To mitigate the impact of false-negative associations caused by negative sampling of DPPs, the proposed SENT-DTI method is inspired by the advantage of negative training (NT) strategy on identification of false-negative samples and design a novel NT strategy that adaptively learns the probability distribution of known DPP features by incorporating a unified high-confidence false-negative association filtering mechanism into a negative loss objective function.
IHLO-DTI, a novel prediction model based on an improved hypergraph neural network and Laplacian matrix optimization, can effectively capture high-order many-to-many interactions between drugs and targets, improving prediction accuracy and robustness.
Guolongwei Dai, Tao Luo, Dandan Li et al.· ACS Synthetic Biology· 0 citations
Drug-target interaction (DTI) prediction is an efficient pre-screening method that uses algorithmic models to assess the binding potential of drug molecules to protein targets. Current research is accelerating towards the integration of heterogeneous graph neural networks, protein language models, and generative artificial intelligence. This review systematically summarizes the latest developments in these technologies, pointing out the problems currently being addressed in research such as data sparsity and cold start, as well as the manifestations of general machine learning challenges such as recommendation systems and noise learning in the biomedical field; Interpret representative models such as Dual Heterogeneous Graph Transformer for Drug-Target Interaction (DHGT-DTI) and Graph Positional encoding and Sequence features for Drug-Target Interaction (GPS-DTI). The combination of dual perspective learning, equivariant graph convolution, and attention mechanism enhances the understanding and reasoning ability of these models in complex biological networks. The generative artificial intelligence diffusion model has opened up a path for developing new drugs through structured and data enhanced approaches. Research has shown that important issues related to computational performance, interpretability, and data compatibility still need to be addressed in existing models. Building a high-performance, multifunctional pre trained model for large-scale biomolecules should be a key direction for development.
Qi-Zhong Yang· ITM Web of Conferences· 0 citations
This paper proposes TextDTI, a multimodal framework that simultaneously exploits sequential and structural representations and enhances feature alignment through adversarial learning and contrastive loss, resulting in robust and high-performance DTI prediction.
Jiaqi Deng, Senyu Tang, Jijun Tang et al.· Journal of Chemical Informat...· 0 citations
SGTL-DDA is proposed, a novel graph transformer framework designed to incorporate structural information and domain-specific knowledge from heterogeneous biological information networks (HBINs) that successfully identifies both known therapeutics and novel repositioning candidates, supported by molecular docking results and literature evidence.
Bowei Zhao, Hui Zhao, Yu-an Huang et al.· IEEE transactions on computa...· 0 citations
A multi-modal deep learning framework to predict drug-target affinity by integrating sequence semantics with graph structural information and design a new symmetric dual cross-attention fusion mechanism for drugs and targets.
Wei Lan, Tian Huang, Guohang He et al.· IEEE journal of biomedical a...· 0 citations
Current computational approaches for drug design typically focus on generating molecules conditioned on specific targets or general molecular properties, often neglecting the influence of disease context on target behavior and therapeutic outcomes. To address this gap, we introduce DrugGen-2, a novel generative model that designs small molecules conditioned on both disease ontology and target protein sequences. DrugGen-2 was developed by fine-tuning a pre-trained GPT-2 model on a curated dataset of approved drugs linked to their diseases and targets, using a two-step strategy of supervised fine-tuning followed by reinforcement learning via group relative policy optimization (GRPO). This process was guided by reward functions optimizing for chemical validity, novelty, diversity, and high predicted binding affinity. When evaluated on five protein targets relevant to diabetic nephropathy, DrugGen-2 significantly outperformed baseline models (DrugGPT and DrugGen). It demonstrated a superior capacity to generate unique molecules, exhibited greater structural similarity to approved drugs, and achieved improved predicted binding affinities across all targets. Molecular docking analyses further supported these findings, identifying candidate ligands with strong binding potential, including compounds with predicted affinities (-9.917, -9.485, and -9.367) exceeding those of reference drugs such as enalapril for angiotensin-converting enzyme (-8.283). By integrating disease-specific context into molecular generation, DrugGen-2 advances AI-assisted drug discovery, offering a powerful tool for de novo design and drug repurposing that accounts for the complex interplay between diseases and molecular targets.
Ali Motahharynia, Mohammadreza Ghaffarzadeh-Esfahani, Mahsa Sheikholeslami et al.· 0 citations