Ultra-processed foods, not just missing vegetables, may be the bigger threat to gut diversity
By Alena Pribyl, Lead Scientist at Microba
We have known for years that diet shapes the gut microbiome. What has been less clear is exactly which foods, nutrients and eating patterns matter most — and whether those effects hold up over time. A study published in Nature Medicine now offers some of the most detailed answers yet, drawing on dietary records and advanced stool DNA sequencing from over 10,000 adults in Israel’s Human Phenotype Project.
Participants logged their food intake via a smartphone app, while researchers profiled their gut microbiomes using shotgun metagenomics (a technique that reads all microbial DNA in a stool sample, identifying bacteria down to the species level). Machine learning models then tested how well diet could predict the makeup and function of each person’s microbial community. The results were striking: diet was significantly associated with over 92% of the 724 bacterial species tested and nearly 98% of microbial metabolic pathways analysed. Some food–microbe links were remarkably specific — coffee consumption, for instance, was strongly tied to higher levels of the butyrate-producing bacterium Lawsonibacter asaccharolyticus, yogurt to Streptococcus thermophilus, and milk to several Bifidobacterium species.
Perhaps the most clinically relevant finding was about broader eating patterns. The degree of food processing emerged as the single strongest dietary predictor of microbial diversity. Diets built around whole, minimally processed foods — whether plant-based or animal-based — were associated with greater microbial diversity, while higher ultra-processed food intake was linked to lower diversity. This challenges a simple “plant versus animal” framing and suggests that avoiding ultra-processed foods may be just as important for gut health as including specific food groups. Encouragingly, these diet–microbiome relationships were not short-lived: they persisted over four years of follow-up, and the models held up when validated in an independent Australian cohort and an Israeli dietary intervention trial.
The researchers also built an exploratory simulation tool that suggested personalised dietary tweaks — such as swapping white bread for fruit or nuts — predicted to shift the microbiome toward profiles associated with better cardiometabolic health markers. However, this remains a computer model that has not yet been tested in a real-world intervention. It is also worth noting that the study population was predominantly healthy Israeli adults aged 40–70, and dietary data was self-reported, which may limit how broadly the findings apply. Still, this is one of the most detailed large-scale investigations of everyday food choices and gut microbial composition to date, and it adds weight to a simple message worth sharing with patients: when it comes to the microbiome, how processed your diet is may matter as much as what’s on the plate.
