Fermented foods offer numerous health benefits that are not only attributed to the microorganisms they contain, but also to their metabolic products – known as microbial metabolites. These microbial metabolites significantly depend on the selection of bacterial strains used and their metabolic capacities. To explore this metabolic capacity, we compared the genomes of 600 strains derived from the Liebefeld Culture Collection, comprising >15,000 isolates mainly originated from the dairy industry, with the metagenome of the human gut microbiome in an in-silico study. We found that 24 strains from this collection were sufficient to cover 89% of all annotated enzymatic reactions of an average human microbiome. Therefore, we hypothesized that fermented foods could be specifically produced to functionally supplement the metabolic activity of a dysbiotic gut microbiota. Consequently, we enhanced the production of certain bioactive substances (e.g., vitamins, amino acid derivatives) in fermented milk by selecting strains based on the presence of genes encoding enzymes in a targeted metabolic pathway. The functionality of the fermented milk was subsequently tested in mouse models. In particular, we showed that the immune regulatory activity of the gut microbiota could be rescued by feeding germ-free mice with a fermented milk enriched in microbial tryptophan catabolites. Finally, we developed Scoary2, an ultra-fast microbial genome-wide association studies (mGWAS) software, to link the production of specific metabolites (e.g., carnitines) to the genetic content of the bacterial strains used for the fermentation. Our approach demonstrates that the combined analysis of bacterial genomes and food metabolomes using modern bioinformatic tools offers promising potential for the targeted development of functional fermented foods.