Rong-Qi Hong

GPCR Ligand Bioactivity Prediction with Physics-Informed Dual-State Query Learning

Dual-State Query is proposed, a physics-informed multimodal architecture that explicitly embeds the Monod-Wyman-Changeux model of allostery within a neural network that outperforms other baselines and provides useful thermodynamic inductive bias, while also exposing a clear limitation on low-homology receptors.