A new clustering method for building multiple trees using deep learning.

Each gene has its own evolutionary history which can substantially differ from the evolutionary histories of other genes. For example, some individual genes or operons can be affected by specific horizontal gene transfer or hybridization events. Thus, the evolutionary history of each gene should be represented by its own phylogenetic tree which may display different evolutionary patterns from the species tree, or Tree of Life, that represents the main patterns of vertical descent. Here, we present a new efficient method for inferring single or multiple consensus trees and supertrees for a given set of phylogenetic trees (i.e. additive trees or X-trees). The output of the traditional tree consensus methods is a unique consensus tree or supertree. Here, we show how Machine Learning (ML) models, based on some interesting properties of the Robinson and Foulds topological distance, can be used to partition a given set of trees into one (when the data are homogeneous) or multiple (when the data are heterogeneous) cluster(s) of trees. We adapt the popular Accuracy, Precision, Sensitivity, and F1 scores to the tree clustering. A special attention is paid to the relevant, but very challenging, problem of inferring alternative supertrees that are built from phylogenies defined on different, but mutually overlapping, sets of species. The use of an approximate objective function in clustering makes the new method faster than the existing tree clustering techniques and thus suitable for the analysis of large genomic datasets.