HBGSA model uses hydrogen bond graph and self-attention for drug-target binding prediction

Researchers have developed HBGSA, a new 3.06M-parameter model designed to improve the prediction of drug-target binding affinity. This model addresses limitations in existing methods by incorporating spatial geometric constraints and hydrogen bond features, which are often overlooked. HBGSA utilizes graph neural networks with self-attention and a novel Pearson correlation loss function to enhance its ability to identify high-affinity compounds. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Improves drug discovery efficiency by prioritizing compounds for experimental validation.

RANK_REASON This is a research paper detailing a new model for drug-target binding affinity prediction.

Read on arXiv cs.LG →

• HBGSA
• PDBbind
• CSAR-HiQ

• paper
• model release