First, models that allow dS to vary across alignment sites consistently provide highly significant improvements in goodness-of-fit, for example, in 9/10 cases examined by Kosakovsky Pond and Muse (2005) and in 42% of the almost 8,000 protein groups analyzed by Dimitrieva and Anisimova (2014) .
← all excerpts
Synonymous Site-to-Site Substitution Rate Variation Dramatically Inflates False Positive Rates of Selection Analyses: Ignore at Your Own Peril.
1
—
—