Speciation over the edge: gene flow among non-human primate species across a formidable biogeographic barrier

Abstract

Many genera of terrestrial vertebrates diversified exclusively on one or the other side of Wallace's Line, which lies between Borneo and Sulawesi islands in Southeast Asia, and demarcates one of the sharpest biogeographic transition zones in the world. Macaque monkeys are unusual among vertebrate genera in that they are distributed on both sides of Wallace's Line, raising the question of whether dispersal across this barrier was an evolutionary one-off or a more protracted exchange-and if the latter, what were the genomic consequences. To explore the nature of speciation over the edge of this biogeographic divide, we used genomic data to test for evidence of gene flow between macaque species across Wallace's Line after macaques colonized Sulawesi. We recovered evidence of post-colonization gene flow, most prominently on the X chromosome. These results are consistent with the proposal that gene flow is a pervasive component of speciation-even when barriers to gene flow seem almost insurmountable.

Keywords: Wallace’s Line; X chromosome; gene flow; genomics; mechanisms of speciation; primate evolution.

Figure 1.

Thirty-five geographical origins of the 40 genetic samples analysed in this study. Numbered localities correspond to the approximate geographical origins of samples as follows, with asterisks denoting samples for which precise provenance is unknown: (1) M. siberu, (2) Ngasang, (3) Kedurang, (4) Malay*, (5) PM665, (6) PM664*, (7) Sukai, Gumgum, (8) PF660, (9) PF1003, (10) PF1001*, (11) PM1000, (12) PF654, (13) PF648, (14) PF651, (15) PM645, (16) PM639, (17) PF644, (18) PF643, (19) PF515, (20) PM565, PM566, PM567, (21) PM561, (22) PM582, (23) PM584, (24) PM592, (25) PM602, (26) PM613, (27) PM618, (28) PM614, PF615, PM616, (29) PF713, (30) PF549, (31) PM545, (32) PM571, (33) PM596, (34) PF625, (35) PF707. Dots are coloured by species as detailed in figures 2 and 3.

Figure 2.

Time-calibrated phylogeny (chronogram) recovered from analysis of autosomal RADseq data. Scale indicates divergence in million years ago (Ma). Black, grey, and white dots over nodes reflect ultrafast bootstrap values from iqtree that are greater than 99, 95 and 90, respectively. Grey bars near each node indicate the 95% CI for divergence times recovered from mcmctree. An inset indicates ranges of Sulawesi macaques and low bootstrap support for one node is indicated with an arrow. Tips are numbered according to their geographical localities depicted in figure 1.

Figure 3.

PCA analysis for all RADseq data (a) and RADseq data from only Sulawesi (b). Individual samples are numbered according to their geographical origins depicted in figure 1.

Figure 4.

Conditional genetic distance to the M. nemestrina genome (H3) as a function of f_DM statistic for autosomes (grey, calculated including sites with heterozygous and homozygous genotypes) and the X chromosome (red, based on genotyping by depth after excluding positions with heterozyous diploid genotypes in males). Statistics are based on WGS data divided into non-overlapping windows of the reference genome spanning five million base pairs. Positive values of f_DM indicate an excess of derived sites (relative to the rhesus macaque) that are shared between the M. tonkeana (H2) genome and H3; negative values indicate an excess of derived sites that are shared between the M. nigra (H1) genome and H3. Genetic distances are ‘conditional’ in the sense that the uncorrected per cent of divergent sites between H2 and H3 or between H1 and H3 is plotted depending on whether f_DM is positive or negative, respectively, for each genomic window. Inset depicts f_DM in 5 Mbp windows on the X chromosome.