Figure 2

Variables associated with beta-diversity in multiple datasets. (a) Variables/combinations significantly associated with changes in bacterial community structure in at least 2 datasets are plotted. Each bar represents the total number of datasets in which a variable (or variable combination) was tested and the red fraction of the bar denotes the number of datasets where the difference in beta-diversity was significant (PERMANOVA, P < 0.05). The first bar in each panel represents the variable “Subset: None; Autism Spectrum Disorder—FALSE over TRUE.” Remaining variables are listed in Supplementary file 1: Table S3. (b) Proportions of datasets with significant tests (non-aggregated data) for each variable combination are plotted. Columns are used to arrange subset values while rows segregate test variables. Results from aggregated data (genus and species levels) are presented in Supplementary file 2 (Figs. S4, S5).