Changli Li

Adaptive Self-Attention Graph Pooling for Drug–Target Affinity Prediction

The Adaptive Self-Attention Graph Pooling (ASAGPooling) mechanism is proposed, which introduces a learnable pooling ratio that dynamically adjusts node retention during training and develops ASAG-DTA, a multi-modal framework that integrates GNNs with Transformers to jointly model molecular graphs, protein contact maps, SMILES sequences, and FASTA sequences.