Fragrances – promising mystery, intrigue and forbidden excitement – are mixed by master perfumers, recipes are kept secret. In a new study of odor perception, researchers at the Weizmann Institute of Science were able to clear most of the secrets from mixed odorant mixtures, not by revealing their secret substances, but by noting and mapping how they were perceived. Scientists can now predict how any complex odor will smell only from its molecular structure. This work will not only revolutionize the world of perfumery, but also lead to the ability to digitize and reproduce fragrances by order. The proposed framework for fragrances, created by neurobiologists, computer scientists and master perfumers and funded by the European initiative for Future Emerging Technologies (FET-OPEN), has been published. Nature.
Professor of the Department of Neurobiology of the Institute. “The problem of organizing smells in an organized and logical way was first proposed by Alexander Graham Bell 100 years ago,” says Noam Sobel. “We have a lot of different scents, from violet and rose to asafoetida,” Bell said. But you can’t have any knowledge of smell unless you can measure their similarities and differences. “It simply came to our notice then.
This centuries-old struggle has really highlighted the difficulty of placing odors in a logical system: There are millions of odor receptors in our noses, each consisting of hundreds of different subtypes, each designed to detect specific molecular features. Our brains potentially receive millions of odors in which these single molecules are mixed and mixed at different intensities. Thus, it was very difficult to map this data. However, Sobel and graduate student Aharon Ravia and Dr. Colleagues led by Kobe Snitz found that there was an order that lay beneath the smells. They came to this conclusion by adopting Bell’s concept – that is, to describe the relationship between smells, not the smells themselves.
In a series of experiments, the group presented volunteer pairs of scents and asked them to rate these scents as similar, and ranked the pairs on a scale of similarity between “same” and “extremely different.” In the initial experiment, the group created 14 aromatic mixtures, each consisting of about 10 molecular components, and presented them one to two times to about 200 volunteers, thus evaluating 95 pairs of each volunteer at the end of the experiment.
To turn the resulting thousands of perceptual similarity assessments into a useful design, the team refined a physico-chemical measure they had previously developed. In this calculation, each odorant is represented by a single vector that combines 21 physical dimensions (polarity, molecular weight, etc.). To compare two odorants, each represented by a vector, the angle between the vectors is taken to reflect the similarity of perception between them. A pair of scents with a low angle distance are similar, while those with a high angle distance are predicted differently.
To test this model, the team first met with Bushdid and Prof. at the Rockefeller Institute in New York. Extensive research on odor discrimination by colleagues from Leslie Vosshall’s laboratory was applied to data collected by others. The Weizmann team found that their models and sizes accurately predicted Bushdid’s results: it was difficult to distinguish odors with low angular distances between them; it was easy to distinguish odors with high angular distances between them. Encouraged by the model’s accurate prediction of data collected by others, the team continued testing for themselves.
The team invented new scents and invited a new group of volunteers to inhale the scents, again using their own methods to predict how this set of participants would evaluate pairs – first 14 new blends and then 100 blends in the next experiment. The model performed extremely well. In fact, the results were in the same ballpark as the color perception results – sensory information based on well-defined parameters. This was particularly surprising given that each individual has a specific type of odor receptor subtype, which can vary up to 30% between individuals.
This set of tools can not only predict how a new fragrance will smell, but also synthesize odorants by design.
Since an “odor map” or “metric” predicts the similarity of both odors, it can also be used to predict how an odor will eventually smell. For example, a substance that emits a new odor at a distance of 0.05 radians or less from a banana will have the same odor as a banana. As the new odorant moves away from the banana, it will smell like a banana and will stop looking like a banana after a certain distance.
The team is now developing a web-based tool. This set of tools can not only predict how a new fragrance will smell, but also synthesize odorants by design. For example, you can pick up any perfume with a number of known ingredients, and using a map and metric, you can create a new perfume that has no common components with the original perfume, but smells exactly the same. Such products of color vision, that is, non-overlapping spectral compositions that create the same perceived color, are called color metamers, and here the team created odor metamers.
Findings of the research Prof. Department of Computer and Applied Mathematics, who is also the vice-president of the Israeli Academy of Sciences and Humanities and co-author of the study. computers to digitize and replicate odors. In addition, of course, being able to add real floral or sea aromas to your holiday photos on social media, giving computers the ability to interpret smells the way people do, can have an impact on environmental monitoring and the biomedical and food industries. , to name a few. However, Christophe Laudamiel, a master perfumer who co-authored the study, said he was not worried about his profession yet.
Sobel concludes: “100 years ago, Alexander Graham Bell created a problem. We now answer: The distance between a flower and a violet is 0.202 radians (distant analogues), the distance between violets and asafoetida is 0.5 radans (very different) and the difference between a flower and asafoetida is 0.565 radians (they are even more different). We have counted the perceptions of smell, and this should really develop the science of smell. ”
Reference: “Allows the creation of one-dimensional scented metamers” by Aharon Ravia, Kobe Snitz, Danielle Honigstein, Maya Finkel, Rotem Zirler, Ofer Perl, Lavi Secundo, Christophe Laudamiel, David Harel and Noam Sobel, November 11, 2020, Nature.
DOI: 10.1038 / s41586-020-2891-7
Prof. Noam Sobel is the head of the Azrieli National Institute for Human Brain Imaging and Research; The study is also supported by the Norman and Helen Asher Human Brain Imaging Center; Nadia Jaglom Laboratory for Research in Olfaxion Neurobiology; Rob and Cheryl McEwen Brain Research Foundation; and Sonia T. Marschak. Prof. Prof. Sobel Sara and Michael Sela. He is the head of the neurobiology department.
Prof. David Harel’s research is supported by Emile Mimra’s estate. Prof. Harel, William Sussman Professor is the Head of the Department of Mathematics.