The degree to which loci promoting reproductive isolation cluster in the genome – i.e. the genetic architecture of reproductive isolation - can influence the tempo and mode of speciation. Tight linkage between these loci can facilitate speciation in the face of gene flow. Pheromones play a role in reproductive isolation in many Lepidoptera species, and the role of endogenously-produced compounds as secondary metabolites decreases the likelihood of pleiotropy associated with many barrier loci. Heliconius butterflies use male sex pheromones to both court females (aphrodisiac wing pheromones) and ward off male courtship (male-transferred anti-aphrodisiac genital pheromones), and it is likely that these compounds play a role in reproductive isolation between Heliconius species. Using a set of backcross hybrids between H. melpomene and H. cydno, we investigated the genetic architecture of putative male pheromone compound production. We found a set of 40 significant quantitative trait loci (QTL) representing 33 potential pheromone compounds. QTL clustered significantly on two chromosomes, chromosome 8 for genital compounds and chromosome 20 for wing compounds, and chromosome 20 was enriched for potential pheromone biosynthesis genes. There was minimal overlap between pheromone QTL and known QTL for mate choice and color pattern. Nonetheless, we did detect linkage between a QTL for wing androconial area and optix, a color pattern locus known to play a role in reproductive isolation in these species. This tight clustering of putative pheromone loci might contribute to coincident reproductive isolating barriers, facilitating speciation despite ongoing gene flow.

This dataset consists of n major data subsets: pheromone analysis of parent species Heliconius cydno and H. melpomene and F1 and backcross individuals between the two species (raw GC-MS output files and data tables; note that genital data are included here but parental and cross wing pheromone data are separately archived in Dryad at https://doi.org/10.5061/dryad.crjdfn31b); androconial analysis of the same individuals (data table and image files); and QTL mapping information (linkage maps and QTL analysis scripts).

GC-MS data were analyzed with AMDIS to produce data tables with compound amounts. Image files of wing area were analyzed using GIMP (GNU Image Manipulation Program) for hindwing and hindwing androconial area. QTL linkage maps were produced from whole genome data (separately archived on ENA under project PRJEB34160) and scripts for analysis of QTL using the R/qtl2 package are included.

Raw GC-MS files may be viewed and analyzed in the free analysis software AMDIS. "MP" in files and filenames refers to H. melpomene (MP from "melpomene Panama") and "CP" to H. cydno.