Quantifying patterns of microbial community assembly processes in bioreactors using different approaches leads to variable results

Abstract

Engineered bioreactors play a vital role in many processes to convert wastes to resources, such as biological wastewater treatment, bioremediation, and conversion of solid waste to methane in landfills. These biological systems rely on communities of microbes to convert waste to valuable resources. A central aspect of the design and operation of bioreactors involves an understanding of microbial community composition and dynamics, including the assembly processes through which they form. However, there remains a significant gap in our fundamental understanding of microbial community dynamics and microbial community assembly (MCA) processes, especially in engineered bioreactor settings. Here, we propose and employ a tool set that can be used by the research community, assess multiple bioreactor systems across a range of process types and ranges, and connect MCA patterns to relevant microbial groups in each bioreactor system. We applied multiple MCA assessment methods using available tools, layering on a trait-based approach, to seven experiments involving different engineered bioreactor systems. The calculated relative contributions of MCA processes varied by the method used, with null modeling approaches estimating a higher influence of stochastic MCA than neutral modeling. While most patterns of MCA were not discernible by general rules, anaerobic generalists assembled more deterministically than anaerobic specialists. Finally, statistical modeling of confidence levels suggests a minimum of 30–40 samples should be used for neutral modeling while a minimum 50–60 samples should be used for null modeling. Overall, we suggest caution when applying and interpreting the results of any one MCA assessment method.