Designing synthetic bacterial consortia for landfill leachate treatment based on community matrices and regression tree analysis

The performance of microbial communities exploited by industry are largely optimised by manipulating process parameters, such as flow rates, growth conditions, and reactor parameters. Conversely, the composition of microorganisms used are often viewed as a black box. This is mostly due to the relatively high costs and technical expertise required to identify and quantify the microbial consortia, as well as limited tools to create functional assemblages. Unknown details about the interactions among species may impose a limit on how much microbial function can be optimised for industrial purposes. Here, a new workflow was developed for studying microbial consortia using high throughput, species and community specific measurements of growth rates and yields. Growth rate and yield among all single, pairwise, triple, quadruples, quintuple and sextuple combinations of six bacterial isolates on landfill leachate were evaluated. Additive, antagonistic (e.g. competitive) or synergistic (+/-) interactions can be inferred from the rate and yield data. We found that antagonistic interactions, which hinder growth and yield, were the dominant interaction type, with only a few synergistic interactions observed. Mixed effects models were used to investigate the relationship between interaction type and species richness (biodiversity). Community identity was found to be a more important factor in predicting yield determining interactions but not rate determining interactions. Species richness was a good predictor of rate determining interactions, with the most positive interactions happening at a low species richness. Regression tree analysis identified Lysinibacillus sp. as a keystone species, a genus previously associated with bioremediation. Its presence led to a drastic change in the function of the synthetic ecosystem, with both positive yield and rate determining interactions. We were able to infer interactions about specific pairs of species, and the competitive/synergistic tendencies of single species from only basic top-down growth measurements. In this way, we have demonstrated how factorial experiments using isolated microorganisms can be used to ultimately design synthetic consortia with desirable traits for industry.