Cross-cohort single-nucleotide-variant profiling of gut microbiota suggests a novel gut-health assessment approach

Abstract

ABSTRACT Adaptive evolutionary changes can precede the ecological changes in the gut microbial communities under constant host selection pressure, yet their association with host diseased status was underexplored. Explore shared disease-associated single-nucleotide variants (SNVs) in gut microbes spanning diverse human diseases. We performed a meta-analysis of 1,711 gut metagenomic samples from 16 case-control studies spanning 12 human diseases and included an additional 446-member cohort for validation. Overall, healthy individuals carried more mutated resident gut microbes, and more SNVs, mainly involving the short-chain fatty acids (SCFAs) producing bacteria. Furthermore, the widespread differences in base variant bias of gut microbes were observed between healthy and nonhealthy subjects suggesting divergent gut microbial evolutionary directions under different host medical conditions. We further found that nonsynonymous SNVs can lead to the loss of function of four SCFA-production genes in nonhealthy populations, among which two genes (i.e., ack and scpC ) from Faecalibacterium prausnitzii C and Bacteroides stercoris, respectively, were externally validated. Subsequently, we developed a novel gut microbiome health index based on the SNV rate of all mutated strains, classifying host health states with 74.23% accuracy, and validated with high accuracy (AUROC = 69.28%). Our study highlights the importance of employing the genetic variability in gut microbiome to characterize the gut microbial adaptation that can also predict human chronic diseases. IMPORTANCE Most studies focused much on the change in abundance and often failed to explain the microbiome variation related to disease conditions, Herein, we argue that microbial genetic changes can precede the ecological changes associated with the host physiological changes and, thus, would offer a new information layer from metagenomic data for predictive modeling of diseases. Interestingly, we preliminarily found a few genetic biomarkers on SCFA production can cover most chronic diseases involved in the meta-analysis. In the future, it is of both scientific and clinical significance to further explore the dynamic interactions between adaptive evolution and ecology of gut microbiota associated with host health status.