How we smell

Carl Zimmer has an interesting article on the biology of smell. It is truly remarkable how our noses can detect the presence of a tiny amount of one type of molecule in the vast expanse of other molecules.

He first describes the process by which we register the presence of a smelly molecule.

When you smell a lily or a cigar or a jug of spoiled milk, you are grabbing their molecules out of an ocean of air. You have exposed nerve endings dangling deep inside your nostrils, each of which is studded with proteins called olfactory receptors. Each neuron is covered in one type of receptor, the shape of which allows it to grab tightly onto certain odor molecules and weakly to others, while letting many others drift by.

Snagging a molecule causes the receptor to squirm, leading to a falling-domino-like series of reactions that ends with the neuron firing an electric signal into your brain. The brain gets signals from thousands of neurons in our noses, creating a distinctive signature for each kind of smell we perceive.

We can distinguish between a vast number of different odors, thanks in part to the vast number of olfactory receptor genes our neurons can choose from. So far scientists have identified 390 different genes in the human genome that encode olfactory receptors.

I don’t know that I would describe 390 as a ‘vast number’ but the point is that being able to detect 390 different kinds of smells is pretty impressive.

He then goes on to discuss how in addition to these functioning genes, the human genome has 468 olfactory receptor genes that neurons cannot use to make a receptor protein and hence do not serve any smelling purpose. These are called pseudogenes and the number of them varies depending on the species.

Animals that depend heavily on smell often have huge numbers of olfactory genes, and relatively few pseudogenes. Rats, for example, have 1207 working olfactory receptor genes, for example, and only 508 pseudogenes–roughly a quarter knocked out. Humans have far fewer olfactory receptor genes, and about 55% of them no longer make receptors. Some other species have suffered even greater losses. Chickens have just 554 genes, of which 476 (86 percent) are pseudogenes. Sperm whales, whose ancestors returned to the sea, lost all their working olfactory receptor genes. They only have pseudogenes left.

So why are these pseudogenes there at all? They are the byproduct of the evolutionary process and relating the olfactory genes and pseudogenes to those of other species in our evolutionary history, we get to understand how they happened to appear.

Olfactory receptor genes are located in regions of our chromosomes that are especially prone to copying mistakes. From time to time, an olfactory receptor gene’s DNA gets copied out twice, creating a duplicate of the original.

After the duplication, one of the two identical genes may mutate. Sometimes the mutation disables the gene, turning it into a pseudogene. Losing the gene isn’t a catastrophe, however, since the other copy of the gene still survives. Sometimes a copying error may delete the pseudogene altogether.

In other cases, the new copy mutates, but not fatally. It may continue making the same olfactory receptor, despite the slight change its DNA. Or the receptor’s structure changes. Its different shape alters its grip on odor molecules. That change may trigger a subtle shift in the smells an animal can sense.

Run this operation over and over again for millions of years, and you end up with the diversity of smelling genes found today in living creatures. Like humans, other mammals have hundreds of working genes and hundreds more pseudogenes. Functional or not, the genes in different species belong to different “families” that originated in duplications millions of years ago. The fact that pseudogenes are “cousins” to functional genes is evidence that they once functioned.

One consequence of this is that there can be considerable variation within species too, resulting in different people having different sensitivities to different smells. My intuitive sense is that the women in my family and workplace seem to be far more sensitive to smells than me, often being able to detect the specific ingredients that went into cooked foods. While there is some support for this view and that the ability to smell gets better with repeated exposure, in my case, its pronounced effect may not be due to any general gender discrepancy but simply that I happen to have a fairly poor sense of smell. I don’t know if this is because I have fewer of the working olfactory receptor genes or whether some of them are kind of flabby and send signals to my brain that are too weak to register.

Fascinating stuff.


  1. StevoR : Free West Papua, free Tibet, let the Chagossians return! says

    I can’t tell how anyone smells over the internet.

    I’m kinda grateful for that.

    But this piece about the mechanics of how we process aromas – pretty interesting, thanks.


    Hear about the dog with no nose?
    How did it smell?

  2. ChasCPeterson says

    When you smell a lily or a cigar or a jug of spoiled milk, you are grabbing their molecules out of an ocean of air.

    ,,,and they go into solution in the aqueous film covering the olfactory epithelium, and that goes for a turd or a fart as well. (ew)

    I don’t know that I would describe 390 as a ‘vast number’ but the point is that being able to detect 390 different kinds of smells is pretty impressive.

    The vast number comes from the fact that nearly all ‘odors’ consist of combinations of these 390+ basic chamical stimuli.

  3. karmacat says

    I always wonder if dogs feel sorry for us because we are smell-challenged compared to them. : )

  4. Matt G says

    Mice have around 1200 functional genes and only a handful of pseudogenes. The thing is, they don’t smell more things than humans – they just discriminate better. When the human says “it could be odor 1 to 20 out of 100”, the mouse says “it’s odor 13”.

  5. Matt G says

    While humans have fewer than 400 odor *receptors*, there are far more odorants (i.e., full and partial agonists for odor receptors) – the estimates range from 10,000 to 100,000 last I knew.

  6. Mano Singham says

    Following up this post, I found this paper that says that the authors have identified that dogs have 1,094 olfactory genes of which only about 20% are pseudogenes.

  7. Matt G says

    My memory is a bit off. According to a 2002 paper, the total number of OR genes in mice is 1296, with 20% pseudogenes. I thought the number of pseudogenes was much lower for some reason.

Leave a Reply

Your email address will not be published. Required fields are marked *