Authors: Francesco Asnicar, Paolo Manghi, Gloria Fackelmann, Gabriel Baldanzi, Elco Bakker, Liviana Ricci, Gianmarco Piccinno, Elisa Piperni, Katarina Mladenovic, Federica Amati, Alberto Arrè, Sajaysurya Ganesh, Francesca Giordano, Richard Davies, Jonathan Wolf, Kate M. Bermingham, Sarah E. Berry, Tim D. Spector, Nicola Segata
Journal: Nature
Year: 2025
Volume: 650
Pages: 450–458
DOI: 10.1038/s41586-025-09854-7
Study at a Glance
- Study type: Cross-sectional observational study with validation in two dietary intervention trials
- Number of studies: Five main cohorts plus two intervention trials
- Participants: Over 34,000 people in main analysis (UK and USA), with additional validation in 7,800 people from public datasets and 746 intervention participants
- Population studied: Generally healthy adults from the UK and USA
- Intervention or approach studied: Observational analysis of links between gut bacteria, diet quality, and health markers; validated in prebiotic blend and personalised diet interventions
- Comparison group: Various body mass index categories and health versus disease comparisons
- Follow-up duration: Cross-sectional for main cohorts; 18 weeks for METHOD trial; 12 weeks for BIOME trial
- Main outcome(s): Ranking of gut bacterial species by their association with health markers and diet quality
Why was this study done?
Poor diet has been linked to the to rising prevalence of disease in Western countries, and gut bacteria may also play an important role. The gut microbiome can be influenced by diet and has been associated with blood sugar levels, cholesterol, and body weight. However, exact relationships between diet, gut bacteria, and health markers remain unclear.
Asnicar and colleagues aimed to identify which gut bacterial species are most strongly linked to healthy diets and favourable health markers by analysing data from over 34,000 people across the UK and USA. They validated their findings in additional populations and two dietary intervention trials.
How was the study performed?
The authors collected stool samples, dietary information, and health measurements from 34,694 participants across five UK and USA cohorts. They used advanced metagenomic sequencing techniques to identify bacterial species in stool samples.
Dietary information came from written food questionnaires and mobile phone app food logs. To evaluate the quality of participants’ diets, they were scored using validated measures including the healthy eating index and plant-based diet indices. Health markers for each participant were also recorded, including body mass index, blood pressure, blood sugar, cholesterol, triglycerides, and inflammation markers.
All the data collected was used to identify the bacterial species consistently associated with more or less favourable health and diet scores.
What did the researchers find?
Strong, consistent links emerged between gut bacterial species and both diet quality and health markers across all five cohorts. Machine learning showed that an individual’s microbiome composition could be used to moderately predict blood sugar, cholesterol, triglycerides, inflammation markers, body mass index, and diet quality.
The authors identified and ranked 661 bacterial species appearing in at least 20% of participants. The most favourably ranked species were poorly known: 22 of the top 50 have never been grown in a laboratory. The 50 most unfavourably ranked species were generally well known, several previously linked to negative health effects.
Most highly ranked species belonged to Firmicutes bacteria. The 50 most favourable species comprised around 6% of total gut bacteria on average, while the 50 most unfavourable accounted for around 14%.
People with healthy weight carried more favourably ranked and fewer unfavourably ranked species than those with obesity. Across public datasets, healthy-weight individuals carried around five more favourable species than those with obesity.
Disease analyses showed people without inflammatory bowel disease, colorectal cancer, type 2 diabetes, impaired glucose tolerance, or cardiovascular disease carried more favourably ranked and fewer unfavourably ranked species than those with disease.
What did the authors conclude?
The findings provide strong evidence for associations between gut bacterial species, diet quality, and health markers, however the study design cannot conclude the microbiome causes better health outcomes. Large-scale intervention studies introducing specific bacterial species or dietary changes will be needed to establish causation.
Many bacterial species most strongly linked to better health are poorly understood. Future research should focus on these under-investigated species to understand their potential health role.
Key takeaways
- The study identified rankings for over 600 gut bacterial species based on associations with health markers and diet quality across more than 34,000 people.
- Many favourably linked bacterial species are currently unknown or uncharacterised, highlighting scientific knowledge gaps.
- People with healthy weight carried more favourably ranked and fewer unfavourably ranked species than those with obesity.
- Rankings were validated in over 7,800 additional people, showing consistent links with body mass index and disease status.
This Scientific Publication Summary is an objective summary of the published study for personal and educational use. It does not constitute medical advice, endorsement of the intervention, or a recommendation to alter clinical practice.