PocketDTA: A pocket-based multimodal deep learning model for drug-target affinity prediction

Jiang Xie, Shengsheng Zhong, Dingkai Huang, Wei Shao

Abstract: Highlights•A novel model named PocketDTA is proposed for drug-target affinity prediction.•PocketDTA combines sequence and structural data to improve prediction performance.•PocketDTA overcomes the traditional limitations of lack of spatial information.•Experimental results demonstrate that PocketDTA outperforms SOTA models.