Document resource
Background Differential diagnosis between Crohn’s disease (CD) and ulcerative colitis (UC) still remains clinically challenging. Gut microbiome dysbiosis is implicated in disease pathogenesis but precise disease-specific microbial signatures capable of non-invasive differentiation remain poorly defined. We hypothesized that high-resolution 16S rRNA profiling combined with machine learning could identify distinctive microbial signatures for accurate CD-UC discrimination and resolve the controversial richness-diversity relationship in IBD.Methods We prospectively enrolled 42 IBD patients (23 UC, 41.3±14.2 y.o.; 19 CD, 37.2±11.5 y.o.) and 8 age-matched healthy controls (47.1±19.8 y.o.). Fecal samples underwent Oxford Nanopore Technologies full-length 16S rRNA gene sequencing (1400-1600 bp, Q-score >10). Bioinformatic pipeline included Porechop v0.2.4 adapter trimming, Chopper v0.8.0 quality filtering, and EMU abundance-based taxonomic classification. Statistical analyses included: alpha-diversity (Shannon index, Shapiro-Wilk, Mann-Whitney U tests); beta-diversity (PERMANOVA, five distance metrics, FDR correction); species-level differential abundance testing (p<0.05); Random Forest machine learning classification with Mean Decrease Accuracy validation.Results A novel microbiome richness-diversity paradox was identified: CD exhibited the highest species richness (92 species) versus UC (78 species) and healthy controls (37 species), and showed the lowest Shannon diversity (CD 3.12 vs UC 3.30 vs controls 3.66; CD vs controls p=0.036, Mann-Whitney U). This paradox ( IDDF2026-ABS-0340 Figure 1. The Microbiome Richness-Diversity Paradox Distinguishes Crohn’s Disease from Ulcerative Colitis. A. Species richness comparison across study groups. B. Shannon diversity index comparison. C. Explanation of the richness-diversity) suggests pathobiont invasion with commensal collapse, explaining contradictory literature findings. Beta-diversity confirmed robust CD-UC separation (PERMANOVA Canberra R=3.09%, p=0.021, FDR=0.063). Disease-specific taxonomic signatures emerged with diagnostic potential: dramatic Enterococcus faecium enrichment (269-fold versus UC: 2.69% vs 0.01%, p<0.001) and Klebsiella pneumoniae expansion (36-fold: 0.72% vs 0.02%, p<0.001) were demonstrated in CD, contradicting previous reports of E. faecium as UC-predominant pathobiont. UC also exhibited elevated Clostridioides difficile colonization prevalence (5.6%). Both CD and UC showed expected depletion of Faecalibacterium prausnitzii (7.9% residual) and Akkermansia muciniphila (2.4% residual). Random Forest classification identified Bacteroides stercoris in UC, Klebsiella pneumoniae and Oscillibacter valericigenes in CD as top discriminatory features, validating clinical applicability.Conclusions This study establishes two paradigm-shifting findings: 1) the richness-diversity paradox mechanistically explains IBD dysbiosis as pathobiont invasion rather than simple diversity loss; 2) E. faecium is suggested as a CD-specific (not UC-specific) pathobiont with diagnostic utility. These microbiome signatures enable non-invasive CD-UC differentiation, potentially reducing endoscopy dependence and guiding precision therapeutics.Abstract IDDF2026-ABS-0340 Figure 1