
Computational methods for nutrition science, built in the open¶
NUTRIx is a network of early-career researchers working at the interface of nutrition, epidemiology, and data science. We meet, learn, and build a shared, openly licensed knowledge base of reproducible methods — so that good computational practice becomes part of everyday nutrition research.
Explore the knowledge base Come to a meeting Meet the network
Welcome¶
Nutrition science is being reshaped by digital tools and high-dimensional data — digitised dietary records, continuous glucose and activity monitoring, metabolic and inflammatory biomarkers, and multi-omics layers such as the microbiome and metabolome. Making sense of these data takes computational skills that are still rarely taught in nutrition programmes.
NUTRIx exists to close that gap. Our members first met at a 2024 spring school on machine learning in nutrition and recognised a shared set of methodological challenges. This site is where we work through them in public.
This is a living resource
The knowledge base grows at each network meeting. Pages are added, revised, and improved over time — and contributions from the wider community are welcome.
What you'll find here¶
-
Diet & Survival
Machine learning and survival analysis for dietary intake and time-to-event data: prediction versus causal inference, confounding and DAGs, substitution models, and N-of-1 designs.
-
Reproducible ML
Transparency and rigour in ML-driven nutrition research: avoiding data leakage, estimating stability and uncertainty, clear documentation, and FAIR, shareable workflows.
-
Multi-omics Integration
Integrative analysis of microbiome, metabolome, and other omics layers with dietary and phenotypic data: harmonisation, dimensionality reduction, and network-based interpretation.
Get involved¶
-
Join an in-person meeting
Each year the network hosts an in-person meeting on one of its three themes. The first two days are open to the wider nutrition and data science community — keynotes, expert talks, and hands-on workshops.
-
Contribute a page
Tutorials, worked examples, and methodological notes are all welcome. Start with the contribution guide.