Drug target interactions predicted using combined chemical and protein language models
Drug-Target Interaction Prediction via Hierarchical Sequential Cross-Attention over Chemical and Protein Language Models
Machine LearningArtificial Intelligence
Summary
Predicting how drugs interact with proteins is important for developing new medicines. The authors designed a method that looks at the sequences of drugs and proteins together, rather than separately, to better understand their interactions. They used two language models trained on drug and protein sequences and combined them through a special attention mechanism to spot detailed patterns. Their approach showed strong results on several datasets and can work well even on unseen drugs or proteins.
What this means in practice
- •For drug discovery teams: Improve early-stage screening by predicting drug and protein binding using sequence-only models that capture detailed interaction patterns.
- •For bioinformatics engineers: Integrate hierarchical sequential cross-attention modules into pipelines to enhance prediction of molecular interactions from raw sequence data.
Authors
Khadidja Henni, Hamza Abdelali, Abdelkrim Aries, Neila Mezghani, Brigitte Vannier, Sara Magdouli, Lina Abou-Abbas
Abstract
Predicting Drug-Target Interactions~(DTIs) is a central task in computational drug discovery, with direct applications in virtual screening, drug repurposing, and therapeutic candidate prioritization. Although recent deep learning methods have improved DTI prediction, many sequence-based models still process drugs and proteins independently and only combine their representations at a late prediction stage. This limits their ability to explicitly model cross-molecular dependencies between chemical substructures and protein sequence regions. In this paper, we propose a sequence-only DTI prediction architecture that combines two pre-trained language models, ChemBERTa for drug SMILES strings and ESM-2 for protein amino acid sequences, with a hierarchical interaction module. The proposed model first extracts contextual representations using pre-trained encoders, then applies 1D convolutional layers to condense local sequence patterns, followed by a sequential bidirectional cross-attention mechanism inspired by the induced-fit view of molecular recognition. Finally, attention-based pooling constructs fixed-size interaction-aware vectors for binary prediction. Experiments on BIOSNAP, Davis, and BindingDB show that the proposed model achieves the best performance on BIOSNAP, matches the best AUROC on Davis, and remains competitive on BindingDB while using only 25.2 million trainable parameters. Ablation results confirm the contribution of both the CNN and cross-attention modules, and cold-start experiments indicate promising generalization to unseen proteins and drugs.