Persistent gene flow suggests an absence of reproductive isolation in an African antelope speciation model

Abstract

African antelope diversity is a globally unique vestige of a much richer world-wide Pleistocene megafauna. Despite this, the evolutionary processes leading to the prolific radiation of African antelopes are not well understood. Here, we sequenced 145 whole genomes from both subspecies of the waterbuck, an African antelope believed to be in the process of speciation. We investigated genetic structure and population divergence and found evidence of a mid-Pleistocene separation on either side of the eastern Great Rift Valley, consistent with vicariance caused by a rain shadow along the so-called 'Kingdon's Line'. However, we also found pervasive evidence of not only isolated and recent, but also widespread historical gene flow across the Rift Valley barrier. By inferring the genome-wide landscape of variation among subspecies, we found 14 genomic regions of elevated differentiation, including a locus that may be related to each subspecies' distinctive coat pigmentation pattern. We investigated these regions as candidate speciation islands. However, we observed no significant reduction in gene flow in these regions, nor any indications of selection against hybrids. Altogether, these results suggest a pattern whereby climatically driven vicariance is the most important process driving the African antelope radiation, and suggest that reproductive isolation may not set in until very late in the divergence process.

Estimating admixture pedigrees of recent hybrids without a contiguous reference genome

“…We used low‐depth (3X) WGS data for 38 waterbuck ( Kobus ellipsiprymnus ) samples (Wang et al, 2022) from four populations in East Africa. Two of the populations, Kidepo Valley National Park (KVNP, Uganda) and Maswa (Tanzania), belong to the defassa waterbuck subspecies ( K. e. defassa ).…”

“…Two of the populations, Kidepo Valley National Park (KVNP, Uganda) and Maswa (Tanzania), belong to the defassa waterbuck subspecies ( K. e. defassa ). The other two populations, Samburu and Nairobi (Kenya), are in the range of the common waterbuck subspecies ( K. e. ellipsiprymnus ), but are known to have received recent gene flow from the defassa waterbuck (Figure 3a) (Lorenzen et al, 2006; Wang et al, 2022).…”

Estimating admixture pedigrees of recent hybrids without a contiguous reference genome

“…We used low depth (3X) whole genome sequencing data for 38 waterbuck ( Kobus ellipsiprymnus ) samples (Wang et al, 2022) from four populations in East Africa. Two of the populations, Kidepo Valley National Park (KVNP, Uganda) and Maswa (Tanzania), belong to the defassa waterbuck subspecies ( K. e. defassa ).…”

“…Two of the populations, Kidepo Valley National Park (KVNP, Uganda) and Maswa (Tanzania), belong to the defassa waterbuck subspecies ( K. e. defassa ). The other two populations, Samburu and Nairobi (Kenya), are in the range of the common waterbuck subspecies ( K. e. ellipsiprymnus ), but are known to have received recent gene flow from the defassa waterbuck (Figure 3A) (Wang et al, 2022; Lorenzen et al, 2006).…”

“…We got access to low coverage (3X) sequencing data of waterbucks ( Kobus ellipsiprymnus ) mapped to the defassa waterbuck ( K. e. defassa ) draft genome (DFW) and a list of quality controlled sites (Wang et al, 2022). From that we estimated genotype likelihoods with ANGSD (Korneliussen et al, 2014) using the GATK genotype likelihood model (McKenna et al, 2010).…”