Dr. Begun's basic question is what kind of natural selection is acting on DNA sequence variation. While a graduate student he discovered what has become the single most valuable contrast in this important area of inquiry. He found that regions of the D. melanogaster genome that experience low levels of crossing-over also reveal low levels of DNA sequence variation. He also showed that this could not be explained by variation in the rate of mutations in the two regions, as the rate of DNA divergence between species does not vary with recombination rate. Dr. Begun realized that this pattern must be due to some kind of interaction between natural selection and linkage. Since the time of his initial discoveries, quite a number of papers and research programs have been dedicated to the question of what kind (or kinds) of natural selection are the cause of the patterns that Dr. Begun discovered. At the moment, the principle contenders are the hitchhiking model, whereby beneficial mutations sweep through populations and remove variation from regions tightly linked to the site of the beneficial mutation, and the background selection model, whereby deleterious mutations remove haplotypes from having potential to leave descendants. Under the background selection model, the deleterious mutation rate may be quite high for a large block of linked genes, and so many, perhaps most, haplotypes may be linked to deleterious mutations. Only the fraction that are not have much chance of segregating variation at appreciable frequency, and this fraction may be quite small. The application carefully dissects the difficulties of testing these models, and it reviews the problematic nature of measuring recombination and mutation rates. Dr. Begun reasons that for many questions, the most useful contrasts that can be made are in Drosophila simulans which is likely to have less extraneous factors affecting recombination rates (e.g. segregating inversions) than does D. melanogaster. The core of the proposal is to carefully measure synonymous, and replacement polymorphisms in four well characterized regions of the genome in both melanogaster and simulans. These regions each have many genes that have already been sequenced, and they have been characterized for recombination rates in both species. Dr. Begun expects that the data from D. simulans in particular will reveal with sufficient confidence the relative levels of X and autosome variation. Under the background selection model, if deleterious mutations are partially recessive (as is probably true) then the magnitude of the effect of background selection should be greater for autosomes than for X-linked genes. Gene copies with linked recessive deleterious mutations cannot ultimately get into high frequency on either the X or an autosome, but they can drift higher in frequency on autosomes to the extent that they are sheltered by recessivity. Under hitchhiking of favorable mutations, the degree of difference expected between X-linked and autosomal genes depends on the dominance of the mutations: if they are recessive then they are more likely to fix if they are X-linked; if they are additive then the X and autosomes should experience the same rate of fixation (and thus of hitchhiking); and if they are dominant, then the autosomes should experience a larger effect. Another major question is how does the ratio of replacement to silent variation differ between the species as a function of recombination. A critical observation in this regard is that D. simulans appears to be segregating a fairly large excess of synonymous site (i.e. within protein coding regions but not affecting amino acid sequence) variation, relative to replacement (i.e. changing amino acid) variation, on the autosomes. D. simulans has slightly more variation of both categories as does D. melanogaster on the X chromosome (the ratio is the same in both species). But on the autosomes, D. simulans has proportionately much more synonymous variation, either given melanogaster levels or given autosomal replacement levels in D. simulans. Given this observation, Dr. Begun expects D. simulans to have a higher ratio or replacements to synonymous in regions of low recombination than it does regions of high recombination. Dr. Begun hopes to sequence approximately 100 loci (about 1000 bp each) in 14 lines from each species (in each, 7 from African and 7 out of Africa). These data will also reveal a wealth of information on variation among loci for different kinds of polymorphisms, as well as the way this variation corresponds to divergence between the taxa.
