Fig. 3: Identification of immunomodulatory species.
From: Gut microbiome-mediated metabolism effects on immunity in rural and urban African populations
a Whole blood from study participants was stimulated ex vivo with one fungal and seven bacterial stimuli and expression of five cytokines was measured. Heatmaps show significant Spearman correlations (P < 0.05) between samples in cytokine responses to all pairs of stimuli. Diagonal cells display significantly increased levels in urban (U) or rural (R) samples (Wilcoxon rank-sum two-sided test, P < 0.05). b Volcano plot of the species effects in cytokine responses, based on the log-linear model log yp,m,i = β0 + βddepi + βmmicm,i + βeeffm + εp,m,i, where εc,s,i is the noise, β0 the intercept, βd the subject sequencing depth, βm microbe abundance, and βe the microbes’ immunomodulatory (“Methods”). Immunomodulatory species are further split by negative (13, blue) and positive (21, red) effects subject to PBonferroni < 0.005 (two-tailed t-test) and prevalence >20%. c Distribution of p-values for each species resulting from the log-linear model used in TZ (top) and NL (bottom) cohorts. d Comparison with the same log-linear model applied to intersecting ex vivo blood cell stimulation from the NL cohort. Macrophages were stimulated with LPS, C. albicans, S. typhi, and M. tuberculosis, and IL-6 and TNF-α were measured. Point colors refer to directionality of immunomodulatory species as in (b). e Individual effects of immunomodulatory microbes. Coefficients are obtained from a linear model log(Cytokine expression) ~ age + gender + microbe fit to each individual cytokine and stimulus combination and only significant values are shown (two-tailed t-test, unadjusted P < 0.05; see Supplementary Data 9 for exact P-values). Source data are provided in the Source data file.
