A study led by researchers in the University of Virginia’s Department of Biomedical Engineering (a joint program of the School of Medicine and the School of Engineering and Applied Science) has systematically catalogued the bacterial composition of 352 over-the-counter probiotic products sold at three of the largest U.S. pharmacy chains—CVS, Walgreens, and Walmart. The analysis, published in Nature Microbiology in June 2026, found that the entire product set collectively represented just 36 unique bacterial species, with more than half of the products built around a single species and no product exceeding 17 strains.
What the Study Analyzed
The team, led by Jason Papin, PhD, together with Glynis Kolling, PhD, and Emma M. Glass, catalogued label and formulation data from commercially available probiotic supplements marketed for uses ranging from general gut health to vaginal health. Rather than running a lab assay for viable colony counts or contamination, the study’s primary contribution was computational: mapping which species and strain combinations the commercial market actually offers, and comparing that against what is mechanistically understood about each organism’s function in the human body.
Key Findings
- The 352 products collectively represented only 36 unique bacterial species—a narrow slice of the many thousands of species known to inhabit the human microbiome.
- More than half of the products contained a single probiotic species.
- The most complex formulation on shelf contained 17 strains; strain combinations were largely non-standardized across brands.
- Lactobacillus species were the most commonly represented genus overall.
The researchers frame this less as a mislabeling problem and more as an evidence gap: commercial formulation choices are not obviously anchored to the specific microbial functions a product claims to support. As the study’s authors note, dietary supplements in the United States are not regulated as strictly as drugs, so there is no requirement that a product’s composition be justified by mechanistic evidence before it reaches a shelf.
HaPaPro: A Computational Tool for Rational Formulation
To address that underlying evidence gap, the team built HaPaPro, a resource of more than 1,000 genome-scale metabolic models covering a wide range of bacterial species. The models let researchers predict, computationally, which organisms are metabolically suited to specific niches in the human body before they are formulated into a product. As a proof of concept, the team applied HaPaPro to identify candidate probiotics with the potential to support vaginal health, an application area the researchers describe as under-served by the current commercial catalog relative to consumer demand.
Why It Matters for Research Integrity and Evidence-Based Practice
This study sits at an intersection relevant to CASRAI’s audience even though its object of study is a consumer product rather than a peer-reviewed claim. It is an example of an academic team applying rigorous, reproducible computational methods to independently characterize a commercial product landscape that operates with comparatively little pre-market evidentiary obligation. That kind of independent, methodologically transparent audit of real-world product claims sits alongside standards bodies, registries, and open-evidence practices as part of the broader accountability infrastructure that research-integrity work depends on.
Funding and Publication
The study was published in Nature Microbiology (DOI: 10.1038/s41564-026-02380-w) on June 18, 2026, and was supported by the National Science Foundation (grant 1842490) and the National Institutes of Health (grants T32 GM-145443-1, R01-AI154242, and R01-AT010253).
