Fig. 2: Travelers’ gut microbiome compositions change throughout travel, with diarrhea-causing large-scale disruptions. | Nature Communications

Fig. 2: Travelers’ gut microbiome compositions change throughout travel, with diarrhea-causing large-scale disruptions.

From: Impact of international travel and diarrhea on gut microbiome and resistome dynamics

Fig. 2

a Taxonomic diversity (Shannon index) of subjects’ microbiomes throughout their length of stay, as days post-arrival. Points are individual samples, colored by sample type with dotted lines connecting samples from the same subject. Solid lines show the best fit line for different sample types: HT (blue), diarrheal (red), non-diarrheal preTD (green), and postTD samples (yellow) (n = 617, LMM, all P > 0.05), and the gray shading represents 95% confidence interval (CI). b Boxplots of the taxonomic diversity of subject-matched diarrheal and non-diarrheal samples collected within 2 weeks (n = 171, LMM, P[DiarrheaTD vs Non-diarrhea PostTD] = 3.4e-02). c Bray–Curtis dissimilarities between consecutive samples from each subject, plotted throughout their length of stay with dotted lines connecting samples from the same subject. Solid lines show the best fit for different traveler types (n = 291, LMM, P[Traveler type] <0.001) d Bray–Curtis dissimilarities between each subjects’ samples and their 1st-week baseline sample (n = 291, LMM, P[Time difference] <0.001) e Boxplots of Bray–Curtis dissimilarities calculated between 1st week baseline sample and a non-diarrheal sample collected 1 month after arrival in HT and TD subjects’. TD subjects were further sub-divided into Early TD (who experienced diarrhea <1 month of arrival) and Late TD (who experienced diarrhea >1 month after arrival) (n = 36, Wilcoxon test). Boxes in the boxplots show median and quartiles; error bars extend to the values within 1.5 interquartile range. f Microbial species significantly associated with diarrhea samples detected using MaAsLin227 where subjects were included as random effects and other metadata variables as fixed effects. The significant associations were corrected for multiple-hypothesis testing using Benjamini–Hochberg method with FDR < 0.25. The first column depicts the log normalized FDR value calculated by −sign(coeff)*log(qval), the second column shows the phylum and the barplot shows the effect size of each species. g The density ridgeline plot of significantly associated species in different sample types normalized by the median relative abundance of non-diarrheal HT samples. Left barplot, fraction of samples below detection limit. Underlying data are provided in the Source Data file.

Back to article page