Genomic screening of bacteria is common practice to select strains with desired properties. However, 40-60% of all bacterial genes are still unknown, making capturing the phenotype an important part of the selection process. While omics-technologies collect high-dimensional phenotypic data, it remains challenging to link this information to genomic data to elucidate the impact of specific genes on phenotype. To this end, we present Scoary2, an ultra-fast software for microbial genome-wide association studies (mGWAS), enabling integrative data exploration. As proof of concept, we explore the metabolome of 44 yogurts with different strains of Propionibacterium freudenreichii, discovering two genes affecting carnitine metabolism.
