Fig. 3: Putative genomic hotspots containing different flatwing-associated loci in cricket populations on three Hawaiian islands.

a Flatwing association mapping. Data from a previously published flatwing QTL²¹ are shown on top, with red symbols showing flatwing-associated SNPs and blue vertical lines the extent of the QTL on the X-linkage group (left-most block of black markers). Horizontal dashed lines indicate FDR-corrected significance thresholds, and alternating shades of black and blue indicate different linkage groups. For the lower three panels, the most significant markers are defined as described in the main text using the strictest criterion and plotted in red. b Venn diagram showing candidate X chromosome scaffolds associated with the flatwing phenotype for each population, all detected using the most stringent Oahu criterion. c Unrooted maximum-likelihood consensus tree for the two flatwing-associated scaffolds shared by all islands. d Rooted maximum-likelihood consensus tree for all 1 kb flanking regions of all flatwing-associated SNPs. c, d Normal-wing males are represented with solid circles and flatwing males with open circles. Clustered individuals (≥2) are highlighted using filled (normal-wing) and open (flatwing) boxes if they were from the same island population and shared the same wing phenotype. Black triangles represent the individual used as the reference genome²¹. Orange rectangle represents seven Australian individuals, which were used as an outgroup. e Violin plot of net genetic distance (d) outside and within flatwing-associated regions on chromosome X between flatwing Kauai and flatwing Oahu samples. Distances were calculated in 10 kb sliding windows with 2.5 kb steps (n = 66,211 and 437 windows, respectively). Values for upper quartile, median, and lower quartile, plus 1.5× interquartile ranges, are indicated with box plots in the middle of the violin plots. Additional statistical details are provided in Supplementary Fig. 26a.