Prediction of Community Compositions in Stoichiometric Community Models

Microbial communities are ubiquitous in nature and play a major role in ecology, medicine and some industrial processes. One example involving a complex microbial community is anaerobic digestion for production of biogas. We use stoichiometric metabolic models to investigate communities that represent the last two stages of anaerobic digestion, namely acetogenesis and methanogenesis.
The organisms chosen to represent those stages are Desulfovibrio vulgaris, Methanococcus maripaludis, and Methanosarcina barkeri. We applied the previously described concept of balanced growth [1], which assumes an equal specific growth rate for all organisms in the community model. One of the open questions in modeling of microbial communities is which objective function might be suitable to describe the community. In order to predict community compositions we successively performed two optimizations: the primary objective is the maximization of the community growth rate. We found that, in realistic applications, this leads to a range of optimal community compositions. As a secondary optimization (applied to the optimal solutions of the first optimization), we therefore searched for the solution(s) where all organisms grow with their respective maximal biomass yield. This additional optimality criterion reduced the range of optimal solutions considerably (in many cases even to a single one) compared to only maximizing the community growth rate. Thus, this approach facilitates better predictions on community compositions.
With a two species model consisting of D. vulgaris and M. maripaludis we investigated the influence of different substrates on the predicted community composition and show that predictions from our optimization approach agree well with measured data. Additionally, we show that the ATP maintenance coefficient influences significantly the predicted community composition, especially for small growth rates.
Finally, we also analyzed a three-species community additionally including M. barkeri. In this community there is not only syntrophy but also competition between the methanogenic organisms, and the possibility to use different substrates. These dependency scenarios create different patterns in predicted community compositions.
In summary, our study presents some new techniques for stoichiometric modeling of microbial communities in general and provides valuable insights in interdependencies of bacterial species involved in the biogas process.

[1] KHANDELWAL RA, OLIVIER BG, RÖLING WFM, TEUSINK B AND BRUGGEMAN FJ. 2013. Community Flux Balance Analysis for Microbial Consortia at Balanced Growth. PLoS ONE 8.