A principal odor map unifies diverse tasks in olfactory perception

Lee, B. K. ORCID: https://orcid.org/0000-0002-0920-3520, Mayhew, E. J. ORCID: https://orcid.org/0000-0001-7881-2306, Sanchez-Lengeling, B. ORCID: https://orcid.org/0000-0002-1116-1745, Wei, J. N. ORCID: https://orcid.org/0000-0003-3567-9511, Qian, W. W. ORCID: https://orcid.org/0000-0003-0726-575X, Little, K. A. ORCID: https://orcid.org/0009-0001-3455-0217, Andres, M. ORCID: https://orcid.org/0009-0004-7787-7473, Nguyen, B. B., Moloy, T. ORCID: https://orcid.org/0000-0002-8372-560X, Yasonik, J. ORCID: https://orcid.org/0000-0003-3307-7955, Parker, J. K. ORCID: https://orcid.org/0000-0003-4121-5481, Gerkin, R. C. ORCID: https://orcid.org/0000-0002-2940-3378, Mainland, J. D. ORCID: https://orcid.org/0000-0002-5056-4598 and Wiltschko, A. B. ORCID: https://orcid.org/0000-0001-9947-1213 (2023) A principal odor map unifies diverse tasks in olfactory perception. Science, 381 (6661). pp. 999-1006. ISSN 1095-9203 doi: 10.1126/science.ade4401

Abstract/Summary

Mapping molecular structure to odor perception is a key challenge in olfaction. We used graph neural networks to generate a principal odor map (POM) that preserves perceptual relationships and enables odor quality prediction for previously uncharacterized odorants. The model was as reliable as a human in describing odor quality: On a prospective validation set of 400 out-of-sample odorants, the model-generated odor profile more closely matched the trained panel mean than did the median panelist. By applying simple, interpretable, theoretically rooted transformations, the POM outperformed chemoinformatic models on several other odor prediction tasks, indicating that the POM successfully encoded a generalized map of structure-odor relationships. This approach broadly enables odor prediction and paves the way toward digitizing odors.

Item Type	Article
URI	https://reading-clone.eprints-hosting.org/id/eprint/113304
Item Type	Article
Refereed	Yes
Divisions	Life Sciences > School of Chemistry, Food and Pharmacy > Department of Food and Nutritional Sciences
Uncontrolled Keywords	Cheminformatics, Humans, Olfactory Perception, Smell, Neural Networks, Computer, Odorants
Publisher	American Association for the Advancement of Science
Download/View statistics	View download statistics for this item