G-K BertDTA: A graph representation learning and semantic embedding-based framework for drug-target affinity prediction

Xihe Qiu, Haoyu Wang, Xiaoyu Tan, Zhijun Fang

Abstract: Highlights•A novel DenseSENet is designed to capture key protein features with all layer inputs.•We enhance SMILES feature extraction with GIN and CNNs to tackle missing labels.•KB-BERT learns SMILES’s semantic features, considering previously ignored aspects.•The framework uses graph learning and semantic embeddings for precise DTA prediction.