Abstract
In a recent study in Current Biology1, we investigated in detail the population genomics of 25 insular or continentally-distributed bird species using whole-genome sequences. As part of our investigations, we included simple ratios of non-synonymous to synonymous mutations, a commonly used measure of the strength of purifying selection on coding regions. Concerns regarding the statistical limits associated with the use of this ratio were raised by Kratochvíl and Rovatsos2. Here, our objective is to continue the debate, not only by introducing the advantages and the drawbacks of this measure, but also by proposing suitable alternatives. Applied to our datasets, this allowed us to strengthen our initial conclusions, in particular regarding the support for the nearly neutral theory.