Conclusion
Gene-flow between closely related groups is increasingly being implicated as a key driver of adaptation. Here we have sought to address whether this process has resulted from admixture events which we previously inferred in African populations within the last 4,000 years. Our approach uncovered regions of the genome previously known to have undergone selection such as the LCT/MCM6 gene complex in relation to lactose persistence, and DARC in relation to P. vivax malaria in the Fula of West Africa. In both cases, combining these observations with our understanding of the admixture history of the Fula demonstrates that changes in local ancestry can help pinpoint regions of the genome where natural selection may have been working. In addition, we identified other regions of the genome where local ancestry is deviates from expectations [Fig. \ref{fig:figSharedAncestry}], raising the possibility that adaptive gene-flow may be a more general, albeit rare, phenomenon. However, to evaluate how widespread this phenomenon is further work is needed to develop approaches that are able to detect more subtle signatures of ancestry deviations, and to account for the problems of non-independence highlighted above. Nevertheless, whilst adaptive introgression (i.e. adaptation resulting from interspecific gene-flow) has a deep literature, both theoretically \citep{Lewontin1966HybridizationEnvironments} and empirically (for example in animals reviewed in \citet{Hedrick2013AdaptiveVariation}), examples of such an effect within species on a recent timescale are less common. As large genome-scale datasets from humans and other species become further available, the framework and methodology described here can be used to understand how widespread the phenomenon of adaptive gene-flow is.