Ancient tree-topologies and gene-flow processes among human lineages in Africa

The deep history of human evolution in Africa remains intensely debated, with increasingly complex models being proposed. To investigate this, we sequenced and analysed 73 novel high-quality whole genomes from 14 Central and Southern African populations with diverse cultural practices. Using extensive simulations and machine-learning Approximate Bayesian Computation (ABC), we jointly reconstructed their demographic history of divergences and migrations. We find extensive genome-wide diversity within and among populations, including substantial local genetic differentiation not fully explained by geography or cultural practices. These patterns highlight the importance of explicitly considering local genomic diversity when reconstructing human evolutionary history. We find that tree-like population histories with long periods of drift separated by short pulses of unidirectional gene-flow better explain the data than continuous gene-flow. Without invoking archaic admixture, our models accurately fit observed genomic variation and identify multiple episodes of gene-flow coinciding with major ecological and cultural changes in Sub-Saharan Africa.