Combined metabolomic and metagenomic analysis reveals inflammatory bowel disease diversity in pediatric and adult patients

Figure 1 Total headspace-gas chromatography-mass spectrometry data obtained for inflammatory bowel disease and control groups of pediatric patients. A and B: Principal component analysis and Orthogonal Partial Least Squares Discriminant Analysis data represent two independent groups of samples [inflammatory bowel disease (IBD) kids and control kids] according to the relative concentration of all detected volatile compounds; C and D: Principal component analysis and Orthogonal Partial Least Squares Discriminant Analysis data represent two independent groups of samples (IBD kids and control kids) according to the relative concentration of stable detected volatile compounds; E: Comparison of volatile organic compound composition for two independent groups of samples (IBD kids and control kids). Relative concentrations in the vapor phase were used. PCA: Principal component analysis; IBD: Inflammatory bowel disease; OPLS-DA: Orthogonal Partial Least Squares Discriminant Analysis.

Figure 2 Total headspace-gas chromatography-mass spectrometry data obtained for ulcerative colitis and Crohn’s disease and control groups of pediatric patients. A and B: Principal component analysis and Orthogonal Partial Least Squares Discriminant Analysis data represent three independent groups of samples [ulcerative colitis (UC) kids, Crohn’s disease (CD)-kids, and control kids] according to the relative concentration of all detected volatile compounds; C and D: Principal component analysis and Orthogonal Partial Least Squares Discriminant Analysis data represent three independent groups of samples (UC kids, CD-kids, and control kids) according to the relative concentration of stable detected volatile compounds; E: Comparison of volatile organic compound composition for three independent groups of samples (UC kids, CD-kids, and control kids). Relative concentrations in the vapor phase were used. UC: Ulcerative colitis; CD: Crohn’s disease; PCA: Principal component analysis; IBD: Inflammatory bowel disease; OPLS-DA: Orthogonal Partial Least Squares Discriminant Analysis.

Figure 3 Total headspace-gas chromatography-mass spectrometry data obtained for ulcerative colitis and Crohn’s disease and control groups of pediatric patients. A: Comparison of the volatile organic compound composition of stable detected metabolites for two independent groups of samples (inflammatory bowel disease kids and control kids). Relative concentrations in the vapor phase were used; B: The nonparametric Mann-Whitney test was used for the primary comparisons between groups, inflammatory bowel disease kids, and Control kids. Statistical significance was determined by a two-sided P-value of less than 0.05. False discovery rate correction was also applied; C: Comparison of the volatile organic compound composition of stable detected metabolites for three independent groups of samples (ulcerative colitis kids, Crohn’s disease kids, and control kids). Relative concentrations in the vapor phase were used; D: The nonparametric Mann-Whitney test was used for the primary comparisons between groups: Ulcerative colitis kids, Crohn’s disease kids, and control kids. Statistical significance was determined by a two-sided P-value of less than 0.05. False discovery rate correction was also applied. UC: Ulcerative colitis; CD: Crohn’s disease; IBD: Inflammatory bowel disease.

Figure 4 Total headspace-gas chromatography-mass spectrometry data obtained for the inflammatory bowel disease group of pediatric and adult patients. A and B: Principal component analysis and Orthogonal Partial Least Squares Discriminant Analysis data represent two independent groups of samples [inflammatory bowel disease (IBD)-kids, IBD-adults] according to the relative concentration of all detected volatile compounds; C: Comparison of volatile organic compound composition for two independent groups of samples (IBD-kids, IBD-adults). Relative concentrations in the vapor phase were used; D: The nonparametric Mann-Whitney test was used for the primary comparisons between groups. Statistical significance was determined by a two-sided P-value of less than 0.05. False discovery rate correction was also applied. UC: Ulcerative colitis; CD: Crohn’s disease; IBD: Inflammatory bowel disease.

Figure 5 Total headspace-gas chromatography-mass spectrometry data obtained for Crohn’s disease group of pediatric and adult patients. A: Principal component analysis and Orthogonal Partial Least Squares Discriminant Analysis data represent two independent groups of samples [Crohn’s disease (CD)-kids, CD-adults]. According to the relative concentration of all detected volatile compounds; B: The nonparametric Mann-Whitney test was used for the primary comparisons between groups. Statistical significance was determined by a two-sided P-value of less than 0.05. False discovery rate correction was also applied; C: Box plots the quantitative differences in the relative contents of metabolites detected in the analyzed groups (CD-kids, CD-adults); D: Risk coefficient calculation. UC: Ulcerative colitis; CD: Crohn’s disease; PCA: Principal component analysis; IBD: Inflammatory bowel disease; OPLS-DA: Orthogonal Partial Least Squares Discriminant Analysis.

