New research uncovers gut microbial markers common across multiple diseases, paving the way for more accurate disease prediction and personalized treatments.

Study: A population-scale analysis of 36 gut microbiome studies reveals universal species signatures for common diseases. Image Credit: Kateryna Kon / Shutterstock

A study published in the Nature Portfolio journal Biofilms and Microbiomes finds significant variations in gut microbial compositions across common human diseases in the Chinese population.

Background

The gut microbiota plays a vital role in human health and disease. Because of its remarkable plasticity, the overall composition of gut microbiota often remains stable even after acute changes in the physiological system. However, chronic and prolonged exposure to stressors can lead to gut microbiota dysbiosis, an imbalance in gut microbial composition that favors the selection of harmful pathogenic microorganisms over beneficial microorganisms.

Gut microbiota dysbiosis has been observed in several diseases, including autoimmune, cardiometabolic, infectious, psychiatric, and cancer. However, the precise cause of microbial dysbiosis in different disease conditions is not fully known due to several challenges, including the lack of unified reference databases, the low accuracy of bacterial species annotation and quantification in high-throughput sequencing datasets, and the highly variable experimental and analytical methods used in studies.

In this study, scientists re-analyzed publicly available fecal metagenomes from 36 case-control studies to investigate the differences in gut microbial composition and diversity between cases and controls in each study.

Study Design

The study analyzed raw sequencing data from publicly available metagenomic datasets, which included 6,314 human fecal samples from 36 case-control studies of the Chinese population. These datasets comprised 3,728 patients with 28 diseases or unhealthy conditions and 2,586 healthy individuals.

A comparative analysis was conducted to explore the differences in overall microbial structure between patients (cases) and healthy individuals (controls). The study conducted an integrated meta-analysis to identify universal microbial signatures across various diseases. Based on the abundance of microbial signatures, advanced machine-learning classifiers were established to investigate the potential of generic gut microbial features in predicting disease states.

Important Observations

The analysis of the gut microbial composition of all samples identified Bacteroidetes and Firmicutes as the most abundant phyla, followed by Proteobacteria and Actinobacteria. A highly diverse proportion of these phyla was observed in the analyzed studies, which may be attributed to differences in geographical locations or experimental methods.

At the genus level, Phocaeicola, Bacteroides, Prevotella, Faecalibacterium, Alistipes, and Roseburia were identified as the dominant genera, showing a similar distribution pattern across all studies.

Several members of Phocaeicola, Bacteroides, Prevotella copri, and Faecalibacterium prausnitzii were the most dominant species in all analyzed samples.

The comparative analysis of case-control studies revealed that distinct gut microbial structures were associated with most diseases, with many showing reduced microbial richness and diversity, while some showed increased richness and diversity compared to controls.

Further analysis revealed that the disease state significantly affects the overall gut microbial composition. Diseases like Crohn’s disease, polycystic ovary syndrome, atrial fibrillation, Grave’s disease, systemic lupus erythematosus, liver cirrhosis, pulmonary tuberculosis, or coronavirus disease 2019 (COVID-19) exhibited the most significant changes in microbial composition.

Gut Microbial Signatures

The integrated meta-analysis identified 277 microbial species that differed in relative abundance between cases and controls. A total of 194 species showed higher abundance in healthy individuals compared to patients. Similarly, 83 species were more enriched in patients compared to healthy individuals.

A significant difference in taxonomic distribution at the phylum and genus levels was observed between disease-enriched and control-enriched microbial species. Notably, patients exhibited a significant enrichment of opportunistic pathogens and a depletion of beneficial microbes.

Microbial species that showed reduced abundance in patients are the most important producers of short-chain fatty acids (SCFAs) in the human gut. This indicates that a reduced SCFA biosynthesis ability is a shared feature across many human diseases.

At the genus level, the analysis identified 107 genera that differed in relative abundance between cases and controls. Of these genera, 73 were more abundant in controls and 34 in patients. Similar to the species-level analysis, a notable reduction in beneficial SCFA-producing genera and an increase in opportunistic genera were observed in patients.

Gut Microbial Signatures for Disease Prediction

A machine-learning algorithm-based analysis showed that these universal gut microbial signatures can accurately distinguish between patients and healthy individuals with a high degree of accuracy. The model could also differentiate between high-risk patients and healthy individuals.

For further validation of the prediction model, the study analyzed fecal metagenomes from three independent public cohorts, including bipolar depression, colorectal cancer, and end-stage renal disease. The findings revealed that the prediction model was particularly effective in distinguishing high-risk diseases, though its ability to detect psychiatric diseases was comparatively lower.

Study Significance

The study identifies universal gut microbial signatures for common human diseases in the Chinese population. Overall gut microbial structure was strongly associated with common human diseases. These findings suggest that generalized disease-associated gut microbial signatures can accurately classify multiple disease states from a healthy state, opening possibilities for future diagnostic tools and personalized interventions.