麻豆传媒

鈥淭oying with our sense of smell is like playing a piano with 400 keys鈥� (Image: Paola Nogueras/Monell Chemical Senses Center)

How do you go about predicting and manipulating smells?
Humans have about 400 types of working odour receptors in the nose but each person has a slightly different set, meaning that odours are often perceived differently by people. This makes prediction a hard challenge. And the truth is, science has been pretty bad at predicting what something will smell like just from its molecular structure. To do it, you need to understand how molecules interact with smell receptors. My team started by pinning down how single molecules interact, and then moved on to cocktails of molecules.

Can you give an example of individuals perceiving a smell in contrasting ways?
There鈥檚 a substance called androstenone, found in sweat and urine, and its receptor is called OR7D4. About 30 per cent of the population can鈥檛 smell it at all. As for the other 70 per cent, half smell it as urinous 鈥� really strong 鈥� and half smell it as sandalwood and kind of weak. So there are big differences in its perception: pleasant versus unpleasant, strong versus weak and then people who essentially can鈥檛 detect it at all.

So it鈥檚 a tough task you鈥檝e set yourselves.
Certainly. But we have a computer model that can take receptor activation into account, and we explore the predictions it makes using our purpose-built olfactometer 鈥� a machine that mixes odours and presents the result to the nose.

What is your ultimate aim?
In the same way that primary colours can make any colour you want, we are in the process of identifying a palette of 40 鈥減rimary odours鈥� that we hope can be combined in various ways to create the smell of anything that you might experience in daily life. If we can work out how to manipulate each of our 400 or so receptors, it means I can 鈥減lay鈥� your olfactory system to perceive any smell I want. It鈥檚 complicated because very few odours activate just one receptor and not another. And there are other drawbacks 鈥� it鈥檚 like playing a piano with 400 keys and you have to hit 100 simultaneously. Sometimes you just won鈥檛 have enough fingers.

Are you making progress?
It鈥檚 going to be pretty ugly for a while yet, but we are getting somewhere. As an example, we can take four molecules that don鈥檛 smell like strawberry, mix them together using the olfactometer and they smell like strawberry.

If we could reproduce everything in the human 鈥渟mellscape鈥� using primary odours, what are the likely applications?
When you know how to reproduce all odours, you can then digitise that information. So for starters you could archive odours, transmit them through the internet to be reproduced with suitable hardware, or add them to television and movies. Such technology could even eliminate supply chains that, for example, destroy sandalwood ecosystems. It鈥檚 hard to imagine the full potential.

This article appeared in print under the headline 鈥淵our nose will be my piano鈥