The MozAtlas orthology framework is similar to, but less restrictive, than other genome-wide orthology pipelines currently available , and includes a set of outgroup species specifically selected to represent the Insecta. Gene-family relationships were identified using all-vs-all Blast (Blossum 42 matrix) searches and Inparanoid clustering, with the longest available translations for annotated proteins. Best reciprocal hits between species were grouped together via single linkage clustering and the sequences aligned with MUSCLE. Tree topologies were subsequently reconstructed with both dS (synonymous substitution rate), dN (nonsynonymous substitution rate), nucleotide and protein distance measures using TreeBest (NJ and PHYML). Orthology classifications were finally described as one-to-one, one-to-many and many-to-many gene relationships, while gene paralogues were characterized on the basis of when divergence occurs.