On the utility of short intron sequences as a reference for the detection of positive and negative selection in drosophila

The detection of selection, both positive and negative, acting on a DNA sequence or class of nucleotide sites requires comparison with a reference sequence that is unaffected by selection. In Drosophila, recent findings of widespread selective constraint, as well as adaptive evolution, in both coding and noncoding regions highlight the difficulties in choosing such a reference sequence. Here, we investigate the utility of short intron sequences as a reference for the detection of selection. For a set of 119 Drosophila melanogaster genes containing 195 short introns (≤120 bp), we analyzed polymorphism and divergence at 1) 4-fold synonymous sites, 2) all sites of introns ≤120 bp, 3) all sites of introns ≤65 bp, 4) bases 8-30 of introns ≤120 bp, and 5) bases 8-30 of introns ≤65 bp. The last class of sites shows the highest levels of both interspecific divergence and intraspecific polymorphism, suggesting that these sites are under the least selective constraint. Bases 8-30 of introns ≤65 bp also have the lowest ratio of divergence to polymorphism, which may indicate that a small proportion of substitutions in the other classes of sites are the result of adaptive evolution. Although there is little signal of selection on the primary sequence of short introns, patterns of insertion-deletion polymorphism and divergence suggest that both positive and negative selection act to maintain an optimal intron length.