Related Experiment Video
Updated: Sep 5, 2025
Evaluation of Colorectal Cancer Risk and Prevalence by Stool DNA Integrity Detection
Published on: June 8, 2020
Integrated Analysis of Colorectal Cancer Reveals Cross-Cohort Gut Microbial Signatures and Associated Serum
Renyuan Gao1, Chunyan Wu2, Yefei Zhu3
1Diagnostic and Treatment Center for Refractory Diseases of Abdomen Surgery, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China; Institute for Intestinal Diseases, Tongji University School of Medicine, Shanghai, China.
Background & Aims:
Studies have reported abnormal gut microbiota or circulating metabolome associated with colorectal cancer (CRC), but it remains a challenge to capture the CRC-relevant features consistent across geographic regions. This is particularly the problem for metabolic traits of CRC because the analyses generally use different platforms and laboratory methods, which poses a barrier to cross-dataset examination. In light of this, we sought to elucidate the microbial and metabolic signatures of CRC with broad population relevance.
Methods:
In this integrated metagenomic (healthy controls [HC], n = 91; colorectal adenoma [CRA], n = 63; CRC, n = 71) and metabolomic (HC, n = 34; CRA, n = 31; CRC, n = 35) analysis, CRC-associated features and microbe-metabolite correlations were first identified from a Shanghai cohort. A gut microbial panel was trained in the in-house cohort and cross-validated in 7 published metagenomic datasets of CRC. The in-house metabolic connections to the cross-cohort microbial signatures were used as evidence to infer serum metabolites with potentially external relevance. In addition, a combined microbe-metabolite panel was produced for diagnosing CRC or adenoma.
Results:
CRC-associated alterations were identified in the gut microbiome and serum metabolome. A composite microbe-metabolite diagnostic panel was developed and yielded an area under the curve of 0.912 for adenoma and 0.994 for CRC. We showed that many CRC-associated metabolites were linked to cross-cohort gut microbiome signatures of the disease, including CRC-enriched leucylalanine, serotonin, and imidazole propionate; and CRC-depleted perfluorooctane sulfonate, 2-linoleoylglycerol (18:2), and sphingadienine.
Conclusions:
We generated cross-cohort metagenomic signatures of CRC, some of which linked to in-house CRC-associated serum metabolites. The microbial and metabolic shifts may have wide population relevance.
