A metaanalysis of bat phylogenetics and positive selection based on genomes and transcriptomes from 18 species

Hawkins, John A.; Kaczmarek, Maria E.; Müller, Marcel A.; Drosten, Christian; Press, William H.; Sawyer, Sara L.

Third, we developed the Mismatching Isoform eXon Remover (MIXR) software package to detect and remove exons with alternate consensus runs directly (available at http://github.com/ hawkjo/mixr). By alternate consensus run, we mean the situation as shown in Fig. 2A , where some subset of species has a long run of amino acids which are internally consistent but which disagree from the majority of the sequences. While no substitution pattern in a single column of an alignment can distinguish real biological mutations from artifactual mismatching of isoforms, runs of an alternate consensus sequence are exponentially unlikely as a function of run length according to all site substitution models, so we want to remove them before analysis. Briefly, the MIXR algorithm works as follows. First, we define the alternateconsensus-run score, designed to give high scores to runs where some subset of species is internally consistent but differs from the majority of species. We then find the maximum alternateconsensus-run score for all species bipartitions (divisions of the species into two subsets) in a given alignment, such as the bipartition into transcriptomic and genomic species in Fig. 2A . Finally, exons with significant scores are removed. See Methods for additional details.