Variation in the gut microbiome has been linked to colorectal cancer (CRC), as well as to host genetics. However, we do not know whether genetic mutations in CRC tumors interact with the structure and composition of the microbial communities surrounding the tumors, and if so, whether changes in the microbiome can be used as a predictor for tumor mutational status. Here, we characterized the association between CRC tumor mutational landscape and its proximal microbial communities by performing whole exome sequencing and microbiome profiling in tumors and normal colorectal tissue samples from the same patient. We find a significant association between loss of function mutations in relevant tumor genes and pathways and shifts in the abundances of specific sets of bacterial taxa. In addition, by constructing a risk index classifier from these sets of microbes, we accurately predict the existence of loss of function mutations in cancer related genes and pathways, including MAPK and Wnt signaling, solely based on the composition of the microbiota. These results can serve as a starting point for understanding the interactions between host genetic alterations and proximal microbial communities in CRC, as well as for the development of individualized microbiota targeted therapies.