Statistical Tools to Understand the Functional Landscape of Microbial BPA Degradation

Bisphenol-A (BPA) is a carcinogenic and endocrine-disrupting pollutant that shows poor promise as a bioremediation target due to bacterial susceptibility to toxicity, slow rates, and incomplete BPA metabolism. Although synthetic communities may overcome these challenges by having strains complement each other, there is a lack of understanding of the mechanism of BPA degradation as well as how degradation dynamics change with changing BPA concentrations. Here, we use statistical tools to uncover the BPA landscape to show that interactions become increasingly important as BPA concentration increases. We isolated 16 strains from various soil samples enriched in BPA and created random combinations of communities from them. We measured the BPA degradation of these communities from five initial BPA concentrations. PCA shows that the change in BPA degradation across concentrations is low-dimensional and has predictive properties. Ridge models show changes in strain coefficients consistent with increased epistasis and decreased importance of individual degraders as concentration increases. These results allow for the synthetic design of optimally BPA-degrading communities, and solidify a new approach to statistically understanding microbial community function landscapes