Search for a command to run...
The emergence of automated methods for the generation of genome-scale metabolic models (GEMs) has enabled the use of these models to study metabolic differences between large sets of different microorganisms. Current methods are often optimized for either handling large sets of strains or highly accurately reflecting the metabolism of selected strains; however, both aspects are necessary for analyzing the metabolic differences among large numbers of strains of the same or closely related species. In this study, we present a workflow for the high-throughput generation of high-quality GEMs of closely related strains, which has been applied for the generation of 439 GEMs of <i>Lactococcus lactis</i> and <i>Lactococcus cremoris</i> strains. Comparison of the resulting GEMs under different growth conditions revealed metabolic differences between the strains in carbon source utilization, fermentation products, and nutrient requirements. Notably, <i>L. lactis</i> and <i>L. cremoris</i> showed differences in xylose and ribose utilization pathways, with over 90% of <i>L. cremoris</i> strains unable to utilize xylose or showing limited ribose utilization due to a lack of key enzymes in the pentose phosphate pathway. Additionally, the strain-specific GEMs predicted differences in amino acid auxotrophies known in <i>Lactococcus</i>, including cysteine, which was possible to synthesize in only 11% of the strains and was found advantageous for growth in milk. The workflow presented in this study enables the generation of GEMs that can be used for comparing a large number of closely related strains and facilitates the assessment of their suitability in different biotechnological applications.IMPORTANCEComparative studies of genome-scale metabolic models (GEMs) for a large set of different strains of the same bacterial species can play a crucial role in uncovering species-specific metabolism and understanding strain-specific metabolic variations. Additionally, results from these analyses can be used as an aid in industrial microbiology for selecting strains with the desired metabolic traits. In this study, we present a method for high-throughput generation and comparison of strain-specific GEMs and apply this method to 439 <i>Lactococcus</i> strains, which is a highly important species used worldwide for food fermentation. Comparison of the strain-specific GEMs revealed metabolic differences between strains relevant for industrial application and furthermore showed the potential of these models for understanding microbial interactions between strains of the same species in co-cultivation.