Development of a Novel Metagenomic Biomarker for Prediction of Upper Gastrointestinal Tract Involvement in Patients With Crohn's Disease.

Abstract

The human gut microbiota is an important component in the pathogenesis of Crohn’s disease (CD), promoting host–microbe imbalances and disturbing intestinal and immune homeostasis. We aimed to assess the potential clinical usefulness of the colonic tissue microbiome for obtaining biomarkers for upper gastrointestinal (UGI) tract involvement in CD. We analyzed colonic tissue samples from 26 CD patients (13 with and 13 without UGI involvement at diagnosis) from the Inflammatory Bowel Disease Multi-Omics Database. QIIME1, DiTaxa, linear discriminant analysis effect size (LEfSe), and PICRUSt2 methods were used to examine microbial dysbiosis. Linear support vector machine (SVM) and random forest classifier (RF) algorithms were used to identify the UGI tract involvement-associated biomarkers. There were no statistically significant differences in community richness, phylogenetic diversity, and phylogenetic distance between the two groups of CD patients. DiTaxa analysis predicted significant association of the species Ruminococcus torques with UGI involvement, which was confirmed by the LEfSe analysis (P = 0.025). For the feature ranking method in both linear SVM and RF models, the species R. torques and age at diagnosis contributed to the combined models. The L-methionine biosynthesis III (P = 0.038) and palmitate biosynthesis II (P = 0.050) were under-represented in CD with UGI involvement. These findings suggest that R. torques might serve as a novel potential biomarker for UGI involvement in CD and its correlations, in addition to a range of bacterial species. The mechanisms of interaction between hosts and R. torques should be further investigated.