Sanchez-Roige said the influence of genetics on coffee consumption was the first of two questions the researchers wanted to answer.
“The second thing is something that coffee lovers really want to learn,” Sanchez-Roige said. “Is drinking coffee good or bad? Is it associated with positive health outcomes or not?”
The answer is not definitive. The group’s genome-wide association study of 130,153 23andMe study participants in the US was compared with a similar database from the UK Biobank, which includes 334,649 Britons, revealing consistent positive genetic associations between coffee and harmful health outcomes such as obesity and substance apply. A positive genetic association is a link between a specific gene variant (genotype) and a specific condition (phenotype). A negative genetic association, on the other hand, is an apparent protective trait that discourages the development of a condition. The results become more complicated when it comes to psychiatric conditions.
“For example, look at the genetics of anxiety, or bipolar disorder and depression: in the 23andMe data set, they tend to be positively genetically correlated with the genetics of coffee consumption,” Thorpe said. “But in the UK Biobank, you see the opposite pattern, where they are negatively genetically correlated. That’s not what we expected.”
She added that there were other cases where the 23andMe kit did not match the UK Biobank, but the biggest discrepancies were for mental disorders.
“It’s common in this field to combine similar data sets to enhance research power. This information paints a pretty clear picture that combining these two data sets was not a really clever idea. And we ended up not doing that,” Thorpe said. She explained that combining databases can mask effects, leading researchers to draw incorrect conclusions — or even cancel each other out.
Sanchez-Roige says the researchers have some ideas about how the differences in results came about. For starters, there was an apples-to-oranges aspect to the surveys. For example, the 23andMe survey asked, “How many 5-ounce (cup-sized) servings of caffeinated coffee do you consume each day?” Compare that to the UK Biobank question, “How many cups of coffee do you drink each day? (Including decaf)”
Apart from portion size and the split between caffeinated and decaffeinated coffee, the study did not take into account different ways of serving coffee. “We know that in the UK people generally prefer instant coffee, whereas in the US people prefer ground coffee,” Thorpe said.
“And then there are frappuccinos,” Sanchez-Roige added, citing the American trend of drinking coffee with lots of sugary additives. Palmer mentioned other caffeinated beverages, and particularly in the context of the UK Biobank, tea, none of which were included in the GWAS, which looked only at coffee. Palmer added that the GWAS study shows that the association between genotype and phenotype is more distinct than the association between coffee and tea.
“Genetics influences a lot of things. Like how statuesque you can be,” he said. “And that kind of thing would probably be very similar whether you live in the U.S. or the U.K. But coffee is a decision that people make.”
Sanchez-Roige noted that coffee comes in a variety of forms, from instant to frappuccino, and is consumed within cultural norms that vary from place to place. A person with a given genotype might have a completely different phenotype living in the UK than in the US.
“And that’s what the data are telling us,” she said. “Because unlike height, which your behavior has little to do with, your behavior and the choices you make in your environment play out in different ways. So the interaction between genotype and environment complicates the picture.”
The collaborators stressed the need for further research to elucidate the links between genetics and environment, with particular attention paid not only to coffee/caffeine consumption but also to other issues related to psychoactive substance apply.
Hayley HA Thorpe is on the Neuropsychopharmacology Special Projects Team. Pierre Fontanillas, the 23andMe Research Team, and Sarah L. Elson are employees of 23andMe, Inc., and Pierre Fontanillas and Sarah L. Elson hold stock or stock options in 23andMe. Abraham Palmer serves on the advisory board of Bright Genomics, for which he receives stock options. The remaining authors have nothing to disclose.