North Africa demographic history through an ABC-DL approach.

Previous studies conducted on North African human populations have identified a complex demographic scenario in the region due to the location as a crossroad between the Mediterranean Sea, Middle East and the Sahara desert. The presence of an autochthonous genetic component, plus extensive migrations and gene flow from the Middle East, Europe and Sub-Saharan Africa, have shaped the genetic composition of its people through time. We built a demographic model based on deep learning in an Approximate Bayesian Computation (ABC-DL) framework to deduce the evolutionary history of North Africa populations, from the origin of the population to infer the effect of the admixture events from the surrounding populations, using whole genome data. The ABC framework eludes the need of estimating the likelihood of the data given the demographic model, like in traditional Bayesian methods, which could become too difficult in complex demographic scenarios like the one in North Africa. Our results support a model where North Africans present a most recent common ancestor with Middle Easterns and that increasingly complex models, with multiple demographic events, perform better than simpler ones in explaining the data.

Dec 15, 2022 12:00 AM

X Bioinformatics and Genomics Symposium