Project Summary/Abstract The Microbiota-Diet-Host interactions are critical to the balance of health and disease. Understanding this axis will be important to the development of nutritional precision medicine. Although there are some important discoveries where microbes make molecules that dramatically affect health, most of the molecules produced by microbiota still remain structurally uncharacterized. In addition, even when a molecule is known and the activity is known, this knowledge tends to be buried in papers and there is not a systematic organization of this knowledge that can be readily leveraged by data scientists. This prevents a deep mechanistic understanding of the microbiome. The goal of PAR-21-253 to which this application is applying to and the accompanying RFA (RFA-DK-21-014) is to build a microbial metabolite knowledgebase that can be used by the larger scientific community. In order to build the knowledgebase, the R01’s funded under PAR-21- 253 are discovery grants that will provide the knowledge of new microbial metabolites and their bioactivities. This proposal will 1) obtain MS/MS signatures for up to 5,000,000 molecules synthesized, using combinatorial/diversity driven synthesis, using diet and host precursors that are accessible to human microbiota. The synthesis is biased towards common bio-transformations. This will be the largest metabolomics reference data set ever assembled, even when all public and non-public libraries are combined; 2) we will use our big data mining strategies, especially our mass spectrometry search tools, to not only discover what molecules are made by microbes, but also understand phenotype, disease, organ/tissue/biofluid, food associations; 3) the newly discovered microbial molecules that are up or down regulated in inflammatory bowel disease (and other gut metabolic disorders) will be further screened in cell-based assays and colitis animal studies to understand the biological effects of the newly discovered microbial metabolites; 4) we will transfer all the knowledge obtained by this R01 to the knowledgebase that will be created by the accompanying RFA. The PIs and Co-Is are well suited to help build the proposed knowledge base, due to their expertise in microbial metabolomics (Dorrestein), synthesis (Dorrestein, Siegel), microbiome (Dorrestein, Knight, Raffatellu), data science (Dorrestein, Knight) and having a track record of doing very large projects and share newly discovered knowledge publicly for the benefit of the community – generally made available years before any publication (Dorrestein, Knight).

Project Narrative In the human body, for 100 human cells that are present, there are 130 microbial cells. Those microbes often define the balance of human health and disease, yet the molecules by which they promote this balance are unknown. We have designed a reverse metabolomics and big data mining strategy to discover microbial metabolites with emphasis on microbial metabolites associated with inflammatory bowel disease and other gut associated metabolic disorders.