Elucidating the genomic architecture of quantitative traits is essential for our understanding of adaptation and for breeding in domesticated organisms.Penicillium roquefortiis the mold used worldwide for the blue cheese maturation, contributing to flavors through proteolytic and lipolytic activities. The two domesticated cheese populations display very little genetic diversity, but are differentiated and carry opposite mating types. We produced haploid F1 progenies from five crosses, using parents belonging to cheese and non-cheese populations. Analyses of high-quality genome assemblies of the parental strains revealed five large translocations, two having occurred via a circular intermediate. Offspring genotyping with genotype-by-sequencing (GBS) revealed several genomic regions with segregation distortion, possibly linked to degeneration in cheese lineages. We found transgressions for several traits relevant for cheese making, with offspring having more extreme trait values than parental strains. We identified quantitative trait loci (QTLs) for colony color, lipolysis, proteolysis, extrolite production, including mycotoxins, but not for growth rates. Some genomic regions appeared rich in QTLs for both lipid and protein metabolism, and other regions for the production of multiple extrolites, indicating that QTLs have pleiotropic impacts. Some QTLs corresponded to known biosynthetic gene clusters, e.g., for the production of melanin or extrolites. F1 hybrids constitute valuable strains for cheese producers, with new traits and genetic diversity, and allowed identifying target genomic regions for traits important in cheese making, paving the way for strain improvement. The findings further contribute to our understanding of the genetic mechanisms underlying rapid adaptation, revealing convergent adaptation targeting major regulators.