Assembling the hamlet genome de novo to uncover the genomic architecture of divergence in Caribbean reef fish (Hypoplectrus spp)

Studies on adaptive radiations have helped to shape our understanding of speciation processes. Although extensive research has been done on a vast variety of organisms, the marine environment has been generally neglected and studies that investigate marine systems are urgently needed. In the era of next generation sequencing, molecular based evolutionary research is increasingly shifting to whole genome based approaches. Therefore, systems used to fully address the marine knowledge gap should be studied from a genomic perspective. The Caribbean hamlets (Hypoplectrus spp., Serranidae) present a promising system spanning the whole speciation continuum from genetically (still) indistinguishable to well diverged species and present an ideal setting to study rapid speciation processes. They already have a rich history of genetic research but currently a reference genome is still missing. Here we present a high quality chromosome level assembly of the barred hamlet genome (Hypoplectrus puella) that will serve as reference for future whole genome based studies on the serranids. The assembly has a total size of 612 mega bases, is comprised of 24 linkage groups and the mitochondrial genome and shows a high level of completeness (93.1% of expected core genes detected). It has a convincing degree of synteny with the stickleback (Gasterosteus aculeatus) and shows a remarkable stretch of low recombination (possibly an inversion) on LG08. To demonstrate the potential provided by this new reference genome, this thesis also includes an exemplary re-analysis of previously published RAD data. The re-analysis expanded the catalog of genes of interest regarding the hamlet speciation since it showed that a whole suite of genes believed to affect reproduction, coloration and behavior are located in direct proximity of divergence peaks. Finally the analysis also showed that methods providing finer resolution than RAD are needed to depict drivers of divergence in hamlets.