NutriChem

NutriChem maps plant-based foods to their phytochemical components and human disease phenotypes by text mining over 21 million MEDLINE abstracts to support nutritional systems biology.

Key Features:

• Extensive Data Repository: Contains data on 17,772 plant-based foods, associations with 7,898 phytochemicals, links to 751 diseases covering 10,066 food items, and predicted associations for 548 phytochemicals and 252 diseases.
• Mechanistic Insights: Links the chemical space of plant-based foods directly with human disease phenotypes to enable exploration of mechanistic consequences of dietary behaviors.
• Text Mining and Classification: Systematically assembles knowledge spaces using text mining and Naïve Bayes classification to distinguish associations indicating disease prevention/amelioration from those indicating progression.
• Novel Discoveries: Analyzes frequently occurring phytochemical–disease pairs to aid identification of novel bioactive compounds with potential drug-like properties, illustrated by a colon cancer case study.
• Research Applications: Supports nutritional systems biology research by providing structured data for investigating the impact of diet on disease and for studying both well-characterized and lesser-known phytochemicals.

Scientific Applications:

• Mechanistic exploration: Study how specific phytochemicals from plant-based foods mechanistically influence human disease phenotypes.
• Bioactive compound discovery: Identify frequently associated phytochemical–disease pairs to prioritize compounds with potential therapeutic or drug-like properties.
• Disease-specific investigation: Use literature-derived associations to generate hypotheses for therapeutic targets and dietary strategies, exemplified by analyses on colon cancer.
• Diet–disease profiling: Classify and compare associations that indicate disease prevention/amelioration versus progression for nutritional intervention studies.

Methodology:

Text mining of over 21 million MEDLINE abstracts; assembly of knowledge spaces linking foods, phytochemicals and diseases; Naïve Bayes classification to distinguish prevention/amelioration versus progression associations; frequency analysis of phytochemical–disease pairs and inclusion of predicted associations for 548 phytochemicals and 252 diseases.