Fig. 1: The composition of the gut microbiome significantly differs between PD and controls (CTR). | Nature Communications

Fig. 1: The composition of the gut microbiome significantly differs between PD and controls (CTR).

From: Machine learning-based meta-analysis reveals gut microbiome alterations associated with Parkinson’s disease

Fig. 1

Distance-based redundancy analysis (dbRDA) was performed on Bray-Curtis dissimilarities calculated for the 16S (a, c) and SMG data (b, d) with the percentage of variance explained annotated along the axes. Raw distances (a, b) and distances conditioned by the study of origin to remove study-specific (batch) effects are shown (c, d). Boxplots depict the sample distribution along the two components and are coloured to indicate the different datasets used (e, f) or the disease status (g, h). Boxes delineate the interquartile range (IQR), with the middle thick segment indicating the sample median. Whiskers extend to the most extreme values within 1.5 × IQR. Data beyond this range are not reported. The sample size (n) for each boxplot corresponds to the number of samples reported in Table 1. Results of permutational analysis of variance (PERMANOVA) performed for each data type are shown within (a) and (b). The significance of the clustering was calculated for both the study of origin (16S: df = 16; F = 49; p-value = 0.0005; SMG: df = 6; F = 18.2; p-value = 0.0005) and the disease status (16S: df = 1; F = 31.4; p-value = 0.0005; SMG: df = 1; F = 7.9; p-value = 0.0005).

Back to article page