DeepETD and EMTDD

Metabolites participate in nearly all fundamental biological processes in vivo. Accumulating new phenotypes of classic metabolites have been revealed; yet the underlying targets and mechanisms remain largely unexplored. Here, we developed a deep learning model named DeepETD (Deep learning-based Endogenous metabolites Target Discovery) that integrates bioinformatic data and an attention mechanism to predict the functional targets of specific metabolite phenotypes. Using this model, we constructed a publicly accessible database named EMTDD (Endogenous Metabolite Target Discovery Database) containing potential targets for 3,382 common human metabolites. Through the application of this model, we identified norepinephrine as a selective peroxidase inhibitor of peroxiredoxin 1 (PRDX1) for atherosclerosis aggravation in model validation experiments.

Citation

Norepinephrine aggravates atherosclerosis by inhibiting PRDX1: Evidence from a new deep-learning model

PMID : XXXXX