Collaborative multi-agent intelligence uncovers subtype-selective allosteric sites at GPCR-lipid interaction interface

This bioRxiv preprint describes a collaborative multi-agent system that systematically analyzes class A GPCRs to find protein-membrane-interface sites that differ between closely related subtypes. The workflow integrates dMaSIF-derived surface fingerprints with Ballesteros-Weinstein (BW) position alignment to compare structurally equivalent membrane-facing residues. Because orthosteric pockets are highly conserved, the lipid interface offers an underutilized source of subtype selectivity. The approach could accelerate the design of allosteric modulators with fewer off-target effects across GPCR families. The authors demonstrate its utility in identifying divergent sites suitable for selective ligand binding, providing a computational blueprint for future experimental validation.