Figure 6 Results of exploratory analysis of gut microbiota diversity in experimental groups (ulcerative colitis and Crohn’s disease, and control). A: The most abundant genera in the gut microbiota of patients are presented as a heatmap. Colors denote the relative abundance of species obtained by the emu pipeline after log transformation using pseudo counts. The figure shows the top 90 by relative abundance species. The rows correspond to the samples/patients; the phylum-level taxonomic ranks are denoted with a right color bar. Hierarchical clustering was performed using the Euclidean distance and complete linkage; B: Shannon index distribution of stool samples across experimental groups. Different colors denote different experimental groups. Horizontal brackets show the P-value obtained by comparing corresponding groups using the Wilcoxon rank-sum test; C: Non-metric multidimensional scaling biplot obtained using taxonomic profiles at the species level of patients’ and healthy controls’ stool samples. Different groups are shown by different color dots. Permutational multivariate analysis of variance analysis results are presented in the lower left corner; D: Bar plots denote the distribution of phylum abundance across experimental groups. The X-axis showed experimental groups, Y-axis showed relative abundance. Different colors show different phyla; E: Bar plots denoted the distribution of genus abundance across experimental groups. The X-axis showed experimental groups, Y-axis showed relative abundance. Different colors show different genera. F: Results of identifying differentially abundant species between ulcerative colitis and coronavirus disease 2019 patient groups using Maalsin2 with the following thresholds: Minimum abundance = 0.01, minimum prevalence = 0.25, maximum significance = 0.05. The Y-axis shows microbial species, the X-axis - the Maaslin2 regression coefficient. Color indicates taxonomic affiliation at the order level for each identified species. ^aP > 0.05, obtained by comparison corresponded groups using the Wilcoxon rank-sum test. ^bP > 0.01, obtained by comparison corresponded groups using the Wilcoxon rank-sum test. UC: Ulcerative colitis; CD: Crohn’s disease; PERMANOVA: Permutational Multivariate Analysis of Variance; MDS: Multidimensional scaling ordinate; NS: Not significant.

Figure 7 Gut microbiome analysis of pediatric and adult patients. А and В: Nonmetric multidimensional scaling based on Bray-Curtis dissimilarity; C: Boxplots representing Shannon’s alpha diversity index; D: Bar plots of bacterial family relative abundance. ^aP < 0.05, assessed by Kruskal-Wallis test. UC: Ulcerative colitis; CD: Crohn’s disease; NMDS2: Non-metric multidimensional scaling ordinate 2.

Figure 8 Difference between the taxonomic composition of the gut microbiome of pediatric and adult patients with inflammatory bowel disease. A: Results of identifying differentially abundant species between pediatric and adult inflammatory bowel disease patient groups using Maalsin2 with the following thresholds: Minimum abundance = 0.01, minimum prevalence = 0.25, maximum significance = 0.05; B: Results of identifying biomarker species of pediatric and adult inflammatory bowel disease patient groups using linear discriminant analysis effect size. The colors indicate biomarker species for the studied groups. IBD: Inflammatory bowel disease; LDA: Linear discriminant analysis.

Figure 9 Difference between taxonomic composition of gut microbiome of paediatric and adult patients with ulcerative colitis and Crohn’s disease. A: Results of identifying differentially abundant species between pediatric and adult ulcerative colitis patient’s groups using Maalsin2 with following thresholds minimum abundance = 0.01, minimum prevalence = 0.25, max significance = 0.05; B: Results of identifying differentially abundant species between pediatric and adult Crohn’s disease patient’s groups using Maalsin2 with following thresholds minimum abundance = 0.01, minimum prevalence = 0.25, maximum significance = 0.05; C: Results of identifying biomarker species of pediatric and adult controls, ulcerative colitis and Crohn’s disease patient’s groups using linear discriminant analysis effect size. The colors indicate biomarker species for the studied groups. UC: Ulcerative colitis; CD: Crohn’s disease; LDA: Linear discriminant analysis.

Figure 10  Spearman correlation between metabolite levels and relative abundance of microbial species. P < 0.05. A: Crohn’s disease; B: Ulcerative colitis; C: Shannon’s index diversity of metabolites in ulcerative colitis and Crohn’s disease groups. UC: Ulcerative colitis; CD: Crohn’s disease.