3D Reality Check: Developing Structural Support for Interpreting Microbial "Omics" Data and Predicting Microbial Function

The microbial activities on which humanity and ecosystems depend are determined by microbial community membership and by the resources and conditions in the microbes’ local environments. In order to effectively develop insight into controls over microbial community contributions to ecosystem function, a pervasive disconnect must be addressed: The length of a typical bacterium or archaeon is approximately one to five micrometers, yet environmental sampling meant to provide contextual insight into environmental conditions surrounding, and resources available to, soil microbes is often (by necessity) carried out at much larger scales. “Metadata” (pH, nutrient concentrations, availability of organic matter etc.) meant to capture conditions and resources available to microbes are measured from homogenized, coffee-mug-size (or even larger) cores of soil. Yet microscale heterogeneity of soil conditions associated, for example, with the presence of plant roots, fungal hyphae, and soil fauna have long been recognized to spur booms and busts of microbial taxa and biogeochemical function. The scale discrepancy between environmental data gleaned from large homogenized samples and the “omics” (transcriptomic, metagenomic, proteomic) signals composed of information from millions of individual microbes makes it impossible to infer the micron-scale environmental conditions that the microbes are actually reacting to, and influencing.
This project aims to develop a computational approach in which a genome-scale model of microbial function (methane production by Methanosarcina barkeri) can be embedded within a high resolution 3D, pore-scale framework to predict methanogenesis in heterogeneous microenvironments. M. barkeri’s genome has been fully sequenced and annotated by JGI, and a detailed genome-scale model has been developed for its metabolism. M. barkeri has the unusual capacity to use the solute acetate or the dissolved gases hydrogen plus carbon dioxide (H2+CO2) to produce methane. In the test case proposed here, the genome-scale model for M. barkeri will be embedded in the 3D pore-scale framework and used to predict how heterogeneous substrate availability at microscales affects the expression of two methanogenic pathways in unsaturated, anaerobic, artificial soil. Those predictions will then be compared with JGI transcriptomic data derived from microcosms populated by M. barkeri in the laboratory and provided with acetate and/or H2+CO2 under the modeled conditions. Simulation of 3D pore domains with the resolution necessary to capture environmental variation at the microbe level requires millions of discretization points, and therefore such simulations are extremely computationally intensive and require the high-performance computing capabilities of EMSL’s Molecular Sciences Computing Facility.
Success of this test case will encourage more broad application of the combined 3D pore-scale modeling framework with embedded biogeochemistry. The modeling approach can accept prescribed 3D structural input from e.g. tomographic or other soil imaging scans. The long-term goal of coupled “omics”-3D pore scale modeling will be to address two fundamental questions in microbial ecology:
• How does 3D microenvironmental structure affect, and how is it affected by, microbial community structural diversity and expression of microbial function?
• How does environmental microheterogeneity affect resulting process rates measured at larger scales, and our ability to predict them, e.g. in bioremediation, ecosystem function, food, or fuel production?