TY - JOUR
T1 - Predicting human olfactory perception from chemical features of odor molecules
AU - Keller, Andreas
AU - Gerkin, Richard
AU - Guan, Yuanfang
AU - Dhurandhar, Amit
AU - Turu, Gabor
AU - Szalai, Bence
AU - Mainland, Joel D.
AU - Ihara, Yusuke
AU - Yu, Chung Wen
AU - Wolfinger, Russ
AU - Vens, Celine
AU - Schietgat, Leander
AU - De Grave, Kurt
AU - Norel, Raquel
AU - Stolovitzky, Gustavo
AU - Cecchi, Guillermo A.
AU - Vosshall, Leslie B.
AU - Meyer, Pablo
PY - 2017/2/24
Y1 - 2017/2/24
N2 - It is still not possible to predict whether a given molecule will have a perceived odor or what olfactory percept it will produce.We therefore organized the crowd-sourced DREAM Olfaction Prediction Challenge. Using a large olfactory psychophysical data set, teams developed machine-learning algorithms to predict sensory attributes of molecules based on their chemoinformatic features.The resulting models accurately predicted odor intensity and pleasantness and also successfully predicted 8 among 19 rated semantic descriptors ("garlic," "fish," "sweet," "fruit," "burnt," "spices," "flower," and "sour"). Regularized linear models performed nearly as well as random forest-based ones, with a predictive accuracy that closely approaches a key theoretical limit.These models help to predict the perceptual qualities of virtually any molecule with high accuracy and also reverse-engineer the smell of a molecule.
AB - It is still not possible to predict whether a given molecule will have a perceived odor or what olfactory percept it will produce.We therefore organized the crowd-sourced DREAM Olfaction Prediction Challenge. Using a large olfactory psychophysical data set, teams developed machine-learning algorithms to predict sensory attributes of molecules based on their chemoinformatic features.The resulting models accurately predicted odor intensity and pleasantness and also successfully predicted 8 among 19 rated semantic descriptors ("garlic," "fish," "sweet," "fruit," "burnt," "spices," "flower," and "sour"). Regularized linear models performed nearly as well as random forest-based ones, with a predictive accuracy that closely approaches a key theoretical limit.These models help to predict the perceptual qualities of virtually any molecule with high accuracy and also reverse-engineer the smell of a molecule.
UR - http://www.scopus.com/inward/record.url?scp=85013999403&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85013999403&partnerID=8YFLogxK
U2 - 10.1126/science.aal2014
DO - 10.1126/science.aal2014
M3 - Article
C2 - 28219971
AN - SCOPUS:85013999403
SN - 0036-8075
VL - 355
SP - 820
EP - 826
JO - Science
JF - Science
IS - 6327
ER -