Pivotal Dominant Bacteria Ratio and Metabolites Related to Healthy Body Index Revealed by Intestinal Microbiome and Metabolomics

Abstract

Various body indexes, especially body fat percentage (BFP), are widely used as effective indicators to measure our health. BFP is used in medicine to assess obesity, which is a body fat mass disorder accompanied with changes of the gut microbiota. However, the relationship between BFP and the gut microbiota has not been studied so far. To address this problem, we examined how gut microbiota and metabolome associated with body indices in healthy people. Microbial and metabolomics data based on 16S rDNA sequencing and LC-MS were obtained from stool samples of 20 healthy adults. Bioinformatics analysis was performed to explore the correlations between the body indices and gut microbial characteristics. Significantly different microbes were further validated via qPCR. Differential characteristics were filtered by building machine learning models to predict body status. Our data showed that abundance of Prevotella and the Prevotella/Bacteroides (P/B) ratio in the gut were markedly higher in high-BFP individuals than in low-BFP individuals. Microbial and metabolomics data consistently suggested significant differences in fatty acid metabolism in stool samples from the two groups. The P/B ratio and fatty acids are discriminative for people with different index levels by cross validation tests with machine learning models. These results suggest using Prevotella and fecal fatty acids as predictors may offer an alternative method for evaluating health status or weight loss.

Supplementary information: The online version contains supplementary material available at 10.1007/s12088-021-00989-5.

Keywords: Body fat percentage; Gut microbiota; Machine learning; Metabolites; Metagenomics.

Fig. 1

Taxonomic differences in the gut microbiota between the high- and low-BFP groups. A Nonmetric multi-dimensional scaling (NMDS) plot based on the Bray–Curtis distance was used to explore the differences in the microbial structures between the two groups. The stress value was 0.17 (stress < 0.05: excellent representation; stress < 0.1: good representation, stress < 0.2: acceptable representation; stress > 0.3: unsatisfactory representation). B Significantly different gut microbiota identified in the top 15 taxa at the phylum, family or genus level via two-tailed Welch's t-test analysis (*, p < 0.05; **, p < 0.01; ***, p < 0.001). C The top 13 taxa with differentially abundant at the level of phylum, family and genus (corrected p-value < 0.05)

Fig. 2

Metabolic functional analysis of the gut microbiota and faecal metabolites. A KEGG pathway prediction for the gut microbiota via PICRSUT. The bile acid-related pathways are marked with a dashed frame. B Differential metabolites in the faeces between the high- and low-BFP groups with a VIP score > 1. The two bile acids are indicated with dashed frames

Fig. 3

Identification and validation of indicator organisms. A Correlation analysis between BFP-related characteristics of the gut microbiota and body indices. B Box-plot of the Firmicutes/Bacteroidetes and Prevotella/Bacteroides ratios in the two groups. C Box-plot of enterotypes analysis. D Validation of the BFP-related microbiota via qPCR