Fig. 3

Comparison of mouse microbiome and metabolome data based on NJC19. (a,b) In the left panels, the abundances of the propionate (a) and acetate (b) producers in NJC19 were obtained from cecal 16S rRNA gene sequence data³⁵ in ref. ³⁷ Cecal metabolite concentrations in the right panels were obtained from Fig. 3c of the same study. Mouse group I consists of mice six weeks after 10-day cefoperazone treatment on mouse group II, cefoperazone-naive mice. (c) In the left panel, the abundances of the butyrate producers in NJC19 were obtained from fecal 16S rRNA gene sequence data³⁵ in ref. ⁴ Cecal butyrate concentrations in the right panel were obtained from Fig. 2c of the same study. Although we used the fecal 16S rRNA gene sequence data for the left panel due to the limited data availability, fecal and cecal microbial compositions were found to strongly correlate in other samples from that study, allowing us to use the fecal sequence data as a proxy for the cecal ones. All mice were initially germ-free in the study. Mouse groups I to V comprise mice transplanted with an obese twin’s microbiota (Ob; group I), mice co-housed with Ob and Ln mice (group II; see next for the definition of Ln mice), mice transplanted with the lean co-twin’s microbiota (Ln; group III), Ob mice co-housed with Ln and germ-free mice (group IV), and Ln mice co-housed with Ob and germ-free mice (group V). (d–f) In the left panels, the abundances of the succinate (d,e) and isovalerate (f) producers in NJC19 were obtained from cecal 16S rRNA gene copy levels in Fig. 2a of ref. ³⁸ We re-scaled succinate producers in (d) to the total 16S rRNA gene copy levels, because the corresponding succinate concentrations (d) were available per cecal dry weight (i.e., per cecal microbial load). Cecal and portal vein metabolite concentrations in the right panels were obtained from Fig. 2b and Supplementary Table 4 of the same study, respectively. Mouse groups ‘HF/HS’, ‘ZF/HS’, and ‘Chow’ represent gnotobiotic mice on a high-fat/high-sucrose diet, a zero-fat/high-sucrose diet, and a chow diet, respectively. In (a–d), each error bar represents standard deviation across replicates. Error bars are missing for (e,f), as well as for the left panel in (d), and units are missing for the right panels in (e,f), because of the information unavailability from the data sources.