Fig. 2: Meta-analysis of the gut microbiome from 5505 individuals (3288 females and 2216 males) across 18 datasets, revealing sex-associated microbial differences in healthy adults based on stool samples. | Nature Communications

Fig. 2: Meta-analysis of the gut microbiome from 5505 individuals (3288 females and 2216 males) across 18 datasets, revealing sex-associated microbial differences in healthy adults based on stool samples.

From: Meta-analysis of 22,710 human microbiome metagenomes defines an oral-to-gut microbial enrichment score and associations with host health and disease

Fig. 2: Meta-analysis of the gut microbiome from 5505 individuals (3288 females and 2216 males) across 18 datasets, revealing sex-associated microbial differences in healthy adults based on stool samples.The alternative text for this image may have been generated using AI.

a The 30 microbial species and genera with the highest standardized mean difference (SMD) meta-analysis coefficient (FDR = 0.01) between sexes. Effect sizes are calculated as SMDs from a linear model controlling for age, BMI, and sequencing depth, applied to centered log-ratio transformed species relative abundances. Yellow: significant effect size (FDR = 0.2). Light blue: non-significant effect size. Black diamonds: SMD between male and female. Yellow horizontal lines indicate the 95% confidence intervals of the effect size from the meta-analysis. b Mean relative abundance distribution of the 30 taxa in the 18 datasets, grouped by sex. The y-axis is in the log10 scale. Boxplots span the median, interquartile range and 1.5 times the interquartile range or the most extreme value. Values outside of this range are plotted as points. c Area Under the Receiver Operating Characteristic of a leave-one-dataset-out validation predicting sex using a Random Forest algorithm trained on various features: (i) age, BMI, sequencing depth of the samples; (ii) species relative-abundances, with and without (iii) age, BMI and sequencing depth as features; (iv) KEGG-level-collapsed UniRef90 gene families abundances with and without (v) age, BMI and sequencing depth as features. On top: the number of female and male participants in each study.

Back to article page