(815b) Biomog: A Computational Framework for the Modification Or De Novo Generation of Genome-Scale Biomass Compositions

The success of genome-scale modeling is contingent on a model’s ability to accurately predict growth behaviors following perturbation. To date, most research aimed at improving model predictions has focused upon the development of mechanisms to modify the cell’s metabolic and genetic networks by adding or removing reactions and refining gene to protein to reaction maps (GPRs) while little focus has been directed towards developing systematic methods of proposing, modifying and interrogating an organism’s biomass composition. To address this gap, the biomass modification and generation (BioMog) framework was created and used to generate biomass compositions de novo, as well as modify predefined biomasses for models of Escherichia coli (iJO1366) and of Shewanella oneidensis (iSO783) from high throughput growth phenotype and fitness datasets. This is accomplished by determining production blocked metabolites (i.e., metabolites that cannot be produced by a particular cellular mutant in a given minimal medium). By determining these blocked metabolites for each experiment performed, one can populate lists that support a particular metabolic candidate’s inclusion or exclusion from a cell’s biomass composition. Using this approach, BioMog’s de novo biomass proposals have been able to independently recapitulate up to seventy percent of the predefined biomass components and outperformed their predefined counterparts at qualitatively predicting growth phenotypes by up to five percent. Additionally, the BioMog procedure allows for the quantitative assessment of a particular metabolite’s essentiality to a cell. Such measures facilitate the determination of erroneous experiments, and inaccurate reaction network and GPR structures. Additionally, the BioMog framework contains an experiment generation component that allows for the directed design of informative experiments. Using such an approach, we correct experimental results relating to the essentiality of thyA gene to E. coli, as well as, perform phenotype experiments supporting the essentiality of protoheme.