Genomic Science Program
U.S. Department of Energy | Office of Science | Biological and Environmental Research Program

Multiomics-Driven Microbial Model Optimization

Authors:

Janis Shin1,2* (jshin1@uw.edu), Herbert M. Sauro2, James Carothers1

Institutions:

1Molecular Engineering and Sciences Institute, University of Washington–Seattle; 2Department of Bioengineering, University of Washington–Seattle; 3Earth and Biological Science Directorate, Pacific Northwest National Laboratory

URLs:

Goals

The goal is to create genome-scale models of endogenous metabolic pathways and develop metabolic sensitivity maps to identify reactions that dominate the control of flux. Specifically, researchers aim to gain insight into the regulatory mechanisms within pathways and predict outcomes of metabolic interventions at the genome-scale.

Abstract

Biomanufacturing poses a sustainable alternative to producing traditionally petrochemical-derived commodity chemicals. For these biomanufacturing efforts to be commercially viable, the design and engineering process must be targeted and quick. The advent of CRISPRa/i enables the targeted activation and repression of specific genes in organisms. Choosing which genes to target with CRISPRa/i for maximal yield through the metabolic pathway requires identifying which enzymes exert the most influence on the flux through the pathway (Kacser and Burns 1995). Bayesian Metabolic Control Analysis (BMCA) has been developed as a way to integrate omics data and genome-scale models to predict the flux control coefficients of a metabolic pathway (St. John et al. 2019). Yet, BMCA’s predictive accuracy has yet to be quantified. Here, the results show that BMCA reliably predicts elasticity values in the absence of allosteric regulation and that the most informative type of data for BMCA is fluxomics, followed by enzyme concentrations, external metabolite concentrations, and internal metabolic concentrations. These results also demonstrate the fidelity and limitations of BMCA’s predictions given the strength of CRISPRa/i perturbations in the dataset and thereby establish guidelines for maximizing the predictive power of the BMCA method. This method was successfully applied to estimate sensitivities in random model topologies. Researchers anticipate that the insights drawn from this benchmarking study can be extended to other metabolic pathways, such as Pseudomonas putida.

References

Kacser, H., and J. A. Burns. 1995. “The Control of Flux,” Biochemical Society Transactions 23(2), 341–66. DOI:10.1042/bst0230341.

St. John, P. C., et al. 2019. Bayesian Inference of Metabolic Kinetics from Genome-Scale Multiomics Data,” PLoS Computational Biology 15(11), e1007424. DOI:10.1371/journal.pcbi.1007424.

Funding Information

This research was supported by the U.S. DOE Bioenergy Technologies Office (BETO), grant no. DE-EE0008927 and DOE Office of Science, BER program, under Funding Opportunity Announcement DE-FOA-0002600.