Résumé
This article has been peer-reviewed and recommended by Peer Community In EvolutionaryBiology (http://dx.doi.org/10.24072/pci.evolbiol.100037).Lateral gene transfers between ancient species contain information about the relative timingof species diversification. Specifically, the ancestors of a donor species must have existedbefore the descendants of the recipient species. Hence, the detection of a transfer event can betranslated into a time constraint between nodes of a phylogeny if the donor and recipient canbe identified. When a set of transfers is detected by interpreting the phylogenetic discordancebetween gene trees and a species tree, the set of all deduced time constraints can be usedto rank the species tree, i.e. order totally its internal nodes. Unfortunately lateral genetransfer detection is challenging and current methods produce a significant proportion offalse positives. As a result, often, no ranking of the species tree is compatible with thefull set of time constraints deduced from predicted transfers. Here we propose a method,implemented in a software called MaxTiC (Maximum Time Consistency), which takes asinput a species tree and a series of (possibly inconsistent) time constraints between its internalnodes, weighted by confidence scores. MaxTiC outputs a ranked species tree compatible witha subset of constraints with maximum cumulated confidence score. We extensively tested themethod on simulated datasets, under a wide range of conditions that we compare to measureson biological datasets. In most conditions the obtained ranked tree is very close to the realone, confirming the potential of dating the history of life with transfers by maximizing timeconsistency. MaxTiC is freely available, distributed along with a documentation and severalexamples: https://github.com/ssolo/ALE/tree/master/maxtic.