FooDis

FooDis extracts and suggests potential causal or therapeutic relationships between food entities and disease concepts from biomedical literature.

Key Features:

• Automated Relation Extraction: An information extraction pipeline employs natural language processing to mine textual data from biomedical abstracts for food–disease relations.
• High Precision Predictions: Achieves a 90% match rate for common food-disease pairs against the NutriChem database and a 93% match rate against the DietRx platform.
• Integration with Semantic Resources: Suggests candidate relations that can be compared to and used to complement semantic resources such as NutriChem and DietRx.
• Dynamic Discovery: Continuously analyzes new scientific publications to identify emerging food-disease relationships.

Scientific Applications:

• Nutrition Science: Supports systematic identification of diet–health associations relevant to nutritional research.
• Epidemiology: Provides candidate food-disease relations that can inform observational and hypothesis-driven epidemiological studies.
• Public Health: Supplies evidence for potential dietary risk or protective factors that may inform public health investigations and policy-relevant research.

Methodology:

Information extraction using state-of-the-art natural language processing to process and analyze biomedical abstracts and identify patterns and relationships between food entities and disease concepts.