GDEGAN: Gaussian Dynamic Equivariant Graph Attention Network for Ligand Binding Site Prediction

ICLR 2026 Conference Submission25261 Authors

Keywords: equivariant gnns, protein ligand interaction, binding site identification, statistical attention

TL;DR: By recognizing that binding pockets have distinct statistical signatures, GDEGAN improves ligand binding site prediction by 37% and upto 20× faster than current methods while doing inference.

Abstract:

Accurate prediction of binding sites of a given protein, to which ligands can bind, is a critical step in structure-based computational drug discovery. Recently, Equivariant Graph Neural Networks (GNNs) have emerged as a powerful paradigm for binding site identification methods due to the large-scale availability of 3D structures of proteins via protein databases and AlphaFold predictions.
The state-of-the-art equivariant GNN methods implement dot product attention, disregarding the variation in the chemical and geometric properties of the neighboring residues. To capture the variation in properties, we propose GDEGAN (Gaussian Dynamic Equivariant Graph Attention Network), which replaces simple dot-product attention with adaptive kernels that recognize binding sites. The proposed attention mechanism captures variation in neighboring residues using statistics of their characteristic local feature distributions. Our mechanism dynamically computes neighborhood statistics at each layer, using local variance as an adaptive bandwidth parameter with learnable per-head temperatures, enabling each protein region to determine its own context-specific importance.
Our model shows better predictive performance, outperforming existing methods with relative improvements of 37-66 \% in DCC and 7-19 \% DCA success rates across COACH420, HOLO4k, and PDBBind2020 datasets. These advances have direct application in accelerating protein-ligand docking by identifying potential binding sites for therapeutic target identification.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Submission Number: 25261