#15 · Systems biology

Genome-scale models predict microbial cross-feeding, growth laws, and product yield

The dominant thread is the use of genome-scale and constraint-based metabolic models to predict microbial behavior from community to strain level. A seven-member maize root synthetic community simulated in exudate medium is reported to predict cross-feeding, with Enterobacter ludwigii identified as a community hub. A feature-learning approach coarse-grains metabolic networks to identify growth predictors conserved across species. A dynamic plus constraint-based model of Bacillus subtilis LY7.16 simulates sucrose effects on levan yield and substrate efficiency. A separate preprint modifies a microgravity cell-growth simulation to isolate gravity-driven flow and sedimentation. What is new is the coupling of learned representations with metabolic networks for coarse-graining, and the hybridization of dynamic and constraint-based formulations for identifying yield-efficiency strategies.