Fig. 6: Propionate production gene cluster negatively associates with metabolic risk markers in prediabetes/diabetes cohort and reduced fasting glucose in mice fed with live A. rhamnosivorans.

Metagenomic analysis (a) shows that inositol utilisation gene cluster (cluster 1) did not form significant correlation with metabolic biomarkers while propionate production gene cluster (cluster 2) was significantly negatively associated with metabolic risk markers, especially fasting insulin and serum triglyceride (Spearman correlation); (b) shows significant differences in levels of fasting insulin (P = 2.6e-04) and serum triglyceride (P = 3.0e-05) in individuals with low versus high abundances of cluster 2 genes in a Swedish prediabetes cohort (Wilcox rank-sum test). −P < 0.1; *P < 0.05; +P < 0.01; #P < 0.001. The horizontal line in each box represents the median, the top and bottom of the box the 25th and 75th percentiles, and the whiskers 1.5 times the interquartile range. Fasting glucose levels in western diet fed C57BL/6J mice after 6 weeks oral administration of heat-killed or live A. rhamnosivorans without (c) or with myo-inositol (d). N = 10 mice per group. Data are mean ± SEM; statistical analysis was done by unpaired two-tailed Student’s t tests. Pearson correlation (e) between caecal propionate:butyrate ratio and fasting insulin in mice treated with heat-killed (blue circles) or live (orange circles) A. rhamnosivorans plus myo-inositol (P = 0.03).