Additional file 3 of Effects of caloric restriction on the gut microbiome are linked with immune senescence

Additional file 2. The three supplementary Excel files, for colon (colon_cca_xlsx), spleen (spleen_cca_xlsx), and liver(liver_cca_xlsx), comprise each four spreadsheets: The sheets “corr.matrix” and “corr.fdr” contain estimated coefficients of the latent cross-correlation matrix (as shown in Fig. 6 and Supplementary Figure 5 heatmaps) and corresponding FDR-adjusted P-values, respectively. The sheet “asv.cca” contains for all ASVs (with the same order as in the correlation matrix) FDR-adjusted P-values and log-fold change values (columns C, D) from DESeq2 differential abundance testing (as shown in heatmap Fig. 1 C), the corresponding taxonomy table including seven taxonomic ranks (columns E-K), taxa labels used for heatmaps (column L) and two columns indicating the top (upper and lower) ASVs that were highlighted in correlation heatmap and circular chord diagrams (column M) and that were selected by sparse canonical correlation analysis using the Bayesian information criterion (column N), with sparse coefficients’ weights provided in column O. The sheet “imm.cca” contains for all immunological parameters (with the same order as in the correlation matrix) that were assessed in a given organ, FDR-adjusted P-values from GLMM and non-parametric (Wilcoxon-Mann-Whitney) differential abundance testing (columns C, D) between AdLib and CalRes groups. Columns E indicates type of parameter, i.e., whether absolute cell counts, or cell frequencies were determined. Column F denotes which cell population served as input for cluster-based analyses or was manually gated. Columns G – I provide cell subset IDs including prepended cluster numbers, manually annotated population names, and differentiation state. Columns J – L contain labels as displayed in correlation heatmap and circular chord diagrams for subset name, differentiation/antigen-experience, and lineage membership, respectively. Columns M – O indicate the top left and right immune parameters in the correlation heatmap, those selected by sparse canonical correlation analysis using the Bayesian information criterion, and sparse coefficients’ weights, respectively.