The Body Odour Gene

The ABCC11 Gene and rs17822931

The ABCC11 gene acts as a pump in your apocrine sweat glands, the glands clustered in your armpits, groin, and around the nipples. These are not the glands that cool you when you exercise (those are eccrine glands, and they produce mostly salt water). Apocrine glands release a thick, oily fluid that does not actually smell on its own. The smell happens downstream: bacteria on your skin, primarily Corynebacterium and Staphylococcus, eat the secreted precursors and excrete the volatile compounds we recognise as body odour (Natsch & Emter, 2020).

If you have at least one working copy (the G allele, genotypes GG or GA), the pump works. Precursors hit your sweat, bacteria feast on them, and you get body odour alongside wet, sticky earwax. If both your copies are the A allele (AA genotype), the pump is broken. No precursors reach the skin. Bacteria have nothing to work with. You get minimal or no underarm body odour and dry, flaky earwax (Yoshiura et al., 2006).

Your earwax type and your body odour come from the same gene, and that is not a coincidence. The ceruminous glands of the ear canal and the apocrine glands of the armpit both depend on the ABCC11 transporter. Rodriguez et al. (2009) confirmed in the Journal of Investigative Dermatology that a functional ABCC11 allele is essential for axillary odour formation. No working transporter, no raw materials for bacteria, no smell. They also found that AA individuals often still buy deodorant due to social convention, despite not needing it.

An Ancient Mutation — Where and When

The smelly G allele is actually the original, ancestral version. Mammals use apocrine glands to signal each other through scent. Body odour is the evolutionary default. The A allele is a newer loss-of-function mutation that arose approximately 50,000 years ago, likely in the Siberia/Central Asia region. It appears in the Ust'-Ishim lineage (~45,000 years ago) and in the Tianyuan individual from Northern China (~40,000 years ago).

The geographic distribution today is striking:

• East Asian populations: 80–95% carry the AA genotype (no body odour). In Korea and Japan, the frequency approaches 100%.
• European and African populations: 97–100% carry at least one G allele (body odour present).
• South Asian and Central Asian populations: Intermediate frequencies, reflecting historical mixing.

The A allele took over East Asia incredibly fast for a single nucleotide change, a speed that usually points to positive selection (Ohashi et al., 2011). Losing body odour must have given these populations a serious edge in survival or reproduction. Geneticists now use rs17822931 as an ancestry-informative marker because few other SNPs show such a stark continental divide.

Why We Evolved Body Odour in the First Place

If body odour were merely an unfortunate byproduct of sweating, there would be no reason for dedicated glandular machinery to produce it. However, apocrine glands are not random. The biological intent is clear.

These glands wake up at puberty, precisely when sexual signaling becomes relevant. They cluster exclusively in reproductive zones (armpits, groin, areolae), not the palms or forehead. They pump out specific chemical precursors rather than generic sweat, and bacteria convert those precursors into individually distinctive volatile compounds. Our noses are tuned for this: Natsch and Emter (2020) note that humans show "particularly high sensitivity" to axillary odours compared to other environmental scents, even when we consciously try to scrub them off.

Body odour patterns are stable over time, genetically determined, and individually specific. These are not the characteristics of metabolic waste. These are the characteristics of a communication system, one that enables kin recognition, individual identification, and potentially mate-quality assessment.

Why Evolution Then Removed It in Half the World

If body odour served an important signalling function, why did natural selection eliminate it in East Asian populations? Three hypotheses compete:

1. Cold climate adaptation (strongest evidence): Ohashi et al. (2011) found that the frequency of the A allele correlates with absolute latitude — the further north, the more common. In freezing temperatures, sweating less saves crucial body heat and moisture. The populations that migrated through Siberia and into Northeast Asia faced some of the coldest environments humans have ever inhabited.
2. Sexual selection for odourlessness: Natsch & Emter (2020) suggest that preference for low-odour partners may have driven the loss-of-function allele to fixation: "An odourless phenotype became a preferred social attribute early on in ancient Asian cultures." If those ancient cultures liked odourless partners, sexual selection alone could have spread the A allele quickly.
3. Population density in agrarian societies: In a close-knit agrarian settlement, not smelling like a locker room was likely a significant social advantage. When people transitioned from small nomadic groups to crowded villages, having a strong odour ceased to be attractive and began to pose a problem.

These hypotheses are not mutually exclusive. Cold may have initiated the spread, and sexual selection for odourlessness may have completed the task.

Cultural echoes persist today. In South Korea and Japan, pronounced body odour (osmidrosis) is sometimes treated as a medical condition warranting surgical removal of apocrine glands. In Western cultures, the same level of scent would be considered entirely normal.

The Sweaty T-Shirt Experiment — Foundational but Contested

In 1995, Claus Wedekind published the famous sweaty t-shirt study. Men wore the same t-shirt for two nights. Women smelled the shirts and rated them. The finding: women preferred the smell of men whose MHC (Major Histocompatibility Complex) genes were dissimilar to their own. It launched an entire field of research into genetic compatibility (Wedekind et al., 1995).

