Phylomes represent the complete collection of evolutionary histories for each gene encoded in a genome. Reconstruction of such large amount of trees has become feasible in recent years thanks to the increase in computation power and improvements in the reconstruction methods. It is now possible to generate those phylomes in an automatic way, generating thus large amounts of data that need to be stored. PhylomeDB was build in order to store all the data derived from phylome reconstruction. As such it now includes hundreds of thousands of trees and alignments. PhylomeDB has also been complemented in recent years with mappings to other external databases such as UniProt, Ensembl or Genolevures. Orthology and paralogy predictions have been extended to include metaPhOrs predictions (http://orthology.phylomedb.org/), which provides a consensus prediction of evolutionary relationships based not only on one tree but on multiple trees found across numerous databases (PhylomeDB, EnsemblCompara, EggNOG, OrthoMCL, COG, Fungal Orthogroups, and TreeFam). In this last version, phylomeDB has enhanced the tree visualization module in order to show protein domains that were mapped onto the tree sequences. All the data can be downloaded either from the tree page, where information of a particular tree can be obtained, or through our ftp page (ftp://phylomedb.org/phylomedb/) for larger datasets.