Phylogeny-aware simulations suggest a low impact of unsampled lineages in the inference of gene flow during eukaryogenesis.

Abstract

The topologies of gene trees are broadly used to infer horizontal gene transfer events and characterize the potential donor and acceptor partners. Additionally, ratios between branch lengths in the gene tree can inform about the timing of transfers relative to each other. Using this approach, recent studies have proposed a relative chronology of gene acquisitions in the lineage leading to the last eukaryotic common ancestor (LECA). However, a recognised caveat of the branch-length ratio method are potential biases due to incomplete taxon sampling resulting in so-called "ghost" lineages. Here, we assessed the effect of ghost lineages on the inference of the relative ordering of gene acquisition events during eukaryogenesis. For this, we used a novel simulation framework that populates a dated Tree of Life with plausible "ghost" lineages and simulates their gene transfers to the lineage leading to LECA. Our simulations suggest that a substantial majority of gene acquisitions from distinct ghost donors are inferred with the correct relative order. However, we identify phylogenetic placements where ghost lineages would be more likely to produce misleading results. Overall, our approach offers valuable guidance for the interpretation of future work on eukaryogenesis, and can be readily adapted to other evolutionary scenarios.