The human gut microbiota can become imbalanced, contributing to conditions such as obesity, diabetes, and inflammatory bowel disease. Now, researchers have developed a mathematical model of the gut microbiota that identifies dysbiotic states across multiple conditions, providing a non-invasive way to monitor gut health.

The findings, published in Science, suggest that monitoring shifts in microbial interaction networks could help to detect and track gut-related conditions, informing microbiota-based interventions.

Probiotics or fecal transplants aim to restore balance of the gut microbiota, but they are often unpredictable. This is because microbial interactions or how the microbiota shifts between “stable states” is poorly understood. What’s more, predicting outcomes with computational models remains challenging.

To address this challenge, Roberto Corral López at the University of Granada in Spain and his colleagues developed a mathematical model of the gut microbiota that tracks bacterial growth, nutrient use, and interactions.

Two states

Using realistic biological parameters, the model could reproduce key patterns seen in real gut communities, including the variety of species and fluctuations in their abundance. It also showed that different microbial compositions can perform the same metabolic functions, similar to observations in real human gut microbiotas. 

With this model, the researchers identified two distinct types of community states: a “healthy” state with many bacterial species rapidly turning over, and a “dysbiotic” state dominated by only a few species that persist at stable levels. These states mirror real patterns seen in conditions such as inflammatory bowel disease. 

In the dysbiotic state, diversity is lower, fewer metabolic pathways are active, and a few bacteria dominate the community. Healthy states, in contrast, have more overlapping metabolic functions, creating redundancy and stability, the researchers found.

Microbial balance

Further analyses revealed healthy microbiotas are dominated by competition, while disease-associated communities form consortia that exploit available nutrients. The researchers also developed a metric called the ecological network balance index (ENBI), which compares the balance of positive versus negative interactions in a community. ENBI increased in dysbiotic states, reflecting stronger positive interactions among microbes. 

ENBI also correlates with disease progression, suggesting it could serve as a non-invasive early warning tool for disease, although it cannot identify specific diseases on its own, the authors say.

Still, they add, “because the model captures ecological interactions in a broad and general way, it can be readily adaptable to other microbiomes, from other tissues (e.g., the vaginal or oral microbiomes) to plant or soil microbiomes.” The model can also be adapted to include treatments such as probiotics or fecal transplants, the authors say.