Ancient DNA has unlocked keys to our evolutionary past, revealing gene flow between modern humans and extinct hominins such as Neanderthals and Denisovans. We lack fossil remains and hence ancient DNA for our closest living relatives, the great apes. However, the development of novel statistical techniques facilitated the discovery of gene flow from now-extinct lineages in humans, in the absence of ancient DNA. In 2019 Martin Kuhlwilm, Tomas Marques-Bonet and colleagues, demonstrated this complex statistical inference approach could be applied to the genomes of present-day bonobos to reconstruct their evolutionary history including gene flow from a now extinct ‘ghost’ lineage. The question was, could we extend this to gorillas and orangutans, who had been understudied from the perspective of gene flow. 

At this point, several interests converged: Martin’s extensive expertise and interest in archaic gene flow, Tomas as a specialist in primate genomics had access to high-quality gorilla genomes including six newly sequenced genomes from previously unsampled regions, generated in collaboration with the Sanger Institute, and I was interested in population genetics and the evolutionary history of the great apes. I had previously analysed local adaptation in an ancestral chimpanzee population and found this fascinating and challenging. However, moving from adaptation to demographic modelling was a challenge I did not fully appreciate the scale of until much later on.

I began the PhD in March 2020, in hindsight not ideal timing. But life and science continues. Working with non-human data, even of our closest living relatives is non-trivial, particularly when attempting to tackle such complex questions. It became apparent, that despite my best efforts a systematic assessment of archaic gene flow across the great apes would be unfeasible, particularly for orangutans whose tangled demography will need clearer hypotheses, more extensive sampling and more powerful methods to address.

On the other hand, in gorillas we had a clear hypothesis that some of the current uncertainties in their demographic model could be resolved by hidden introgression events, such as from ghost lineages. Due to sample limitations, no previous demographic model had included whole genome data from representatives of all four gorilla subspecies. Developing a ‘null’ demographic model of the four extant subspecies was an arduous task to say the least. We were interested in inferring effective population size changes through time, including bottlenecks, divergence times and levels of gene flow. This involved iteratively improving the demography, implementing different summary statistics, assessing the impact of potential correlations in the summary statistics and more. As it sounds, this was intellectually all-encompassing, but as the only person undertaking this type of work in the group it was also isolating.

Mountain gorilla mother and infant during a rest period. Photograph by Mike Cranfield, Gorilla Doctors.

Once the hurdle of a reasonable null demography was overcome, the project became more manageable. As I began adding the possibility of gene flow from ghost lineages into the common ancestors of eastern or western gorillas and performing model comparisons I was able to take much more fun in the work. From the demographic modelling I was seeing a signal of ghost gene flow into the common ancestor of eastern gorillas, but not into the common ancestor of western gorillas. In revisions, we undertook a wider exploration of the ghost parameter space, and the inference that a model with 2.47% of ghost gene flow to the eastern ancestor best fit the empirical data was proven to be robust.

When we localised the putative introgressed fragments within eastern gorilla genomes, they indeed looked characteristic of an archaic origin. We saw sharing of the putative introgressed material across eastern gorillas, but this was higher within each subspecies, indicative of a younger introgression event than the corresponding event in bonobos, since most of the introgressed material is still segregating in gorillas. We also observed a consistent depletion in introgressed material on the X chromosome across gorillas, bonobos and humans, likely due to purifying selection.

Given my background in adaptation I was keen on exploring the possibility of adaptive introgression in the gorillas. This would not have been possible without the patient and thoroughly helpful input of Derek Setter, first author of the method ‘VolcanoFinder’, which is able to detect adaptive introgression without an archaic reference using distortions of the site frequency spectrum. Intersecting the outliers of adaptive introgression with the putative introgressed regions, we uncovered a particularly interesting candidate related to bitter taste perception. The input of primatologists such as Katja Guschanski, shed light on this signal, beyond the essential role of taste receptors to avoid toxicity, eastern and western gorillas have different diets and as such  bitter taste receptors are a very plausible target of adaptive introgression.

Open questions that remain include, whether there was one ghost introgression pulse into eastern gorillas or multiple? Do further ghost introgression events exist which we currently lack resolution to detect? Deep population structure within western lowland gorillas is an avenue ripe for exploration, likely via denser non-invasive sampling. Likewise, deeper functional characterisation of the introgressed regions, going from the dry lab to functional assays, or knockouts would be exciting, albeit risky.

This has been a long journey. Since I started, Martin has become a PI and the work of Tomas has culminated in a special issue on primate genomics in Science. I am grateful for this opportunity to work on a topic so intellectually stimulating, and for being given the space to troubleshoot and explore (although sometimes it felt endless). I am also very thankful to have received the SMBE Graduate Student Excellence Award and to have had the incredible opportunity to present our findings at SMBE 2023 in Ferrara.

We have added increased resolution on the evolutionary history of gorillas, and uncovered a previously-unknown lineage of gorillas via analysis of variation in present-day gorillas. In the face of the Anthropocene, I hope we can continue to learn more about our charismatic closest relatives before it is too late.

Tomas Marques-Bonet, Martin Kuhlwilm and Harvinder Pawar at SMBE 2023, Ferrara, Italy.

Acknowledgment: Header photograph, mountain gorilla mother and infant together with another adult female during a rest period. Copyright: Mike Cranfield, Gorilla Doctors, https://www.gorilladoctors.org/