2Many animal species are comprised of discrete phenotypic forms. Understanding the genetic mechanisms generating and maintaining such phenotypic variation within species is essential to comprehending morphological diversity. A common and conspicuous example of discrete phenotypic variation in natural populations of insects is the occurrence of different colour patterns, which has motivated a rich body of ecological and genetic research [1][2][3][4][5][6] . The occurrence of dark, i.e. melanic, forms, displaying discrete colour patterns, is found across multiple taxa, but the underlying genomic basis remains poorly characterized. In numerous ladybird species (Coccinellidae), the spatial arrangement of black and orange patches on adult elytra varies wildly within species, forming strikingly different complex colour patterns 7,8 (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint . http://dx.doi.org/10.1101/345942 doi: bioRxiv preprint first posted online Jun. 13, 2018; 3 wild has been attributed to a combination of allelic diversity, interactions between allelic forms, and plastic response to environmental factors 12,13 . Genetic crosses have demonstrated that the majority of H. axyridis melanic forms result from variation of multiple alleles segregating at a single, uncharacterised, autosomal locus 9,10 , hereafter referred to as the colour pattern locus.To identify this colour pattern locus, and the mechanisms underlying discrete colour pattern variation, we used a population genomics approach, taking advantage of the cooccurrence of multiple colour pattern forms in natural populations. To that end, we first performed a de novo genome assembly of the H. axyridis Red-nSpots form (HaxR) using long reads produced by a MinION sequencer (Oxford Nanopore) (see Extended Data Table 1).Then, to fine map the colour pattern locus on this assembly, DNA from 14 pools of individuals (from n=40 to n=100 individuals per pool) representative of the world-wide genetic diversity and the four main colour pattern forms of H. axyridis was sequenced on a HiSeq 2500 (Illumina, Inc.) (Extended Data Table 2). Our aim was to characterize genetic variation associated with phenotypic differences across pool samples, using the proportion of individuals of a given colour form in each pool as a covariate. To do so, we called 18,425,210autosomal SNPs, and we performed a population-based genome-wide association study accounting for the covariance of allele frequencies across pools 14 . Given the dominance hierarchy of colour morph alleles with Black-2Spots > Black-4Spots > Black-nSpots > RednSpots 9,12,13 , we first performed the association study using the proportion of Red-nSpots individuals, carrying two copies of the most recessive allele, in each DNA pool as a covariate.We found 710 SNPs strongly associated with the proportion of the Red-nSpots form (Bayes factor > 30 db), the vast majority (86%) of which are located within a s...