Genome-wide architecture of adaptation in experimentally evolved Drosophila dataset

The molecular basis of adaptation remains elusive even with the current
ease of sequencing the genome and transcriptome. Previous attempts to
address this question have knocked out or knocked down genes, but these
approaches are not applicable to populations with moderate to high genetic
variation. By contrast, experimental evolution offers replicated outbred
populations with well-defined selection regimes and known evolutionary
histories. We used experimentally evolved populations of Drosophila in
conjunction with statistical learning tools to explore interactions
between the genome, the transcriptome, and phenotypes. Our results
indicate that transcriptomic measures from adult samples can predict
phenotypic characters at many adult ages, which is consistent with the
presumed quiescence of the Drosophila transcriptome during adulthood.
Importantly, when comparing the genome and transcriptome in predicting
phenotypic characters, we find that the two types of data are comparably
useful. When using genome sites as predictors for the expression of the
transcriptome, we find that gene expression is influenced by genomic
regions across all large chromosome arms. Conversely, we found many
genomic regions influencing the expression of numerous genes, which is
consistent with widespread pleiotropy. Our results also highlight the
power of the combination of experimental evolution, next-generation
sequencing, and statistical learning tools in exploring the molecular
basis of adaptation.