Application of artificial intelligence to decode the relationships between smell, olfactory receptors and small molecules

Abstract

Deciphering the relationship between molecules, olfactory receptors (ORs) and corresponding odors remains a challenging task. It requires a comprehensive identification of ORs responding to a given odorant. With the recent advances in artificial intelligence and the growing research in decoding the human olfactory perception from chemical features of odorant molecules, the applications of advanced machine learning have been revived. In this study, Convolutional Neural Network (CNN) and Graphical Convolutional Network (GCN) models have been developed on odorant molecules-odors and odorant molecules-olfactory receptors using a large set of 5955 molecules, 160 odors and 106 olfactory receptors. The performance of such models is promising with a Precision/Recall Area Under Curve of 0.66 for the odorant-odor and 0.91 for the odorant-olfactory receptor GCN models respectively. Furthermore, based on the correspondence of odors and ORs associated for a set of 389 compounds, an odor-olfactory receptor pairwise score was computed for each odor-OR combination allowing to suggest a combinatorial relationship between olfactory receptors and odors. Overall, this analysis demonstrate that artificial intelligence may pave the way in the identification of the smell perception and the full repertoire of receptors for a given odorant molecule.

Pred-O3, a web server to predict molecules, olfactory receptors and odor relationships

“…Two deep learning prediction models based on a Graphical Neural Network (GNN) have been added to the website to predict the potential odor and/or the potential interacting human ORs of a query compound. These models have been extracted from the article ( 25 ). The odor predictor can predict one or multiples odors class among a set of 23 odor note with a prc AUC of 0497 and the human ORs predictor can predict the potential interaction of a molecule with one or multiple human olfactory receptors among a set of 74 human OR with a prc-auc of 0,91.…”

“…For example, it is possible to search for chemical substructures to study functional groups connected to particular odors, as well as using similarity structural search methods to look for similar odorant molecules for a query chemical of interest. A prediction tool, recently developed internally, based on a deep learning approach (Graphical Neural Network) was implemented in the aim to predict the odors or olfactory receptors associated for a query compound ( 25 ). Finally, we precomputed 3D protein models for 1572 olfactory receptors and integrated them in the SeamDock server giving the possibility to dock any odorant molecules and to provide the potential mode of binding of an odorant to an olfactory receptor ( 26 , 27 ).…”

Volatile Composition of Fortification Grape Spirit and Port Wine: Where Do We Stand?

“…To advance in this challenging topic, it is crucial to apply high-throughput techniques, such as GC×GC-ToFMS, which seems to be a powerful technique for the analytical coverage of the chemical data on volatile compounds, or even used in combination with olfactometry and/or advanced artificial intelligence techniques. These chemical data seem to be useful in constructing predictive models to provide insights into the human perception of odorants, to predict odor from molecular structure, and to decode the relationships between smell, olfactory receptors and VOCs, among others [ 113 , 114 , 115 ].…”

Combining the Classification and Pharmacophore Approaches to Understand Homogeneous Olfactory Perceptions at Peripheral Level: Focus on Two Aroma Mixtures

“…Establishing the relationships between odors and molecular structures remains challenging [ 10 ]. The recent studies aiming to link the odors to the structure of the odorants involve large databases and the use of computational methods applying machine learning approaches [ 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 ]. These approaches are applied to perform classifications of the odorants considering both their odor notes and structural features.…”