A Scalable Structure-Aware Multimodal Architecture for Accurate Drug-Target Affinity Prediction.

MultiGeo: Predicting Drug-Target Affinity via Adaptive Multi-Conformation Ensemble Learning

DiConSite: A Unified Topology-Adaptive Architecture for Protein Binding Site Prediction Across Ligand Modalities.

Accurate identification of protein binding sites is essential for understanding biological mechanisms and advancing drug design. However, many structure-based predictors rely on spatial graphs whose topology remains fixed throughout message passing, making them sensitive to structural noise and difficult to transfer across ligand modalities. To address this issue, we propose DiConSite, a topology-adaptive and reusable architecture for residue-level binding site prediction across ligand-specific tasks. DiConSite is centered on a Latent Topological Evolution (LTE) module that augments the initial Euclidean graph with a latent functional topology. A Hierarchical Topological Distillation (HTD) objective and a Dynamic Curriculum Distillation (DCD) schedule are further introduced as LTE-dependent optimization stabilizers: they align relational structure across network depths only after the underlying topology has been refined. Extensive experiments across nine benchmarks show that DiConSite achieves consistently strong and often best-performing results, while improving robustness to structural uncertainty and cross-modal variation. By combining protein language model embeddings with topology-adaptive geometric reasoning, DiConSite offers a reusable framework for residue-level protein interaction analysis.

StructureSAFE: A structure-aware chemical language model for unified hit identification and lead optimization

Structure-based generative models (SBGMs) hold great promises for accelerating drug discovery by enabling target-aware molecular design. However, existing approaches face fundamental challenges: three-dimensional graph-based models can explicitly incorporate protein structural information but often generate chemically implausible molecules due to limited training data, while chemical language models (CLMs) produce chemically plausible molecules but struggle to effectively leverage three-dimensional structural information for structure-conditioned generation and hard to incorporate lead optimization functionality due to the nature of SMILES string. Here, we present StructureSAFE, a structure-aware chemical language model that resolves this trade-off by integrating protein structural and evolutionary encoders with the SAFE molecular representation via pretraining and finetuning training scheme, enabling both de novo hit identification and a comprehensive suite of lead optimization subtasks within a unified framework. Comprehensive benchmarking on the MolGenBench dataset demonstrates that StructureSAFE achieves state-of-the-art (SOTA) performance across multiple metrics, with particularly pronounced improvements in chemical plausibility relative to graph-based models lacking pretraining. Evaluation on a rigorously constructed held-out test set further confirms its ability to generate drug-like, synthetically accessible molecules with competitive predicted binding affinities for previously unseen targets on both hit identification and lead optimization setting. In silico case studies across four therapeutically relevant targets validate its capacity to generate chemically plausible molecules that recapitulate key binding interactions of known high-affinity ligands while proposing novel interactions for potential better affinity and exploring previously unknown regions of chemical space. Taking together, StructureSAFE represents a versatile and practical tool to provide high-quality candidate molecules for augmenting medicinal chemistry workflows in both hit identification and lead optimization campaigns.

An Explicit Interaction-Prompted Diffusion Framework for High-Fidelity 3D Molecular Generation.

Current structure-based drug design generative models often struggle to faithfully recapitulate genuine ligand-protein binding interactions. Instead, under the coupling of implicit learning architectures and biased training data, they tend to learn spurious statistical correlations. To address this, we propose EIP-Diff (Explicit Interaction-Prompted Diffusion), an architecture featuring a novel explicit interaction-prompt embedding mechanism that is better suited for real-world target-specific drug design. This architecture replaces biased implicit learning with explicit, residue-level biological guidance, thereby promoting more fine-grained geometric fidelity and more precise interaction-aware conditioning. To fully realize the capabilities of EIP-Diff and provide a reliable basis for performance evaluation, we further constructed CrystalData set, which provides higher-fidelity and less-biased structural supervision than existing data sets. This explicit architecture markedly improves distribution consistency: even when trained on the crossdocked data set, EIP-Diff achieves the highest alignment with authentic pharmacological distributions among evaluated models. Training on CrystalData set further enhances this alignment and improves 3D geometric accuracy, while retaining strong controllability, high chemical space coverage, and near-perfect uniqueness. In addition, target-based validation on KAT6A and YTHDC1 confirmed that EIP-Diff accurately recapitulates native-like binding modes. Furthermore, in a real-world drug design task against IDO1, we successfully designed a novel lead compound with nanomolar potency (IC50 = 0.31 nM). These results demonstrate that the EIP-Diff architecture can explicitly leverage experimentally derived structural data and biologically meaningful interaction information for target-specific molecular generation, thereby enabling its effective application to real-world structure-based drug design.

READ: A Retrieval-Alignment Diffusion Framework for Structure-based Drug Design.

Structure-based drug design (SBDD) models are central to modern pharmaceutical research, enabling the rational exploration of protein-ligand interactions at atomic resolution. However, most existing approaches frame molecular generation as an isolated optimization or a one-to-one matching task, overlooking the shared binding patterns and intrinsic similarities among protein-ligand complexes. This fragmented perspective constrains their ability to capture the fundamental principles governing molecular recognition and binding specificity. Moreover, the limited availability of high-quality experimental data further hampers model generalization and real-world applicability. To address these challenges, we present READ, a retrieval-alignment molecular generation framework that conditions the generative process on small molecules targeting homologous proteins. Retrieved ligands are aligned with a diffusion model across multiple representational spaces and integrated as conditional guidance throughout successive stages of generation. Under a standardized docking-based evaluation protocol, READ achieves consistently strong performance against state-of-the-art SBDD methods. More importantly, it introduces a retrieval-alignment paradigm for structure-based molecular generation, offering a practical framework for early-stage computational hit generation while leaving prospective experimental validation as future work.

A Structure-Aware Multimodal Framework for Drug–Target Interaction Prediction via Heterogeneous Graph Learning

Predicting drug–target interactions is critical for drug discovery, yet many deep learning methods overlook atom–residue–level relationships. We propose Protein Heterogeneous Graph learning for Drug–Target Interaction prediction (PHGDTI), a multimodal framework that integrates sequence and structural cues for binding prediction. Drug and protein sequences are embedded with Mol2Vec and Tasks Assessing Protein Embeddings (TAPE) and refined by a self-attention module. In parallel, a drug–protein graph encoder models three complementary graphs: a drug atom graph, a protein residue graph, and a heterogeneous atom–residue graph. Graph attention layers propagate intra- and intermolecular information, and SAGPooling yields compact structural representations. Fusing these structural and sequence features enables accurate affinity estimation. Experiments on the Davis kinase dataset and GalaxyDB dataset show PHGDTI surpasses competitive baselines, and ablation results highlight the benefit of heterogeneous graph modeling.