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.

Positive Selection Analysis. Finally, we looked within bat genomes for signatures of positive selection acting on protein-coding genes. We used the PAML software package (61) to identify dN/dS > 1 codons within each of the orthologous gene alignments that we created. We used codon models M8 and M8a, which estimate the evolutionary pressures that have acted at specific codon sites over the entire phylogeny of species included. These models do not consider unique evolutionary scenarios that have affected gene evolution along different branches over the phylogenetic tree like some other models. In M8, the data in each multiple sequence alignment is fit to a model where one group of codons from within the alignment is allowed to evolve with dN/dS > 1. M8a is the null version of this model, where that same category of codons is allowed but is constrained to dN/dS = 1. Because there is one additional degree of freedom in M8, the data will usually fit better to the M8 model (as estimated by a likelihood value). A gene was deemed to be evolving under positive selection when the data are a statistically better fit to M8 than to M8a (P < 0.05 after correction for multiple tests; Dataset S1). We performed this analysis on 10,650 genes that met criteria of having at least six species and 30 codons represented in the alignment. Out of these, 181 genes met the statistical criteria for being subject to positive natural selection, and these genes were hand-curated for function (Dataset S1). Table 2 shows a subset of the genes. Nineteen percent are known to be involved in immunity or the replication cycles of pathogens. Surprisingly, 8% are known to be involved in collagen formation, including 10 out of the 27 top scoring genes for positive selection (Dataset S1). We also looked at the gene ontology (GO) classifications which are overrepresented in the list of genes under positive selection and present this data in Dataset S2. The GO categories most enriched for positively selected genes are dominated by immune responses and collagen formation. *These genes are placed in two categories on this table. †Relatively little is known about these genes, but they are expressed solely in the testis.