DeepCYP: an integrated deep learning web server for the holistic “pathway–site–product” prediction of CYP450 metabolism

Yiling Zhou, Sen Yang, Xiaoli Wang, Yuanhang He, Yao Tian, Jiacai Yi, Yikun Wang, Youchao Deng, Dejun Jiang, Dongsheng Cao · Nucleic Acids Research · May 2026

DeepCYP (https://deepcyp.scbdd.com) is a freely accessible deep-learning web server for end-to-end CYP450 metabolic profiling. Trained on an expanded dataset and a mechanism-based reaction rule library, it uses a multi-task graph neural network with multi-scale descriptors to cover the “pathway–site–product” continuum across nine major CYP isoforms in a single pipeline. Benchmarking shows gains over FAME3, SMARTCyp, and BioTransformer 3.0, and the server supports high-throughput batch processing for early assessment of metabolic liabilities.