Linking individual microbial and metabolite biomarkers of Colorectal Cancer using computational models of gut microbiota metabolism

Colorectal cancer (CRC) is the second most deadly cancer in the world, affecting almost 150,000 Americans each year and leading to over 50,000 deaths. Numerous microbial species and the metabolites they produce in the gut have been associated with the development and spread of CRC. However, it remains unknown to what extent the gut microbes, individually or collectively, contribute to the gut metabolome and affect CRC carcinogenesis. To better understand the functional role of individual microbial species in CRC carcinogenesis, we used a bottom-up systems biology approach based on GEnome-scale Models (GEMs) of metabolism to functionally profile the gut microbiota at species and molecular level resolution in CRC. To this end, we used publicly available fecal metagenomic data from 30 subjects with CRC and 30 non- CRC controls to construct GEMs of the gut microbiota metabolism at species-level resolution (spanning 113 microbial species). By computationally simulating these models, we could infer the metabolic activity of each microbial species in the gut, which allowed us to trace back individual microbial species producing specific secreted metabolites. This analysis identified 338 metabolites with differential production levels by the gut microbiota as well as over 400 linkages between specific microbial species and metabolites that were significantly different between CRC subjects and non-CRC controls (Wilcoxon, adjusted p < 0.01). Several of these identified metabolites and species have been previously implicated in CRC, examples of which include chicory inulin, deoxycytidine, and acetate that are produced by Faecallibacterium prausnitzi, Eubacterium eligens, and Bacteroides vulgatus, respectively, amongst other species, according to our models. Overall, our study provides a roadmap for mechanistically linking microbial and metabolite biomarkers of CRC.