DNA Romance's DRom 1.0 algorithm relies on this premise, using 100 MHC SNPs to calculate how dissimilar your HLA is from a potential partner's. Dissimilar MHC, the theory goes, means you will find each other's scent attractive. That is "chemistry."

The evidence, however, has become more complicated since 1995:

• A major review by Natsch & Emter (2020) in Philosophical Transactions of the Royal Society B found "no evidence for HLA-linked patterns" in the composition of odorous compounds. They note that MHC proteins probably do not bind to odorant precursors, so we still do not know exactly how MHC genes would change your scent.
• The original sweaty t-shirt results were "not replicated in a larger study," according to the same review.
• A meta-analysis by Winternitz et al. (2017), covering multiple studies, found no significant overall effect of MHC dissimilarity on odour preference (Zr = −0.024, p = 0.289).
• Derti & Cenik (2010) published a paper titled "Absence of Evidence for MHC-Dependent Mate Selection" in PLoS Genetics, questioning whether the effect exists at all in humans when examined at a population level.

The original sweaty t-shirt experiment opened a genuinely interesting line of inquiry. But the evidence for a direct MHC-to-odour-to-preference pathway is more mixed than the pop-science coverage suggests, and the honest thing is to say so.

The Confound — When There Is No Scent to Evaluate

Now layer the ABCC11 finding onto the MHC-odour hypothesis.

If you both carry the G allele (both produce body odour), the scent-based compatibility test works as advertised. You can both smell each other, and if your MHC genes are dissimilar, each of you should find the other's scent attractive.

But if one partner carries AA and the other carries GG or GA, the channel is one-directional. The AA person's nose works fine; they can smell and evaluate the partner's scent. But the GG/GA partner receives no scent signal back. They are evaluating silence.

If both partners carry AA, the axillary olfactory channel is dead silent. Neither person is broadcasting. For roughly two billion people, this is the default condition.

This is not a criticism of MHC-based matching. It is a genuine confound. If the mechanism depends on scent, its effectiveness depends on whether scent is actually produced.

But MHC Still Matters — More Than Skin Deep

Smell is not the only channel, though. ABCC11 controls axillary (underarm) odour specifically. Saliva, breath, skin lipids, and genital secretions all carry molecular information through ABCC11-independent pathways.

Kissing is a good example. Wlodarski & Dunbar (2015) proposed it as an MHC testing mechanism. It brings two people close enough for non-axillary scent evaluation and exchanges saliva, which carries a rich molecular signature. If MHC information travels through multiple channels, losing one (axillary odour) does not necessarily eliminate the signal.

Three lines of evidence suggest that MHC compatibility matters beyond any single sensory channel:

1. Wu et al. (2018) studied 262 Asian American speed-daters, a population likely to have a high frequency of the ABCC11 A allele. Women preferred MHC-dissimilar partners, and the effect was "comparable to personality" in predicting second-date offers. This was measured through real dating behaviour, not t-shirt sniffing. If MHC-based attraction works only through axillary odour, this result should not exist in a population where many participants produce no underarm scent. Something else is carrying the signal.
2. Immune diversity in offspring: MHC dissimilarity between parents produces children with a broader range of HLA alleles and more robust pathogen resistance. This benefit has nothing to do with whether the parents can smell each other.
3. Reduced miscarriage risk: Ober et al. (1998) found that HLA-similar couples, those sharing more MHC alleles, had significantly higher rates of fetal loss. Too much HLA similarity spikes the risk of pregnancy failure, and that has nothing to do with how anyone smells.

Kromer et al. (2016) also found associations between MHC compatibility and sexual satisfaction in established couples, suggesting that MHC dissimilarity matters not just for initial attraction but for the long term.

What This Means at DNA Romance

This is exactly why DNA Romance includes rs17822931 in our analysis. Our DRom 2.0 trait prediction system already reads this SNP to predict earwax type (wet or dry), the same variant that controls axillary body odour. We know which users carry AA, GA, or GG.

Our DRom 1.0 MHC compatibility score uses 100 SNPs in the HLA region. If you have the smelly genotype (GG/GA), the scent-based "chemistry" prediction is backed by research, even if scientists are still arguing over the exact pathway. For users who carry the AA genotype, MHC compatibility still predicts immune diversity in offspring and reduced miscarriage risk. The Wu et al. (2018) speed-dating study is relevant here: real MHC-based attraction was observed in Asian Americans, a population where many participants likely carry the odourless genotype. Whatever channel they were using, it was not their armpits.

Could we incorporate rs17822931 into how we present compatibility scores? Yes. Knowing whether both partners produce body odour, only one does, or neither does would allow us to provide users with better context regarding what their MHC score signifies for their specific pairing.

We wrote this article because the relationship between MHC genes, body odour, and attraction is more complicated than "dissimilar MHC equals chemistry." We would rather explain the nuance than pretend it does not exist.

Related Reading

• What are Human Leukocyte Antigens (HLAs)?
• The DRom 1.0 Algorithm Predicts Romantic Chemistry
• Ten Ways Scent, Sense and DNA Influence Our Relationships
• The Genetic Blueprint of Love: Unraveling Romantic Chemistry