PYF: a multi-functional algorithm for predicting production and optimizing metabolic engineering strategy in Escherichia coli microbial consortia.

Simulating production in microbial consortia is crucial for optimizing metabolic engineering strategies to achieve high yields. However, existing algorithms for modeling polymicrobial metabolic fluxes, based on genome-scale metabolic networks, often overlook the conflicts and coordination between biosynthesis tasks and self-growth interests, leading to limited prediction accuracy. This study introduces the Polymicrobial cell factory Yield Forecasting (PYF) algorithm, which simulates the relationships between biosynthesis and growth more effectively by incorporating the expression degrees of biosynthesis pathways. PYF was shown to accurately predict the production of Escherichia coli-E. coli consortia under various scenarios, including mono-metabolite exchange, dual-carbon sources, and dual-metabolite exchange. The results revealed a mean relative error (MRE) of 0.106, an average determination coefficient of 0.883, and an average hypothesis testing parameter of 0.930 between predicted and experimental productions. Compared with the recent metabolic simulation algorithm, PYF reduced the MRE by ~61.6%. PYF is adaptable and enables accurate simulation even without enzyme catalytic data. Meanwhile, PYF rapidly analyzed and optimized metabolic engineering strategies through sensitivity analysis. By eliminating the need for specialized division and integration of polymicrobial metabolic networks, PYF greatly simplifies the simulation process, offering a novel approach for predicting and enhancing production in microbial consortia.