Fig. 11

Circos plot showing correlation analysis of immunological parameters with microbiota data. Sparse partial least squares discriminant analysis (sPLS-DA) was used to identify a first and second component based on TCR- and cytokine expression levels or absolute 16S rRNA counts and OTU relative abundance. The most discriminative features that were selected by the model from TCR- and cytokine expression data (grey) and OTU abundance (pink) are shown, where the outermost lines represent the feature abundance or expression level in samples from control tissue (green) and AP tissue (orange). At a correlation cutoff of 0.7, only negative correlations (blue lines) between the features were found. Bacterial taxa presumed to have functional riboflavin biosynthesis pathways are highlighted in blue font