The distribution of integer partitions in human genome-wide genealogies reflects gene flow from a super-archaic lineage

Abstract

Several recent studies have found evidence for ancient gene flow between the ancestors of modern humans and an unsampled super-archaic lineage. Here we present a new, simple approach for characterising this process given genome-wide genealogies sampled from a single population. We summarise genealogies as distributions of integer partitions and show that this captures the temporal signal of tree imbalance left by ancient gene flow. We analyse genealogies from modern humans and find that the integer partition distributions are inconsistent with a history of panmixia but can be explained by gene flow from a super-archaic lineage. Our analysis favours a model of continuous gene flow over pulse-admixture and also recovers a bottleneck in the ancestors of modern humans. This work highlights a clear signal of ancient structure in genealogies of modern humans and provides an inference approach that complements existing methods.