Computer models can predict which probiotics will best colonize a person’s digestive tract, research shows, offering the potential to create bespoke combinations for a particular individual.
The findings, PLOS Biology, reveal how metabolic modeling can guide personalized, microbiome-mediated interventions to create designer probiotics.
Testing the strategy on results from two previous clinical trials, researchers found their model was at least 75% accurate in predicting which probiotic species remained present in each person’s gut.
The model was also able to predict how different probiotics impacted on the production of healthy short-chain fatty acids.
“Here, we bridge the gap between probiotic design and real-world application, using deep mechanistic insight to identify the right intervention for each individual,” said researcher Nick Quinn-Bohmann, PhD, from the Institute for Systems Biology in Seattle, U.S.
His Institute co-worker Sean Gibbons, PhD, added: “This work further demonstrates the potential of microbial community-scale metabolic models (MCMMs) as tools for designing and optimizing personalized probiotic and prebiotic interventions.”
Several factors affect the ability of a probiotic strain to grow and survive in the digestive tract, including having a suitable metabolic niche, interactions with other gut microbiota, and a person’s immune system.
Genome-scale metabolic models have been powerful for estimating microbial growth and metabolism, and recently this approach has been extended to diverse microbiota communities to produce MCMMs.
Quinn-Bohmann and team examined whether MCMMs could retrospectively predict how well probiotic species colonized participants in two placebo-controlled intervention trials.
The first was aimed at improving glucose control and tested a five-strain probiotic combined with the the addition of low-dose inulin, a prebiotic fiber known to support probiotic growth, among people with type 2 diabetes.
The second included an eight-strain probiotic combination aimed at treating recurrent Clostridioides difficile infections.
MCMMs largely agreed with measurements on how well the probiotics were adopted, with 75% to 80% accuracy. The models also captured treatment-driven shifts in predicted short-chain fatty acid production.
The model was tested further in 1786 people who shifted from a low- to high-fiber diet as part of a healthy eating and lifestyle intervention.
This revealed that it was also able to predict how increasing dietary fiber affected gut molecules and cardiometabolic markers of health.
“Taken together, these findings demonstrate the utility of MCMMs as a predictive framework for assessing prebiotic, probiotic, and dietary interventions at the individual and population levels,” the researchers summarized.
They added: “Ultimately, leveraging MCMMs in a clinical setting could enable precision microbiome therapeutics, optimizing probiotic, prebiotic, and dietary intake to more effectively treat a wide range of acute and chronic diseases.”
