For centuries, the human sense of smell has remained a mystery compared to our understanding of vision and hearing. While we can measure the wavelength of light for vision research or the frequency of sound for hearing research, there has been no reliable way to measure or predict the odor of a molecule based solely on its molecular structure. However, scientists have recently made a groundbreaking discovery in the field of olfaction, using machine learning to create a tool that can predict the odor profile of a molecule with remarkable accuracy.
Published in the prestigious journal Science, researchers from the University of Reading, in collaboration with the Monell Chemical Senses Center at the University of Pennsylvania, Arizona State University, and Osmo, a company associated with Google’s machine learning lab, have successfully developed an “odor map.” This revolutionary tool not only identifies molecules that look different but smell the same, but also molecules that appear similar but possess distinct odors.
Traditionally, flavor chemists and perfumers have relied on their own olfactory senses to describe aromas. However, this new machine learning-generated model has the potential to revolutionize the field. Professor Jane Parker, leading the research team at the University of Reading, emphasizes that the model correctly predicts the odor of exceptions, where previous models failed. This breakthrough has tremendous implications for synthetic chemists working in the food and fragrance industries.
One of the most intriguing aspects of this odor map is its versatility. It goes beyond merely working for known odorants and structurally similar molecules. The map can effectively describe unrelated molecules with distinct molecular characteristics. This opens up a wealth of untapped potential odorants that could be explored by researchers in the food and fragrance sectors. The number of possible odor combinations could potentially reach the millions.
As part of the research collaboration, the University of Reading played a crucial role in assessing the purity of the samples used to test the artificial intelligence (AI) model. By employing gas chromatography, the team was able to separate the trace levels of impurities from the target molecule. This process allowed them to evaluate whether any of the detected impurities overwhelmed or masked the odor of the target molecule.
Out of the 50 samples tested, the team discovered a few cases with notable impurities. In one instance, the impurity in question turned out to be traces of the reagent used in synthesizing the target molecule. These traces imparted a distinctive buttery smell that overpowered the intended odorant. This discrepancy between the panel’s description and the model’s prediction underscored the importance of purity in accurately decoding smells.
Once the AI model was trained with data, its ability to predict the smell of a novel compound proved to be excellent. In fact, its predictions matched the average scent scores of human panels. This breakthrough has significant implications for synthetic chemists in their quest to discover new aromas. With this tool, chemists can now screen large numbers of molecules for their aroma potential, similar to how the pharmaceutical industry screens compounds for new medicines.
The development of the odor map marks a turning point in the understanding and utilization of smell. By harnessing the power of machine learning, scientists have opened doors to the production of more sustainable flavors and fragrances. This breakthrough not only allows for the discovery of novel and exciting odors, but it also offers the potential for creating eco-friendly alternatives to existing flavors and fragrances that rely on limited natural resources.
The successful creation of an odor map using machine learning technology represents a major breakthrough for the field of olfaction. This innovative tool has the ability to predict the odor profile of molecules based on their structural characteristics alone, revolutionizing the work of synthetic chemists in the food and fragrance industries. With the potential to generate sustainable flavors and fragrances, the possibilities for future applications of this technology are numerous. The era of scent exploration has just begun, and this odor map is the compass that will guide us.
Leave a Reply