Computational modelling of olfactory receptors

Abstract

Olfactory receptors (ORs), the largest subfamily of G protein-coupled receptors, are essential for detecting and interpreting environmental odorants in animals. Understanding their function is crucial for deciphering olfactory perception and exploring emerging roles in non-olfactory systems. With the recent surge in available sequence data and AI-based structural predictions, computational modelling has become indispensable for investigating OR structure, ligand binding, and activation mechanisms. This review provides a comprehensive overview of computational approaches used in OR research, including homology modelling, molecular docking, molecular dynamics simulations, free energy calculations, pharmacophore modelling, virtual screening, and machine learning-based predictions. Both ligand-based and structure-based pharmacophore modelling are discussed in detail, highlighting their respective applications, strengths, and limitations. While structure-based approaches have gained prominence due to advances in receptor structure prediction tools like AlphaFold, ligand-based pharmacophore modelling remains valuable in scenarios where structural data are limited or uncertain. Case studies illustrate how these techniques have been applied to identify novel OR–ligand interactions, explore receptor dynamics, and support drug discovery. Collectively, these computational strategies offer powerful tools for decoding OR function, guiding experimental validation, and expanding our understanding of olfactory signalling in health and disease.