Metabolic network percolation quantifies biosynthetic capabilities across the human oral microbiome

Area of Science:

• Microbiology
• Computational Biology
• Systems Biology

Background:

• Microbial biosynthetic capabilities are crucial for growth, interactions, and community structure.
• Predicting metabolic functions in diverse microbial communities is challenging due to uncertainties in metabolic pathways and environmental conditions.

Purpose of the Study:

• To develop a probabilistic computational method for quantifying biosynthetic robustness of microbial metabolic networks.
• To create an atlas of predicted metabolic capabilities for human oral microbes and analyze their ecological implications.

Main Methods:

• A probabilistic method inspired by percolation theory was used to assess metabolic network robustness under variable environments.
• The method was applied to genome-derived metabolic networks of 456 human oral microbes to predict the production of 97 metabolites.

Main Results:

• An atlas of predicted biosynthetic capabilities for 97 metabolites across 456 human oral microbes was compiled.
• The atlas revealed taxonomically-related trends in biomass composition and enabled estimation of inter-microbial metabolic distances correlating with co-occurrences.
• A distinct cluster of fastidious/uncultivated taxa, including Saccharibacteria (TM7), was identified with significant metabolic deficiencies.

Conclusions:

• The developed probabilistic approach effectively quantifies metabolic robustness and embraces uncertainty in complex microbial ecosystems.
• The created atlas provides insights into microbial community structure, metabolic interactions, and the characteristics of understudied microbial groups.