Results of the selection analyses for the best alignment method (PRANK), and two others (CLUSTAL and MUSCLE) and the MAP dataset of Solenopsis.

Given are the log-likelihoods of the null hypotheses (H0), which assume no selection, and alternative hypotheses (Ha), which allow for positive selection. Likelihood-ratio tests (LRT) of positive selection are conducted to compare the two hypotheses. Levels of significance are 3.84 at 5% and 6.63 at 1% for the site model and 3.84 at 5% and 5.99 at 1% for the branch and branch-site models, following the χ12 distribution to guide against violations of model assumptions. Statistically significant LRTs for positive selection are indicated by italics and *** for p≪0.01. Note that inference of positive selection greatly depends on the alignment method used.