Adaptive radiations are often stereotypical, as populations repeatedly specialize along conserved environmental axes. Phenotypic plasticity may be similarly stereotypical, as individuals respond to environmental cues. These parallel patterns of variation, which are often consistent across traits, have led researchers to propose that plasticity can facilitate predictable patterns of evolution along environmental gradients. This “flexible stem” model of evolution raises questions about the genetic nature of plasticity, including: How complex is the genetic basis for plasticity? Is plasticity across traits mediated by many distinct loci, or few “global” regulators? To address these questions, we reared a hybrid cichlid mapping population on alternate diet regimes mimicking an important environmental axis. We show that plasticity across an array of ecologically relevant traits is generally morphologically integrated, such that traits respond in a coordinated manner, especially those with overlapping function. Our genetic data are more ambiguous. While our mapping experiment provides little evidence for global genetic regulators of plasticity, these data do contain a genetic signal for the integration of plasticity across traits. Overall, our data suggest a compromise between genetic modularity, whereby plasticity may evolve independently across traits, and low-level but widespread genetic integration, establishing the potential for plasticity to experience coordinated evolution.

Full methods can be found in the article in Evolution. In brief, we used a hybrid mapping cross of cichlid fish to assess the genetic basis of diet-induced plasticity. F3 hybrid animals resulting from a cross between a Labeotropheus fuelliborni female and a Tropheops sp. "redcheek" male were split into two diet treatments. These hybrids were then sacrificed, dissected, & analyzed for a variety of phenotypic traits, including both linear & geometric morphometrics. Continuous traits resulting from the geometric morphometric analyses were subjected to a canonical variates analysis, which produced an axis arraying individuals from more plastic animals at each extreme, to less plastic animals near the origin. These values were then used as the basis for a series of QTL analyses aimed at linking plasticity to its underlying genetic basis. Finally, both the trait values and the genetic associations were used in two separate network analyses to determine whether plasticity was correlated across the body at either the morphological or genetic level.

For each trait, I have provided two TPS files - one with the unadjusted landmark coordinates (named: "Trait X.tps"), and one with the allometry-adjusted landmark coordinates (named: "Trait X SL Corr.tps"). I also include the file with the standard length values for each individual ("SL.csv") - these were used in the size standardizations. I have also provided the files used in the network analyses based on shape ("Shape Scores for Network Analysis.csv") and genetics ("LOD Scores for Network Analysis.csv"). These provide the shape scores for each individual for each trait and the LOD scores at every position sampled along the genome for each individual for each trait, respectively. Finally, I have provided the QTL map with all 21 phenotypes that were mapped ("QTL Map with Traits.csv").