Empirical approaches to detecting the action of natural selection in Drosophila

This dissertation examines two aspects of how natural selection
shapes the amount and pattern of genetic variation within and
between species: (1) the role of positively selected alleles in
shaping the variation within and between subpopulations of a
subdivided species and (2) the influence of epistatic selection
operating on RNA secondary structures. First, the role of natural
selection in shaping the pattern of variation within and between
populations of the subdivided species Drosophila ananassae is
investigated. To delimit the spread of positively selected alleles
and characterize the role of natural selection in genetic
differentiation, sequence data was collected from a locus in a
region of low recombination for 13 populations, spanning a majority
of the species range of D. ananassae. The migration behavior of
this selected locus is compared to that of 10 independent neutrally
evolving loci and tested against alternative models of natural
selection. Second, nucleotide variation at the D. melanogaster
bicoid locus is examined. The presence of a large, conserved
secondary structure in the 3’ untranslated region enables the
relationship between RNA secondary structure and patterns of
standing variation in natural populations to be explored. Variation
within this structure is analyzed with respect to models of
compensatory evolution and recent improvements of these models.
Evidence suggests that bicoid may be the result of a relatively
recent gene duplication in the Dipteran lineage, thus, variation in
the bicoid coding region is also analyzed with respect to the
evolutionary processes that may be ongoing if this gene is still
undergoing diversification and/or refining of its function.
Finally, long-range compensatory interactions between the two ends
of Drosophila alcohol dehydrogenase (Adh) mRNA are investigated by
experimental manipulation. Site-directed mutations were introduced
in the D. melanogaster Adh gene in an effort to explain why
previous mutational analysis failed to fit Kimura’s classical model
of compensatory evolution. The results of the mutational analysis
indicate that a classical result was not observed due to the
pleiotropic effect of changing a nucleotide involved in both
long-range base pairing and the negative regulation of gene
expression.