DSAGLSTM-DTA: Prediction of Drug-Target Affinity using Dual Self-Attention and LSTM

The research on affinity between drugs and targets (DTA) aims to effectively narrow the target search space for drug repurposing. Therefore, reasonable prediction of affinities can minimize waste resources such as human material resources. In this work, a novel graph-based model called DSAGLSTM-DTA was proposed DTA prediction. is unlike previous drug-target model, which incorporated self-attention mechanisms in feature extraction process molecular graphs fully extract its effective representations. features each atom 2D graph were weighted based attention score before being aggregated molecule representation two distinct pooling architectures, namely centralized distributed architectures implemented compared benchmark datasets. addition, course processing protein sequences, inspired by approach GDGRU-DTA, we continue interpret sequences time series their using Bidirectional Long Short-Term Memory (BiLSTM) networks, since context-dependence long amino acid sequences. Similarly, also utilized mechanism obtain comprehensive representations proteins, final hidden states element batch with unit output LSTM, results represented proteins. Eventually, concatenated fed into block evaluated different regression datasets binary classification datasets, demonstrated that superior some state-ofthe-art models exhibited good generalization ability.