The big unsolved problem in microbiome discovery efforts is the millions of microbial gene sequences whose gene products/functions are unknown and uncharacterized. How can we harness this unknown microbiome biology for human health and agricultural benefits?
Second Genome is solving this problem with novel bioinformatics approaches, followed by laboratory assay screens to obtain prioritized listsof molecules with desired properties. For the known gene functions, we have created an integrated ‘omics knowledgebase of known bacterial functions with differential gene abundance in specific studies and compute the interaction linkage to observed host responses. From these linkages, interventions can be proposed. For the unknown microbial genes observed in de-novo assembly of metatranscriptomic NGS reads we measure their expression levels, and determine if their translated protein features (structural or physiochemical) are predictive of their interaction with host targets using Machine Learning approaches. Selected gene products are expressed and purified from bacterial gene sequences. These expressed proteins are screened through application-specific in vitro functional assays (e.g., barrier function, inflammation, tissue growth stimulation). This overall pipeline allows detection of known microbe-host interactions as well as novel interactions.