“Coffee, I Must Have Coffee…”

Whether you reach for a Starbuck’s Venti – the “Big Gulp” of coffee servings – or a wee size cup of Joe in the morning, or none at all, might be influenced by your genes.

Two recently published studies found an association between coffee and caffeine consumption and SNP variants near the same two genes, CYP1A1 and AHR. One study looked at coffee consumption specifically; the other looked at all caffeine consumption including from tea, soda and chocolate, but coffee was by far the main source of caffeine for the participants.

Scientists from deCode Genetics in Iceland studied more than 10,000 coffee drinkers in European descent mainly in Europe. Their results published in  Human Molecular Genetics in February pinpointed two SNPs near the CYP1A1 and AHR genes. The SNP rs2472297 near CYP1A1 had a fairly consistent effect — people with the T version were prone to drink about a quarter cup more of coffee a day, according to the study.

(23andMe customers can look up their data for rs2472297 and rs4410790 using the Browse Raw Data feature. 23andMe also looks at a variant that affects how quickly a person  metabolizes caffeine. )

The other study, which was published in PLoS Biology and included more than 47,000 individuals with European ancestry living in the U.S., made similar findings, noting that people with two copies of a T at rs4410790 near the AHR gene drank   about a third of a cup of coffee more each day than those without any copies.“Now, for the first time, we know specific genes that influence the amount of caffeine that individuals consume,” Dr. Neil Caporaso of the National Cancer Institute told the BBC. Caporaso, the co-author of study in the US, noted that while researchers know about associations between hundreds of genes and specific medical conditions, how those variants relate to dietary consumption is less well known.Beyond that, both studies offer intriguing hints about how these two genes might interact, but more research is needed. In the meantime how people relate to their morning fix of java is well known, perhaps best summed up by Johann Sebastian Bach in his “Coffee Cantata:”“How sweet the coffee tastes, more delicious than a thousand kisses . . . . Coffee, I must have coffee . . . . If I don’t get my three cups of coffee each day, I’ll shrivel up like a piece of roast goat.”
  • David Schlesinger

    Is 1/4 or 1/3 a cup of coffee really significant? And it doesn’t appear that they controlled for the type/brand of caffeinated beverage. The quantity of caffeine in a standard 8oz cup of coffee can vary greatly. Is espresso considered the same as coffee? Are they drinking Coke or Jolt?

    This is exactly the kind of pop culture nonsense science that ends up on CNN and Fox News, reported as fact. It’s a shame because there’s a real need for scientists to educate the public, not confuse them.

    This study can be added to the long list of ridiculous GWAS, including the one linking promiscuity to variants in DRD4.

    • Hi David,

      Thanks for your comment.

      In the deCODE study, the authors limited the analysis to coffee only; in the PLoS study they looked at coffee (they assumed 137mg of caffeine per cup), tea, soda, and “other” in addition to an analysis restricted just to coffee. Differences in caffeine content per cup of coffee would presumably average out with a large sample. While there is always the possibility of an uncontrolled variable exerting an influence on the results, the use of multiple large cohorts (and lack of heterogeneity between them) does give some confidence that this wasn’t the case.

      In terms of significance, it depends on what you mean — the effect was very statistically significant (p-values < 10^-10, and replicated independently), but whether it constitutes a meaningful difference is certainly a matter of interpretation. As is the case with GWAS in general, regardless of whether these findings are useful in everyday life, the associations themselves often provide clues about the underlying biological mechanisms for different health conditions and traits.

  • my both parents love to drink coffee and drink it a lot, and I dont like it at all. I am waiting for analysis results from 23andme soon. Will I be able to find on 23andme website whether I have that coffee gene or not?

    • Hi CSB,

      Your 23andMe health reports will include a report on caffeine metabolism and a report on coffee consumption using the markers described in this blog post. Of course, these variants only have a small effect on coffee consumption (about 1/4 cup more or less, on average), but it’s certainly interesting to think about. Welcome to the service!

  • Hi Scott,

    I personally love coffee and drink a lot of it. This study is very revealing yet does it also explain how, we, who drink excessive amounts of coffee can treat ourselves of this habit!?

    In my case I’d love to drink a lot less coffee than I am currently drinking. BTW, I don’t even drink Coke and the likes that contain caffeine but without coffee I will not survive.