Teleost gill arches are exquisitely evolved to maximize foraging efficiency, and include structures for the capture, filtering, and processing of prey. While both plasticity and a genetic basis for gill arch traits have been noted, the relative contribution of genetics and the environment in shaping these structures remains poorly understood. East African cichlids are particularly useful in this line of study due to their highly diverse and plastic feeding apparatus. Here we explore the gene-by-environmental effects on cichlid GRs by rearing pure bred species and their F₃ hybrids in different foraging environments. We find that anatomical differences between species are dependent on the environment. The genetic architecture of these traits is also largely distinct between foraging environments. We did, however, note a few genomic “hotspots” where multiple traits map to a common region. One of these, for GR number across multiple arches, maps to the ptch1 locus, a key component of the Hedgehog (Hh) pathway that has previously been implicated in cichlid oral jaw shape and plasticity. Since Hh signaling has not previously been implicated in GR development, we explored functional roles for this pathway.  Using a small molecule inhibitor in cichlids, as well as zebrafish transgenic systems, we demonstrate that Hh levels negatively regulate GR number, and are both necessary and sufficient to maintain plasticity in this trait. In all these data underscore the critical importance of the environment in determining the relationship between genotype and phenotype, and provide a molecular inroad to better understand the origins of variation in this important foraging-related trait.

See methods for a full description. The data are largely morphological in nature, and include both counts, which were collected with the use of a dissecting microscope, and skeletal lengths, which were collected from digital images in the software program, Image J. The genetic data are presented as genotypes - "AA", "AB", and "BB". Polymorphisms were identified and genotyped via RAD-seq. 

All data are formatted for direct import into R - e.g., each sheet many be saved as a .txt or .csv file that may read into R. Missing data are represented as "NA". Abbreviations are as follows: Ben = benthic foraging treatment, Gli = zebrafish carrying the dominant-negative Gli2 transgene, GR = gill raker, LF = Labeotropheus fuelleborni, MZ = Maylandia zebra, Pel = pelagic foraging treatment (note that this treatment is refered to as Lim=limnetic foraging treatment in the manuscript; these terms may be used interchangeably), res = residual values, Shh = zebrafish carrying the Shha activator transgene, SL = standard length, TRC = Tropheop sp. "red cheek", WT = wild-type zebrafish.

The genetic data is organized as a matrix [indiviuals x genotype]. The first three cells in each column represent (1) the marker name "scaffold_X.########", whereby "X" refers to the specific scaffold number and "########" represents the base pair on that scaffold where the polymorphism is located, (2) the linkage group number, 1-22, and (3) the genetic position in centimorgans, cM. The LF allele is A, and the TRC allele is B.