Fig. 4: Metabolomics-inferred BMI reflected gut microbiome profiles better than BMI.

a, Overview of study cohorts and the gut microbiome-based obesity classifier generation. CV, cross-validation; RF, random forest. b, Difference in gut microbiome α-diversity between the matched and mismatched groups within the normal or obese BMI class. Each comparison value indicates adjusted P value, calculated from OLS linear regression with BMI, sex, age and ancestry principal components as covariates while adjusting multiple testing with the Benjamini–Hochberg method across the 24 (2 BMI classes × 4 omics categories × 3 metrics) regressions. Data: the standard box plot (Methods); n = 240 (Normal) or 260 (Obese) participants (see Supplementary Data 6 for each sample size). a.u., arbitrary units. c,e, ROC curve of the gut microbiome-based model classifying participants to the normal versus obese class in the Arivale (c) or TwinsUK (e) cohort. Each ROC curve was generated from the overall participants: n = 500 (c, BMI class), 427 (c, MetBMI class), 209 (e, BMI class) or 145 (e, MetBMI class) participants. The dashed line indicates a random classification line. P: P value of two-sided unpaired DeLong’s test. d,f, Comparison of model performance between the BMI and MetBMI classifiers in the Arivale (d) or TwinsUK (f) cohort. Out-of-sample metric value was calculated from each corresponding hold-out testing set. Data: mean with 95% confidence interval, n = 5 models. P: P value of two-sided Welch’s t-test. All exact values of test summaries are found in Supplementary Data 6 and 10.