RT Journal Article
SR Electronic
T1 A Principal Odor Map Unifies Diverse Tasks in Human Olfactory Perception
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 2022.09.01.504602
DO 10.1101/2022.09.01.504602
A1 Brian K. Lee
A1 Emily J. Mayhew
A1 Benjamin Sanchez-Lengeling
A1 Jennifer N. Wei
A1 Wesley W. Qian
A1 Kelsie Little
A1 Matthew Andres
A1 Britney B. Nguyen
A1 Theresa Moloy
A1 Jane K. Parker
A1 Richard C. Gerkin
A1 Joel D. Mainland
A1 Alexander B. Wiltschko
YR 2022
UL http://biorxiv.org/content/early/2022/09/06/2022.09.01.504602.abstract
AB Mapping molecular structure to odor perception is a key challenge in olfaction. Here, we use graph neural networks (GNN) to generate a Principal Odor Map (POM) that preserves perceptual relationships and enables odor quality prediction for novel odorants. The model is as reliable as a human in describing odor quality: on a prospective validation set of 400 novel odorants, the model-generated odor profile more closely matched the trained panel mean (n=15) than did the median panelist. 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.One-Sentence Summary An odor map achieves human-level odor description performance and generalizes to diverse odor-prediction tasks.Competing Interest StatementThe authors have declared no competing interest.