DietRx

DietRx provides a platform for exploring health impacts of dietary ingredients by integrating interrelationships among food and key molecular agents. The resource assimilates dietary factors (food and chemicals), their health consequences (diseases) and genetic mechanisms to facilitate queries for investigating associations among these entities. At the core of the DietRx is the data of 21207 positive/negative food-disease associations for 1781 food entities belonging to 24 categories (vegetable, plant, fruit, meat & egg, herbs & spices etc.) text-mined from biomedical literature (27 million MEDLINE abstracts) using state-of-the-art named entity recognition tools, and a deep learning based relation classification model (Precision=0.87, Recall=0.8, and F1 Score = 0.84) which was trained with significant amount of manually curated data. These data are further interlinked with those involving 6992 food chemicals and 20550 genes, compiled from curated data sources, thereby creating a seamless platform for probing elements central to diet and their health consequences. DietRx facilitates the study of associations among food, disease, chemicals, and genes to enable data-driven inferences to be used for culinary interventions, nutrigenomics insights as well as for drug discovery endeavors.