ME 2.0: A Computational Framework for Constructing and Simulating Genome-Scale Models of Metabolism and Expression

The widely used genome scale modeling of metabolism has recently been extended to integrate Metabolism with gene Expression to form ME models in T. maritima and E. coli. Although these ME-models have increased predictive capabilities and accuracy due to their incorporation of the biosynthetic costs for the machinery of life, they suffer from a significant increase in model size and complexity. For example, the E. coli ME model iOL1650-ME computational model increases the number of columns in in the resulting matrix by a factor of 30 relative to the iJO1366 M model it was built on. This results in models which are both difficult to compute with and challenging to understand conceptually. Additionally, to build these large computational models, both the T. maritima and E. coli ME models each had their own specific independent codebases, which makes it difficult to apply ME techniques to new models. To address these challenges, we have developed a new framework coded in Python and built on top of COBRApy for building and simulating ME Models. This framework, called ME 2.0, models the same biological processes as previous ME models, but strives to keep the resulting computation model simple to streamline computation and analysis. By adapting the reconstruction from iOL1650-ME to this new framework, an ME simulation can be performed in roughly 10 minutes, markedly faster than the six hours previously required. More importantly, however, this framework is organism agnostic, and will enable construction of ME-models of organisms beyond E. coli-K12 MG1655.