Why Does Cilantro Taste Like Soap? The Gene Is a Small Part

If cilantro has ever made your salsa taste like soap, you have company. When 23andMe asked its customers “Does fresh cilantro taste like soap to you?”, 12.4% of the 18,495 who answered yes or no said yes.

The same 2012 study, in the journal Flavour, found a variant linked to that answer near OR6A2, a smell-receptor gene now often called the cilantro soap gene. The authors’ own estimate of how much it explains is far smaller than the nickname suggests.

Why does cilantro taste like soap?

Cilantro tastes like soap to some people because most of the key aroma compounds in its leaves are aldehydes, and aroma chemists describe one group of them, mostly (E)-2-decenal and (E)-2-dodecenal, as soapy, fatty or pungent. Not everyone notices that note, and a gene variant in a cluster of smell-receptor genes accounts for a small, measurable share of the difference.

The soap is in the leaf’s aldehydes

The 23andMe paper describes the split: many people love cilantro “while others claim that it tastes or smells foul, often like soap or dirt.” The authors put the soapy aroma down largely to aldehydes, citing aroma studies that describe the (E)-2-alkenals as “soapy, fatty, ‘like cilantro,’ or pungent.” Aroma chemists pin other food smells on particular compounds the same way: in a 2024 study of table grapes, the candyfloss note of Cotton Candy grapes went with high levels of furaneol.

Two other aldehydes the paper lists in cilantro, decanal and dodecanal, have industrial uses. PubChem, which notes decanal’s use in fragrances and synthetic citrus oils, compiles EPA chemical-use reports that list both decanal and dodecanal under soap, cleaning compound and toilet preparation manufacturing.

The authors left open whether the problem is smell or taste: “It is suspected, although not proven, that cilantro dislike is largely driven by the odor rather than the taste.”

One variant held up, and it explains little

Nicholas Eriksson and seven 23andMe colleagues scanned the genomes of 14,604 customers of European ancestry who answered the soap question, then checked the leading hits in 11,851 who answered only a liking question.

The one that held up, rs72921001, sits on chromosome 11 within a cluster of eight smell-receptor genes. The authors called OR6A2 “the most promising candidate,” because its receptor, much studied in its rat form, responds to aldehydes, including (E)-2-decenal. They named a neighbor, OR10A2, a reasonable candidate too.

The variant’s C form goes with “both detecting a soapy smell and disliking cilantro,” and in Europeans it is the more common; each copy of the other form, A, was linked to odds of tasting soap about a fifth lower.

Despite “the strength of the association,” the authors wrote, the variant “explains only about 0.5% of the variance in perceiving that cilantro tastes soapy.”

Their estimate of how much all the common variants their chips measured could explain was low too, under 9%, a figure they could not statistically separate from zero and treated as a lower bound.

A single gene makes a tidy answer, as tryptophan does for the turkey nap. The paper’s own conclusion is narrower: cilantro dislike “may stem from genetic variants in olfactory receptors.”

The 3 to 21% figure measured dislike, not soap

The range quoted most often, 3 to 21%, comes from a second 2012 paper in Flavour, by Lilli Mauer and Ahmed El-Sohemy of the University of Toronto. They asked 1,381 young adults aged 20 to 29, recruited on campus, to rate cilantro on a 9-point scale from “dislike extremely” to “like extremely”, or to say they had not tried it or would not, and counted ratings of 1 to 4 as dislike.

The percentages cover only the 1,108 who had tried it; 27% of the East Asian participants and 16% of the Caucasian participants never had. The two surveys rank the groups differently:

Group	Disliked it, of those who had tried it (Toronto)	Said it tastes like soap (23andMe)
East Asian	21% (81 of 390)	8.4% of 463
European or Caucasian	17% (85 of 484)	13.0% of 15,186
South Asian	7% (11 of 157)	3.9% of 335
Hispanic or Latino	4% (1 of 26)	8.7% of 898

The people differ too, campus volunteers against genetics-company customers, and neither set describes a whole ethnic group.

The 23andMe team found “the differences in allele frequency across populations do not explain the differences in soapy-taste detection.” Mauer and El-Sohemy judged differences in flavor perception, “possibly attributable to genetic differences,” the likely cause, but said they “cannot rule out” exposure, since cilantro “is most popular” in the cuisines of the groups with the fewest dislikers. The 23andMe authors allowed for experience too: “many people might initially dislike cilantro yet later come to appreciate it.”

A twin study put the genetic share higher

A twin study from the Monell Chemical Senses Center, published in Chemical Senses in 2012, tested twins at the Twins Days Festival in Twinsburg, Ohio. It put the heritability of how pleasant people found cilantro’s smell at 0.38, and of its flavor at 0.52.

Those figures assume, as the paper flags, that a shared upbringing adds nothing, and the ratings included only 15 pairs of non-identical twins. The 23andMe authors noted the gap: the studies measured different things, or rare variants their chips missed may carry more of the effect.

Testing 44 variants in taste and smell genes, the Monell team linked cilantro ratings to variants in three genes: TAS2R50, a bitter-taste receptor; GNAT3, a signaling component shared by bitter and sweet taste; and TRPA1, a receptor for pungent chemicals. Those links were weak (the team ran 1,188 tests), and when the 23andMe team checked the same three variants against its soap question, it was “unable to replicate” any of them.

Bugs are the other complaint, and an old one

Mauer and El-Sohemy sum up the descriptions: people who like cilantro may call it “fresh, fragrant or citrusy,” while those who dislike it “report that it tastes like soap, mold, dirt or bugs, among other descriptors.”

The bugs have a chemical match. (E)-2-decenal “is a major component of the alarm pheromone” of the brown marmorated stink bug, a 2020 study in Frontiers in Physiology notes. Mauer and El-Sohemy report the suggestion that it is one of the compounds people find unpleasant in cilantro, “although this has not yet been tested.” Honey bees also use an alarm scent to rally the nest.

John Gerard’s Herball, printed in London in 1597, opens its chapter on coriander, the plant whose leaves are cilantro, by calling the common kind “a very stinking herbe, smelling like the stinking worme called in Latine Cimex”, cimex being Latin for bedbug. Gerard condemned “the greene and stinking leaues” and praised “the drie and pleasant well sauoring seede.”

The plant’s name may carry the same complaint. Etymonline traces coriander through Latin to Greek koriannon, “often said by botanists to be related to koris ‘bedbug’ from the bad smell of the unripe fruit.”

Gerard faulted the green leaf in 1597, and 23andMe’s liking question still specified “fresh (not dried) cilantro” four centuries later. The aldehydes are in that fresh leaf for whoever eats it; the split lies in the eaters, and most of it has not been traced to any gene. Strictly FYI.