A new study in npj Science of Food introduces FoodAtlas, a large-scale, AI-powered knowledge graph that connects foods to the chemicals they contain—and to potential health and biological effects. Developed by IIFH-affiliated faculty member Dr. Ilias Tagkopoulos and his collaborators from AIFS, the platform brings together data from scientific literature and major biomedical databases into one integrated, searchable resource.
The team extracted tens of thousands of food–chemical relationships from published studies and linked them to disease associations, bioactivity data, and even flavor compounds. The result is a structured map that captures how 1,000+ foods relate to thousands of molecules and health-relevant pathways.
Beyond mapping ingredients, FoodAtlas can reveal clear dietary patterns, from omega-3–rich seafood linked to lower cardiovascular risk to high-fructose products associated with metabolic and nervous system impacts. The platform can even predict antioxidant activity and suggest healthier ingredient swaps, achieving an average 11.9% reduction in predicted disease risk in modeled meals.
By unifying food composition, chemical bioactivity, and disease associations in a transparent, evidence-based system, FoodAtlas offers a powerful new tool for advancing precision nutrition, food innovation, and data-driven dietary guidance.
