Non-invasive genomic sampling uncovers novel connectivities and origins of confiscated gorillas

Abstract Background Gorillas are a group of African great apes with two species and four subspecies that are currently critically endangered or endangered. Previous studies that analysed the genetics of wild gorillas from non-invasive samples, such as faeces or hair, analysed short mitochondrial or nuclear markers, which may not reflect the wider nuclear genome. Recent technical advances in target capture hybridisation, enrich the endogenous DNA content of non-invasive samples, allowing contiguous genomic regions to be sequenced. Results Here, we generated georeferenced genetic data from faecal and hair samples of 280 wild gorillas, sampled from three of the four gorilla subspecies, across large parts of their present-day distributions. With this expanded representation of gorilla genetic diversity in the wild, we detected three population clusters in western lowland gorillas, with the Sangha River and its affluents acting as significant barriers to gene flow. We reconstructed patterns of past population connectivity between western lowland gorillas in the north-eastern distribution range and Cross River gorillas, which may have been facilitated by a migration corridor also used by the Central and Nigeria-Cameroon chimpanzee subspecies. Finally, we predicted the geographic origins of wild-born gorillas, achieving a mean prediction error of 65 km, with a population-level resolution for mountain gorillas and some populations of western lowland gorillas. Conclusion Our work characterises fine-scale population structure in western lowland gorillas, which will be informative for future conservation strategies. This proof of concept in predicting geographic locations of wild gorillas, will be useful for future applications to geolocalise trafficked or rescued gorillas.