Flux sampling and context-specific genome-scale metabolic models for biotechnological applications.

Abstract

Genome-scale metabolic models are used in fields ranging from metabolic engineering to drug discovery and microbiome design. Although these models are often used to predict putatively optimal states, some applications, including modeling human tissues for drug development and microbial communities for synthetic ecology, may require sampling the whole space of feasible fluxes to obtain distributions of biologically relevant states. Additionally, many applications involve using transcriptomic or proteomic data to predict fluxes for specific tissues, diseases, or patients. We revisit different methods used toward these goals and focus on their limitations and challenges, providing guidelines on how to avoid some of the shortcomings of existing approaches and highlighting conceptual barriers that will require new methodologies and offer opportunities for future development.