Multiple sequence alignments
We used Mafft v. 7.475 (Katoh & Standley, 2013) to build three multiple sequence alignments (MSA) per each gene family. First, we generated a MSA (MSAc) with only complete sequences identified in D. silvatica using the L-INS-i algorithm with options –localpair –maxiterate 1000. Then, we built a second MSA (MSAf) for all sequences (the full data set, comprise the complete and partial copies identified in D. silvatica ) using the Mafft –addfragments option (–keeplength) to align fragment (partial) sequences to the previous computed MSAc. Finally, we also used the L-INS-i algorithm to generate a third MSA for each family, which included the complete sequences from D. silvatica and the curated members of the same family annotated in the genome of Drosophila melanogaster Meigen, 1830 (MSAp).