There are a number of molecules which no human should ever have to taste for safety reasons. Still, it could be interesting to know how it interacts with the taste buds. Applications could also include knowing how much to dilute or mix in aroma to medication in the pharmaceutical industry.

What methods are there to determine the taste of a compound without putting a human in danger?

  • 1
    $\begingroup$ The taste is a particular property (the same thing taste different for different persons) that does not indicate danger. To know if a molecule is danger, there are ADME (Adsorption, Distribution, Metabolism and Excretion) and toxicology simulations. $\endgroup$ – Camps Oct 26 '20 at 11:42

An interesting, rather open-ended question. The fact that we don't really have a proper understanding of the mechanism that results in the taste, combined with the fact that taste is also different from individual to individual makes it borderline impossible to set up a good physical model for this question.

However, this has never stopped cheminformaticians before. There's a vast majority of properties, more and less abstract ones, that can be correlated to known properties of the molecules (so-called descriptors). For example you will find a lot of equation that let's you characterize the efficiency of a drug molecule if you know for example its size, water/octanol solubility ratio, acidity constants and similar descriptors. This problem of finding drug efficiency from these rather simple quantities shows that predicting taste should indeed be possible.

So how could someone do this? First of all, since you don't have a physical/chemical theory, you need a somewhat reliable dataset - this is actually available. If you have your data ready, you should select a set of good descriptors, either by using serious statistical tests, or using your own intuition. Then you need to establish a relationship between the descriptors and the predicted properties. This step is usually done using a variety of machine learning tools.

In fact, a quick Googling shows me that this approach had recently been taken up in the literature, and taste prediction of a chemical compound does exist: look here, or here for example.

I hope I explained the idea in layman's terms well enough that you can understand the point of these publications as well as search for similar ones yourself.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.