Skip to content

When you choose to publish with PLOS, your research makes an impact. Make your work accessible to all, without restrictions, and accelerate scientific discovery with options like preprints and published peer review that make your work more Open.


The Coffee-Genes Story Brews Discontent on Thanksgiving

A few weeks ago a report claiming a love for coffee is “in the genes” fleetingly flashed across screens. I thought I’d let the grounds settle and then take a closer look, fitting with my writing about food every Thanksgiving (pumpkins last year, yo-yo dieting in 2016, and turkeys in 2015).

I keep a mental list of dumb things linked to genes.


“Get the giggles easily? Blame your genes: Genetic variant enhances how people react to funny – and sad – situations,” reported the Daily Mail, echoing a paper published in the journal Emotion. It refers to the long and short variants (alleles) of the 5-HTTLPR gene, which encodes the serotonin transporter. Variants of that gene are connected to whether participants laughed at Gary Larsen cartoons or not.

An article in The Telegraph“It’s All In Your Genes: How DNA Explains Your Sense of Humor, Sleep Patterns, and Phobias,” explored arachnophobia, speculating about a genetic advantage handed down from an ancient aversion to spiders. The article links to another Telegraph article about a genetic connection to shopping style.

Soon I found myself sucked into a quagmire of British newspapers parroting each other, but not the technical reports. They evoke “in your genes” to cover quite a technological territory, including twin studies, comparing behaviors to having certain gene variants, to genome-wide association studies, to spinning stories that just make sense, like a fear of scorpions or not eating toxic bugs.

Twin studies are classic, low-tech but intriguing. Basically, a trait shared more often among identical twins than among fraternal twins indicates an inherited component. But the characteristics can veer into vagueness, such as a twin study that examined the trait of “how to vote.” A paper from the Journal of Politics that inspired headlines was more specific. The researchers equated having a particular variant of the dopamine D2 receptor gene with belonging to a political party. Has the current situation in the U.S. overridden a DNA-encoded preference for identifying as Republican or Democrat?

Wikipedia, I discovered, lists “Genopolitics” as “the study of the genetic basis of political behavior and attitudes,” made up of behavioral geneticists, politicians, and psychologists. Not molecular biologists who might explain how DNA base sequences that encode proteins could affect what we find funny, whether we power shop or linger, and whether we continue to vote according to our political party even when an elected leader doesn’t behave quite as we expected.

Tracking Genetic Influence on Coffee Consumption

My unofficial, nonscientific observations suggest that people either love or hate coffee.

The aversion is oft-attributed to a time long ago when the ability to taste and spit out bitter substances was adaptive – poisonous plants and decaying matter. In this nice explanation, Genetics of Taste and Smell: Poisons and Pleasures, Danielle Renee Reed and Antti Knaapila of the Monell Chemical Senses Center in Philadelphia describe our individual “chemosensory landscapes” (the chemical senses being smell and taste): “Genetic studies in humans and experimental animals strongly suggest that the liking of sugar and fat is influenced by genotype; likewise, the abilities to detect bitterness and the malodors of rotting food are highly variable among individuals.”

Taste according to genotype is the basis of the new study. But first came an investigation from 2016. That paper, in Scientific Reports, associated variants of a gene with “habitual coffee consumption” in two populations in Italy, assessed with the metric “number of cups per day.”

Stop. Right. There. A “cup” of coffee is hardly a unit measure! How many times have I stopped someone from making coffee using 2 tablespoons to 8 ounces of water, because a cup of coffee is actually 2 tablespoons of grinds to 6 ounces (ok, 6.5 if you like it weak)? If the participants in these studies do not consistently drink the same number of ounces, well, what does that do to the statistics? Researchers who are meticulous in describing DNA should at least standardize a cup of coffee. Tall, grande, and vente would be more acceptable because at least then we can calculate the ounces.

Anyway, the 2016 report revealed a gene called PDSS2 (prenyl diphosphate synthease, subunit 2) that influences fondness for coffee, somehow, because the encoded protein regulates caffeine metabolism. I looked it up in the geneticist’s bible Online Mendelian Inheritance in Man and PDSS2 turns out to be part of coenzyme Q10. That makes sense, given CoQ10’s energy-boosting rep.

The just-published study, also in Scientific Reports but from a different research team, from Northwestern Medicine and QIMR Berghofer Medical Research Institute in Australia, didn’t look at caffeine metabolism or its jolt, but ability to sense (taste) bitterness. They compared three parts of the genome that harbor single-base markers (SNPs) that in twin studies were associated with “perceived intensity” of three bitter chemicals: caffeine of course, quinine (the malaria drug from tree bark), and PROP (propylthiouracil; a synthetic chemical that elicits a taste similar to that of broccoli).

