Estimating selection by comparing synonymous and substitutional changes.

A higher ratio of substitutional to synonymous changes in between-species than in within-species comparisons has been taken as evidence for positive selection changing amino acids. A model is presented in which a difference of this kind arises as a result of purely neutral mutations, provided that the "species" compared are sufficiently different to approach a steady state between forward and backward mutation. In Neisseria, substitutions are twice as frequent, relative to synonymous changes, in between-species comparisons: it is shown that the data are consistent with the neutral model. The argument does not invalidate evidence for positive selection, for example in Drosophila, when the species compared are fairly similar.