Past Connectivity but Recent Inbreeding in Cross River Gorillas Determined Using Whole Genomes from Single Hairs

Abstract

The critically endangered western gorillas (Gorilla gorilla) are divided into two subspecies: the western lowland (G. g. gorilla) and the Cross River (G. g. diehli) gorilla. Given the difficulty in sampling wild great ape populations and the small estimated size of the Cross River gorilla population, only one whole genome of a Cross River gorilla has been sequenced to date, hindering the study of this subspecies at the population level. In this study, we expand the number of whole genomes available for wild western gorillas, generating 41 new genomes (25 belonging to Cross River gorillas) using single shed hairs collected from gorilla nests. By combining these genomes with publicly available wild gorilla genomes, we confirm that Cross River gorillas form three population clusters. We also found little variation in genome-wide heterozygosity among them. Our analyses reveal long runs of homozygosity (>10 Mb), indicating recent inbreeding in Cross River gorillas. This is similar to that seen in mountain gorillas but with a much more recent bottleneck. We also detect past gene flow between two Cross River sites, Afi Mountain Wildlife Sanctuary and the Mbe Mountains. Furthermore, we observe past allele sharing between Cross River gorillas and the northern western lowland gorilla sites, as well as with the eastern gorilla species. This is the first study using single shed hairs from a wild species for whole genome sequencing to date. Taken together, our results highlight the importance of implementing conservation measures to increase connectivity among Cross River gorilla sites.

Keywords: Cross River gorilla; NGS; bottleneck; gene flow; hairs; inbreeding; non-invasive; wild gorillas.

Figure 1

(A) Map showing the geographic distribution of the four gorilla subspecies. The Cross River and MN gorilla distribution ranges are from [2], and the other subspecies shape files were obtained from the IUCN. The shape files showing the specific Cross River gorilla populations (darker green) have been adapted with permission from [2] (Supplementary File). Inset at the top right of the map shows the Cross River distribution in more detail. Black dots indicate sites for which we have samples in the final dataset. Black crosses indicate sampled sites for which samples were discarded after quality control (Takamanda NP (National Park), CRNP Okwa Hills (Cross River National Park, Okwa Hills), and Mt. Doudou). Samples of unknown geographic origin are not represented in this map. (B) Principal component analysis (PCA) including hair samples (filled circles) and blood samples (empty circles) of the four subspecies. (C) PCA of western lowland gorilla hair samples with color indicating each site (Monte Alen, MoA; Ngaga Camp, Nga; Deng Deng, Den; Bai Hokou, Bai; unknown origin, Unk1). (D) PCA of Cross River gorilla hair samples with color indicating each site (Afi Mountain Wildlife Sanctuary, Afi; CRNP Boshi Extension, Bos; Mbe Mountains, Mbe; and Kagwene Gorilla Sanctuary, Kag). (E) Heatmap of pairwise F_ST of hair samples for both western gorilla species, CR and WL. Site labels are colored to indicate species, with red for WL gorilla sites and green for CR gorilla sites. Cell color indicates Fst value.

Figure 2

(A) FROH (percentage of the genome in runs of homozygosity (ROH) for different ROH size ranges (0.5–1.5 Mb, dark grey; 1.5–2.5 Mb, grey; 2.5–5 Mb, light grey; 5–10 Mb, yellow; and >10 Mb, blue). (B) Inference of recent effective population size from GONE for each subspecies (Cross River gorilla, CR; western lowland gorilla, WL; Grauer’s gorilla, GR and mountain gorilla, MN).

Figure 3

(A) Representation of the main results from D-statistics analysis between Cross River (CR) sites and western lowland (WL) gorilla regions. Arrows indicate which western lowland (WL) sites have higher allele sharing with Cross River (CR) gorillas. Arrow thickness indicates the level of gene flow, with the thicker arrow indicating higher gene flow between WL and CR sites. Empty dots indicate sites (CRNP Boshi Extension and Monte Alen) that have no positive signal of allele sharing with the other gorilla subspecies. (B) F-branch (fb(C)) analysis on excess allele sharing between species/populations/regions. The x-axis and y-axis trees are the species tree and the expanded species trees (with internal branches indicated with dotted lines), respectively. Grey-colored cells indicate comparisons for which gene flow could not be inferred based on the given tree topology. Red color indicates fb value; the darker color indicates greater allele sharing between the expanded tree branch (relative to its sister branch) (b) and the populations on the x-axis (C).