Developed by research teams at the Medical University of Graz, Austria and the Institute for Systems Biology in Seattle, Metagenomic Estimation of Dietary Intake, or MEDI, looks at the DNA in stool samples to estimate what foods a person has consumed and in what quantity, with high levels of accuracy.
Sean Gibbons, associate professor at the Institute for Systems Biology and co-author of the research, says MEDI will allow researchers to track the diets of participants in any study that has collected fecal metagenomic data.
“This should expand the inclusion of diet as a variable in many studies, and it will allow people to go back to prior studies with available stool metagenomes to add an additional dietary data layer,” he told NutraIngredients.
The problem with traditional dietary assessments
Accurate tracking of dietary and nutrient intake is essential to understanding phenotypic heterogeneity in research i.e., the variation in observed characteristics within a study population who share the same disease or condition.
However, researchers assessing dietary intake often rely on food diaries or questionnaires. These methods depend on the individual to precisely recall their own food intake and may therefore be inaccurate.
Survey-free techniques such as analyzing human plasma or serum are often used in clinical settings, however, these methods have a ‘limited breadth of diet-relevant features’, according to the researchers.
“Therefore, there is a demand for approaches that can quantify dietary and nutrient intake patterns without the need for food frequency questionnaires or recall surveys,” they wrote in Nature Metabolism.
Creating and testing MEDI
MEDI uses metagenomic shotgun sequencing (MGS)—a common data type in human gut microbiome research that analyzes the DNA of all organisms in a sample.
Since MGS has not been used to detect food-derived DNA until now, the scientists began by obtaining genomic assemblies for as many single-organism foods as possible.
Using online databases, they catalogued 429 genomes and genomic assemblies, representing 561 food items and their associated strains. This food genome database exceeded commonly used databases by at least five-fold, leading the researchers to develop an efficient mapping strategy: MEDI.
MEDI’s performance was then assessed using simulated data, controlled-feeding studies with defined dietary interventions and observational data from infants and adults.
“MEDI predictions were validated in controlled feeding studies,” Professor Gibbons explained. “For example, we were able to distinguish between people who added an extra avocado to their meal and those who did not. We also saw that MEDI-inferred nutrient intake like total protein, total energy and total carbohydrates was significantly associated with questionnaire-based intake.”
“Finally, we showed how MEDI could be used to distinguish dietary differences between patients with metabolic syndrome and healthy controls. Specifically, MEDI indicated that healthy folks tended to eat more fruits and vegetables and folks with metabolic syndrome tended to eat more pork and chicken,” he added.
Implications for future research
The authors of the research say this newly discovered methodology could have far-reaching implications, from identifying dietary patterns linked to chronic conditions to predicting personalized short chain fatty acid production profiles in the gut.
“Paired microbiome-diet information could be leveraged to build microbial community-scale metabolic models, which are emerging tools that enable the design of personalized nutritional interventions that optimize specific functional outputs from the microbiota, like butyrate production,” Professor Gibbons said.
However, the model does have its limitations. For example, it is likely to be biased towards whole foods and therefore may not capture highly processed foods as accurately. It is also unable to account for different food preparations or the addition of ingredients such as added sugars or cooking oils.
Nevertheless, the authors say MEDI will be a valuable tool for nutritionists, epidemiologists, anthropologists, clinicians and microbiome researchers when applied to the ‘treasure trove’ of existing human stool data.
“By leveraging a common data type that is regularly collected to investigate the composition of the human gut microbiota, MEDI provides a value addition to any past, present or future metagenomic study for which rough estimates of dietary intake would prove useful,” they concluded.
Source: Nature Metabolism. doi: 10.1038/s42255-025-01220-1. “Metagenomic estimation of dietary intake from human stool.” Authors: C. Diener, et al.
