Abstract
Colorectal cancer (CRC) is influenced by both enteroviruses and bacteria, yet current microbial typing schemes rely primarily on bacterial data. To construct a more comprehensive microbial typing scheme for CRC, this study integrated enteroviral and bacterial profiling data from fecal samples of 414 healthy controls (NCs), 151 advanced adenoma (AAs) patients, and 255 CRC patients using Illumina sequencing. Metabolites and trace elements were analyzed via liquid chromatography and ICP-MS, respectively. Microbial subtyping was performed with ConsensusClusterPlus based on combined viral and bacterial sequencing data, leading to the identification of two initial viral subtypes (V1, n=309; V2, n=511). The V2 group was further split into two bacterial subtypes (V2B1, V2B2), yielding three distinct microbial subtypes. Disease ratios (CRCs&AAs/NCs) were 1.06 (V1), 0.67 (V2B1), and 1.29 (V2B2). V1 showed increased Streptococcus agalactiae, Peduvirus, and Imidazopyrimidines; V2B1 had higher CAG_127sp900553925, nickel (Ni), and benzene derivatives; V2B2 exhibited elevated Citrobacter farmeri, Svunavirus, arsenic, and organic sulfonic acids. Gut disease prediction model was more accurate after virus typing (86.54% of accuracy in V2B2 subtype; 73.33% of accuracy without typing). These results suggest that integrating enteroviral and bacterial subtypes offers a more precise framework for CRC identification than bacterial-based typing alone.