Identification of microbial markers across populations in early detection of colorectal cancer

Abstract Associations between gut microbiota and colorectal cancer (CRC) have been widely investigated. However, the replicable markers for early-stage adenoma diagnosis across multiple populations remain elusive. Here, we perform an integrated analysis on 1056 public fecal samples, to identify adenoma-associated microbial early detection of CRC. After adjusting potential confounders, Random Forest classifiers are constructed with 11 discriminate from control (area under ROC curve (AUC) = 0.80), 26 CRC (AUC 0.89), respectively. Moreover, validate in two independent cohorts achieving AUCs 0.78 0.84, Functional reveals that altered microbiome is characterized increased ADP- l -glycero-beta- d -manno-heptose biosynthesis elevated menaquinone-10 These findings validated a newly-collected cohort 43 samples using quantitative real-time PCR. This work proves validity adenoma-specific multi-populations, which would contribute treatment