The marker for PROP is on chromosome 7. Two markers in a cluster of taste genes on chromosome 12 tracked with quinine and caffeine. The study also assessed fondness for tea and alcohol, but coffee is my drinking drug of choice.

Past studies on twins had implicated the influence of these three chemicals on bitter taste perception as being heritable – heritability is perhaps the second most misused genetic term in the media, following genetic code. Even scientists who aren’t geneticists misuse it.

Heritability refers to the degree that genes contribute to the variation in a trait in a population – NOT to the degree to which a trait is inherited. Anyway, variants of these three genes seem to be connected to whether a person is like me and finds coffee bitter and likes it, or my husband Larry, repelled by the bitterness.

The researchers considered the three bitter-sensing gene variants among the 438,870 participants in the UK Biobank, asking them to self-report their coffee consumption. The exact wording: “How many cups of coffee do you drink each day? (including decaffeinated coffee).”

Wouldn’t the caffeine/no caffeine make a difference? And why calculate confidence intervals and standard deviation to presumably assess how much their intake varied from day to day, but not quantify the volume of the cups??? What am I missing? Perhaps in the UK coffee cup volume never varies and is common knowledge. But then again, they have Starbucks. So I decided to do some limited math.

Average coffee consumption, according to the study, was 2.13 cups per day with standard deviation of 2.1. Applying the Starbucks conversion formula of tall (=12 ounces), grande (16 ounces), and venti (20 ounces) and multiplying by 2.13 average number of daily cups gives the possibilities of 25.56, 34.08, and 42.6 ounces, respectively. And if a person were to have, say, a venti in the morning and a quick tall later on, the values would fall in between the extremes, approaching a lovely continuous distribution, like a classic polygenic trait.

Detail aside, plotting reported increasing coffee consumption against the three markers revealed a correlation only for the marker (a chromosome 12 taste SNP) for caffeine, and not the markers for quinine or PROP. But the effect was hardly dramatic enough to justify the wide reporting.


At least at Starbucks you know how many ounces are in a cup.

Drilling down into the details of the report, rather than just cutting and pasting from a news release, reveals an increase in covfefe consumption range from .02 to .15 UK cups per day among the genetically enhanced like myself. I didn’t need a study to tell me that folks tend to like coffee or tea, but not both, with the exception of my eclectic best friend Wendy.

An increase of circa 10% of an undefined cup made headlines? It did. Here and here and here, for example.

This is why I like to wait a few days or weeks after a news release comes out and then go back to the original paper.

The study had a few flaws. Only 5% of the Brits in the database took part, which might have introduced bias in favor of coffee connoisseurs. Participants were mostly or all European whites. More egregious, the investigation apparently didn’t distinguish among Dunkin’ Donuts, Starbucks, and whatever chains are endemic to feed the addiction in the UK. Brew strength is critical to any coffee lover and we all know there’s quite a distinction between Starbucks and Dunkin’.

The findings are considered “causal,” the researchers contend, rather than merely an association, because of a plausible mechanism, presumably the link to a variant of a bitter taste gene. OK.

The next step is to hypothesize a seeming conundrum: if a drink tastes more bitter, why do some of us like it? Here’s where we could perhaps use one of those Genopoliticians, but study co-author Marilyn Cornelis, assistant professor of preventive medicine at Northwestern University Feinberg School of Medicine, explained. “You’d expect that people who are particularly sensitive to the bitter taste of caffeine would drink less coffee. The opposite results of our study suggest coffee consumers acquire a taste or an ability to detect caffeine due to the learned positive reinforcement (i.e. stimulation) elicited by caffeine.” And Starbucks certainly provides a nice environment. That’s the beauty of the scientific method. If findings are unexpected, come up with a new hypothesis, then test it.

To that end, the investigators conclude that “further research is required to validate the causal effects detected in the present study.”

Um, why????

Some people just like bitter-tasting foods.

Happy Thanksgiving!







  1. Ricki, great article, as always. I agree that any (peer-reviewed) scientific study covering something as detailed and complicated as human genetics, should start out of the gate by standardizing volume of substance ingested by a standard, unambiguous (preferably SI) unit of measure. If they don’t know well enough to do that, then the confidence factor on the genetics bits is questionable.

    But the quantities of coffee you cited for small, medium, and large “cups”, seemed large – quite large, to me. The link referenced to the Starbucks equivalencies says that tall, grande, and venti Starbucks cups are 12, 16, and 20 fluid (avoirdupoois) ounces, respectively, not 24, 32, and 40 ounces as stated here. So did I miss a multiplication factor of 2 somewhere ?

Leave a Reply

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

Add your ORCID here. (e.g. 0000-0002-7299-680X)

Back to top