Fig. 5: Effect of staple food type on the gut bifidobacterial community and its association with glycosidase-encoding genes.

a Composition of the bifidobacterial community in each staple food group. Ridge regression (with adjustment for age, sex, ethnicity, altitude, latitude, longitude, sampling month, and urban/rural/pastoral residence) was used to assess differences in the relative abundance of total Bifidobacterium between groups (indicated with p values on the top of the bar plot), as well as differences in the relative abundance of each Bifidobacterium species between groups (indicated with coefficient and p.adj values in the heatmap). In the bar plot, mean values are displayed. R, rice, n = 311; W, wheat, n = 354; RW, rice & wheat, n = 219. b Number of glycosidase types (left) and number of genes encoding glycosidases (right) of B. adolescentis and B. pseudocatenulatum isolates. p values from Mann–Whitney tests are shown. c Differential glycosidase-encoding genes between B. adolescentis and B. pseudocatenulatum isolates. p.adj values from Mann–Whitney tests are shown. Glycosidases that hydrolyze arabinoxylan were highlighted in red. d Illustration of arabinoxylan degradation. Endo-1,4-beta-xylanase from B. adolescentis cleaves arabinoxylan by internally hydrolyzing the 1,4-beta-D-xylosidic linkage between xylose residues in the xylan backbone, and arabinofuranosidases from B. adolescentis and B. pseudocatenulatum remove arabinose substituents from the xylan backbone.