Predictive multispecies constraint-based metabolic modeling: case studies and best practices

The last decade has seen a proliferation of studies and techniques for the modeling and analysis of metabolic interactions between distinct species, including microbes, plants, mice, and humans. Many studies in this area are explanatory, but a subset present predictive models aimed at either achieving specific biotechnological aims or furthering our understanding of the biological systems in question. Recent work suggests that the analysis frameworks and models that have been developed in the area of microbiome research are unreliable for accurately characterizing species-to-species interactions. In light of this development, the development of this subfield of metabolic modeling and its latest stumbling block are put into historical context. Several case studies of validated and predictive multispecies constraint-based metabolic models are then discussed. From these case studies, general principles are extracted that readers and practitioners in this area may find useful for evaluating and carrying out future multispecies metabolic modeling.