Multi-omics Integration for Substrate-Explicit Biogeochemical Modeling and More

Microorganisms control nutrient cycles by secreting extracellular enzymes to decompose complex organic matters into labile forms, which are subsequently consumed for microbial growth and transformed into other chemical compounds. Therefore, understanding the interplay among substrates, enzymes and microbes is the key for accurate prediction of microbially-mediated chemical transformations. However, the current practices of biogeochemical modeling oversimplify microbial processes, which significantly limits the predictive ability of the model. With the advancement of high-throughput omics profiling techniques, the growing recognition of the importance of representing microbial activities in biogeochemical models at sufficient resolution has led to the development of so-called microbial- and enzyme-explicit approaches. These models account for microbial physiology and interactions as well as enzymatic decomposition, but still provide a lumped description of nutrient pools. To fill this gap, we propose a new concept of biogeochemical modeling that enables incorporating high-resolution mass spectrometry data such as those from FTICR-MS. This method allows identification of biogeochemical transformation pathways for all compounds detected by mass spectrometry, and thus is termed substrate-explicit modeling. We demonstrate how thermodynamic theories can be used to convert chemical formulae of organic compounds (provided from mass spectrometry) into stoichiometric and kinetic equations of biochemical transformation. We further present the DOE's KBase (http://kbase.us/) modeling platform that enables incorporating FTICR-MS and other omics data in a more elegant way. Based on these developments, we propose a complete omics-based biogeochemical and ecosystem modeling pipeline from metagenomes to reactive transport models. We have applied our approach in comparative analysis of two riverbank sediments, one with and one without vegetation. Importantly, the proposed substrate-explicit modeling can be synergistically combined with enzyme- and microbial-explicit approaches to produce a platform capable of unprecedented predictions of biogeochemical and ecosystem dynamics.