211 Statistical analysis Descriptive statistics Descriptive data are presented as means ± standard deviation (SD), medians [interquartile range, IQR] or proportions n with corresponding percentages (%). Between-group comparisons were performed using Mann-Whitney U-tests, Pearson’s chi-squared tests or Fisher’s exact tests (if n observations were <10). Nominal P-values ≤ 0.05 were considered statistically significant. Mucosal gene expression analysis Sample gene expression dissimilarity was calculated using Aitchison’s distances. General linear models were used to assess the associations between mucosal gene expression and clinical phenotypes while controlling for potential confounders, which were determined from our previous study (medication included the use of aminosalicylates, thiopurines and steroids).103 In particular, to assess the effect of mucosal inflammation on gene expression, we re-coded the inflammation status in an ordinal fashion as 0, 1 or 2 to represent biopsies from non-IBD controls, biopsies from non-inflamed tissue of patients with IBD and biopsies from inflamed areas of patients with IBD, respectively. Intestinal inflammatory status was thus treated as a continuous variable to account for presence of residual inflammation in biopsies marked as being taken from non-inflamed areas in the intestines. A correction for multiple hypotheses testing was applied using an FDR threshold of 5%. Inflammation-associated genes were identified in three comparisons: (1) CD colonic inflamed tissue vs. CD colonic non-inflamed tissue vs. non-IBD colonic tissue, (2) CD ileocecal inflamed tissue vs. CD ileocecal non-inflamed tissue vs. non-IBD ileocecal tissue and (3) UC colonic inflamed tissue vs. UC colonic non-inflamed tissue vs. non-IBD colonic tissue: Gene ~ intercept + inflammation + age + sex + BMI + medication + batch Clinical phenotype–associated genes were identified using the following model: Gene ~ intercept + Montreal/anti-TNF therapy + age + sex + BMI + inflammation + tissue location + medication + batch Microbial characterization Microbial richness and evenness was determined by calculating the Shannon index representing alpha-diversity of the gut microbiota. Microbial dissimilarity of samples was determined by calculating Aitchison’s distances after clr transformation using the R package Compositions (v2.02). Analysis of paired samples from the same individuals was performed while comparing microbial features between inflammation status, disease location anddisease subtype usingpairedWilcoxon tests. Factors potentially influencing mucosal microbiota were determined using Hierarchical Allagainst-All significance testing (HAllA).104 Associations between microbial features and biopsy inflammatory status, IBD diagnosis, disease location (biopsy origin) and clinical phenotypes were 1. 2. Mucosal host-microbe interactions in IBD
RkJQdWJsaXNoZXIy MjY0ODMw