Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - LocalAdaptation (Detecting Local Adaptation with Climate-Informed Spatial Genetic Models)

Teaser

Local adaptation, whereby individuals of a population exhibit higher fitness in their local environment compared to that experienced by other populations, has the potential to shape the distribution of genetic diversity and influence speciation. However, detecting and...

Summary

Local adaptation, whereby individuals of a population exhibit higher fitness in their local environment compared to that experienced by other populations, has the potential to shape the distribution of genetic diversity and influence speciation. However, detecting and quantifying the extent of local adaptation is challenging, since neutral demographic processes can leave signatures which are hard to distinguish from those of local selection. In this project, I propose to quantify the extent of local adaptation in Anatomically Modern Humans by using climate-informed spatial genetic models (CISGeM) to reconstruct past population sizes, local movements, and range expansions, and thus provide a null model against which the signature of geographically-localised selection can be detected.
In CISGeM, demography is affected by local resource availability, which in turn is defined by paleoclimate and paleovegetation reconstructions. By using these additional lines of evidence, it is possible to generate accurate demographic reconstructions for any number of populations, as well as integrating information from both modern and ancient genomes. Such spatially-explicit reconstructions are key for defining the expected neutral patterns due to complex demography, and thus allow us to isolate the signals of selection from this noisy background with high fidelity. The availability of paleoclimate reconstructions also enables formally testing hypotheses about the drivers of selection, integrating the changes in the strength of selection through space and time.
While this project will be focused on Anatomically Modern Humans, the framework that I will develop will be applicable to the investigation of local adaptation from genomic data in any species. Such tools are very timely, given the ever-increasing availability of large genetic datasets thanks to the decreasing cost of genotyping and sequencing in both model and non-model organisms.

Work performed

In order to detect selection, we need to first model the underlying demography of the species of interest, since certain demographic events can mimic the genetic signatures left by selection. During the initial part of this project, we have focussed on getting a better understanding of a number of key events. Taking advantage of advances in ancient DNA, we were the first to sequence ancient whole-genomes from Africa and East Asia, improving our understanding of migrations in these two understudied areas. We also investigated the mixing of ancestral population groups in the eastern parts of Eurasia. We detected a new, important ancestral group in the Caucasus which later contributed to the Bronze age migrations into Europe. Our work with aDNA and modern genomes highlighted the importance of mountains acting as barriers to geneflow both in the past and in more recent times, and we have worked to incorporate this environmental feature into our models.
Our project relies on a powerful spatial framework that is informed by climate reconstructions. During this initial phase of the project, we have focussed on improving our ability to downscale paleoclimate simulations (which are usually run at a coarse scale to keep computational costs within reason), so that we can accurately reconstruct the variability of climate at the local scale. We have also improved our ability to capture the seasonality of climate. Both features are particularly important to capture the possible role of mountains on affecting geneflow, an element highlighted by our analysis of human genomes. We are now finalising a publication that incorporates all our improvements and provides a new set of improved climate and vegetation reconstructions for the past 120k years suitable to drive advanced ecological analyses. I have also been coordinating an international working group synthesising information about ice sheets over the same time scale, filling a major gap on an environmental feature that likely played a key role in affecting human demography.
Finally, we have improved our proof-of-concept spatial model, and turned it into a more polished tool. Importantly, we are now able to simulate long stretches of the genome, including the effect of both mutations and recombination. This is essential to later simulate selection. Taking advantage of the new climate reconstructions that have generated, we are currently fitting the demographic model to available genetic data (including the ancient data we generated in the early parts of the project. This step will generate an accurate and realistic reconstruction of past human demography, which will then be used as the starting point to explore selection.
Finally, we have started reviewing the literature on selection scans in humans, building a comprehensive overview of genomic regions that have been previously suggested to be under selection, together with the driver that have been suggested to be responsible for each adaptation.

Final results

Our preliminary work to characterise the demography of areas of the world that had been so far neglected, such as Africa and East Asia, has generated several high profile papers, and received a great deal of attention by the international media (six of those papers were in the top 99th percentile of papers covered by media according to Altmetric). Thorough interaction with the press and radio, as well as social media, we have been able to reach a broad section of society (the combined readership of the newspapers that covered us runs into the hundreds of millions). The public is very interested in our past, and our work, which has taken advantage of the latest advances in ancient DNA and developed new approaches to analyse large genomic datasets, has helped to clarify the deep history of several populations.
Our work in improving climatic reconstructions for our genetic modelling is not yet published, but we have already received interest from a number of groups interested in using these data to study Quaternary extinctions. Our efforts in synthesising information for our models has also catalysed efforts by a number of glaciologists who are now working with us to generate a set of detailed reconstructions of ice sheets through time. Once completed, this resource will not just be an important tool for our work, but also provide a potentially important reference in the field of earth sciences for setting future research priorities, as it will highlight data gaps.
We have not been able to reap the rewards of our new spatial model, as most of this first year and a half has been devoted to developing and testing the tool, but I expect a number of publications to come over the next year. Presentation of preliminary results at conferences and workshops had led to a great deal of interest from colleagues, and we are confident that we have a valuable tool that will serve us well for the rest of the project.