigrants from southeastern Asia about 500 CE. Thus, it is likely that these immigrants brought Austronesian goats, cattle and possibly also other livestock from their region of origin to Madagascar.
Other introgressions are more recent. The exceptional Y1AA and Y2A in the English Anglo-Nubian is explained by the documented import during the 19th century of Indian and African goats to England. These served on the ships as source of milk and meat, but surviving males were crossed with English goats, which resulted in the emergence of a popular transboundary breed.
The worldwide popular Boer goat also is of mixed origin (Porter et al., 2016; Vidal et al., 2017) and carries exclusively Y1AA haplotypes. This breed is supposed to be a crossbred of local African and Indian goats, possibly mediated by incrossing of Anglo-Nubian males (Porter et al., 2016). The crossbred origin is consistent with the results of Colliet al. (2018): a separate phylogenetic position of the Boer relative the other African and Asian goats and the K=3 pattern of model-based clustering showing African and west-Asian ancestry. The Indian ancestry is entirely in agreement with a close clustering of the Boer and Pakistani Y1AA haplotypes (Figs. 1B and 2).
Subsequently, the Boer became itself a source of introgression. The same Y1AA haplotypes are closely related to Y1AA haplotypes in local breeds in Uganda, Malawi, Mozambique and Zimbabwe. In these countries crossbreeding with Boer goats from Africa is popular because of its excellent meat production (Banda et al., 1993; Garrine, 2007; Lu, 2011; Onzima et al., 2018). Therefore, it is most likely that the Y1AA in eastern and southern African goats originates from the Boer.
There were three out-of-range findings of Y1B, in the Ugandan Karamonja, in the Korean native goat and in the indigenous goats kept on the Chongmin Island in Shanghai. Because of the popularity of Swiss dairy goats in both Uganda (NAADS, 2005) and Korea (Kim et al., 2019), crossbreeding again is the most likely explanation. There are no data on European admixture in the Chongmin goats (Gao et al., 2020). Thus, exotic occurrence of Y chromosomal variants are direct and sensitive indicators of admixture events, but are to be complemented with quantitative admixture tests, such as model-based clustering, thef 3 and f 4 test or, ideally, identification of introgressed segments across the genome.
The latter approach may also lead to clues on the phenotypic consequences of the introgression via the identification of the admixed genes (Chen et al., 2018; Lv et al., 2014; Wang et al., 2015; Zheng et al., 2020). A more direct link with the Y-chromosomal variation would be provided if this can be linked to male phenotypic traits, but even in human genetics this has scarcely be investigated (Matsunaga et al., 2021; Yang et al., 2018; X. Zhang et al., 2021). Breeds in which different Y-chromosomal haplogroups occur may allow to study an association of haplogroups with typically male traits such as male fertility and dominance behavior. It would be interesting if Y-chromosomal variants can be related to the climate or other environmental features, because this would imply that the geographic differentiation of the Y-chromosomal variation is driven by regional adaptation.
Most introgressions described in this study contribute to the expansion of popular breeds at the expense of the original local breeds. On the one hand, depending on the extent of the gene flow this may decrease the diversity of the genetic resources; on the other hand, it does not necessarily disrupt the environmental adaptation, arguably one of the most consequential components of the phenotypic repertoire. If properly managed, admixture of productive breeds may also contribute to the sustainable conservation of local populations and illustrates that genetic diversity has never been a static phenomenon.
We conclude that the Y-chromosomal variation of goats reveals bottlenecks, expansions and introgressions, illustrating the power of Y-chromosomal markers for inferring the genetic origin of mammalian populations.