Evolution of genes related to temperature adaptation in Drosophila melanogaster as revealed by QTL and population genetics analyses

The fixation of beneficial variants leaves genomic footprints
characterized by a reduction of genetic variation at linked neutral
sites and strong, localized allele frequency differentiation among
subpopulations. In contrast, for phenotypic evolution the effect of
adaptation on the genes controlling the trait is little understood.
Theoretical work on polygenic selection suggests that fixations of
beneficial alleles (causing selective sweeps) are less likely than
small-to-moderate allele frequency shifts among subpopulations.
This thesis encompasses three projects in which we have
experimentally addressed the issue of selective sweeps vs. allele
frequency shifts in the context of polygenic adaptation. We studied
three X-linked QTL underlying variation in chill coma recovery time
(CCRT), a proxy for cold tolerance, in Drosophila melanogaster from
temperate (European) and tropical (African) environments. The
analysis of these QTL was performed by means of selective sweep
mapping and quantitative complementation tests coupled with
expression assays. While the results of the selective sweep mapping
approach identified a gene (CG4491) that is unlikely to be
affecting CCRT, quantitative and gene expression analyses revealed
two linked candidate genes (brk and CG1677) that appear to differ
in their evolutionary histories. We found that the difference in
expression of the gene brk between populations affects CCRT
variation. Cold tolerant flies from the temperate zone have a lower
expression of this gene than cold sensitive flies from the tropics.
We found that a likely cause of this difference is variation in a
cis-regulatory element in the brk 5’ enhancer region. Sequence
variants in this element exhibit moderate frequency differences
between populations from temperate and tropical environments,
forming two latitudinal clines: one from the equator to the north
and another one in opposite direction to the south. In contrast,
the other gene within the same QTL (CG1677), which is linked to
brk, showed no measurable effect on cold tolerance but is a likely
target of strong positive selection leading to a selective sweep in
the European population. These results are consistent with the
aforementioned theoretical predictions about footprints of
selection in polygenic adaptation. They are also proof of the
conceptual bias incurred when identifying candidate genes within a
QTL via selective sweep mapping, at least in naturally evolving
populations. The challenge for the evolutionary genetics community
in the coming years is to develop statistical tools that are as
powerful and robust as those already available to map selective
sweeps to identify sites in the genome where allele frequency
shifts have occurred due to adaptive evolution at the phenotypic
level. Finally, the last section of the results is a report of a
new population genetics dataset. It consists of a collection of 80
inbred lines from a natural D. melanogaster population in Sweden
and 19 full genome sequences derived from this sample. We hope this
material will provide us with further insight into the processes
underlying adaptation to novel and stressful environments.