Abstract
Adaptation to human-induced environmental change has the potential to profoundly influence the genomic architecture of affected species. This is particularly true in agricultural ecosystems, where anthropogenic selection pressure is strong. Heliothis virescens feeds on cotton in its larval stages and populations in the Southern United States have been declining since the widespread planting of transgenic cotton in the late 1990s. These cotton cultivars endogenously express proteins derived from the bacterium Bacillus thuringiensis (Bt), which are lethal to H. virescens. No physiological adaptation to Bt toxin has been found, so adaptation to this altered environment could involve: 1) shifts in host plant selection mechanisms to avoid cotton, or 2) changes in detoxification mechanisms required for cotton-feeding versus feeding on other host plants. A decline in pyrethroid use in Bt cotton landscapes likely also led to reversion to susceptible alleles at loci involved in expression of pyrethroid resistance. Here we begin to address the question of whether such changes occurred in H. virescens populations between the years 1997 and 2012. In a proof of concept experiment, we confirmed that allele frequency changes at the voltage-gated sodium channel gene, a pyrethroid resistance locus, could be detected through a genomic scanning technique that depends on linkage disequilibrium between molecular markers and gene targets of selection. A direct PCR approach first confirmed a decline in frequency of the pyrethroid resistance allele in H. virescens populations over time. We then tested the hypothesis that this known genetic change could be detected via a ddRAD-seq enabled genome scan in concert with our new H. virescens genome assembly. One ddRAD-seq marker was physically linked to the sodium channel gene, and the rate of allele frequency change at that marker was similar to that of the sodium channel resistance allele. We then identified additional ddRAD-seq loci with significant allele frequency changes over the 15 year study period. Genes near these ddRAD-seq loci were identified and their potential contributions to adaptive phenotypes are discussed.