Dan Hartl was inducted into the National Academy of Sciences of the USA in 2005 for his contributions to the field of evolutionary genetics. His inaugural article as an Academy member was published this week in the Proceedings of the National Academy of Sciences (PNAS, pronounce pee-nas, hehe). Hartl and colleagues compared levels of polymorphism in 91 genes from Drosophila melanogaster with divergence from D. simulans. They were interested in determining the effect of natural selection on nonsynonymous mutations, kind of like what was done in this paper (reviewed here).
Hartl and colleagues estimate that nearly all (~95%) potential amino acid changes are deleterious and the majority of amino acid polymorphisms (~70%) are deleterious. But the vast majority (~95%) of amino acid changes between species are adaptive. Most are, however, only mildly beneficial, suggesting that they are nearly neutral (see, these models aren’t complete bullshit). In comparison, Shapiro et al estimate that nearly a third of all amino acid fixations between Drosophila species are adaptive.