Abstract
Inferring gene trees is difficult because alignments are often too short, and thus contain insufficient signal, while substitution models inevitably fail to capture the complexity of the evolutionary processes. To overcome these challenges species tree-aware methods seek to use information from a putative species tree. However, there are few methods available that implement a full likelihood framework or account for horizontal gene transfers. Furthermore, these methods often require expensive data pre-processing (e.g., computing bootstrap trees), and rely on approximations and heuristics that limit the exploration of tree space. Here we present GeneRax, the first maximum likelihood species tree-aware gene tree inference software. It simultaneously accounts for substitutions at the sequence level and gene level events, such as duplication, transfer and loss and uses established maximum likelihood optimization algorithms. GeneRax can infer rooted gene trees for an arbitrary number of gene families, directly from the per-gene sequence alignments and a rooted, but undated, species tree. We show that compared to competing tools, on simulated data GeneRax infers trees that are the closest to the true tree in 90% of the simulations in terms relative Robinson-Foulds distance. While, on empirical datasets, GeneRax is the fastest among all tested methods when starting from aligned sequences, and that it infers trees with the highest likelihood score, based on our model. GeneRax completed tree inferences and reconciliations for 1099 Cyanobacteria families in eight minutes on 512 CPU cores. Thus, its advanced parallelization scheme enables large-scale analyses. GeneRax is available under GNU GPL at https://github.com/BenoitMorel/GeneRax.
