Genetic influences on coffee consumption were the first of two questions the researchers wanted to answer, Sanchez-Roige said.
“The other topic that coffee lovers really want to know about is,” Sanchez-Roige said. “Is drinking coffee good or bad? Is it associated with positive health outcomes or not?”
The answer is not certain. The group's genome-wide association study of 130,153 U.S.-based 23andMe research participants, compared with the UK Biobank database of 334,649 Britons, revealed consistent positive genetic links between coffee and harmful health outcomes such as obesity and drug use. A positive genetic linkage is an association between a specific gene variant (genotype) and a specific condition (phenotype). In contrast, a negative genetic linkage is an apparent protective trait that discourages the development of a condition. The findings become more complex when it comes to psychiatric conditions.
“For example, look at the genetics of anxiety or bipolar and depression: in the 23andMe data set, they are positively genetically correlated with coffee intake genetics,” Thorpe said. “But then, in the UK Biobank, you see the opposite pattern, where they are negatively genetically correlated. This is not what we expected.”
He said there were other instances in which the 23andMe group did not match the UK Biobank, but the biggest disagreements were in psychiatric conditions.
“It's common in this field to combine similar datasets to increase study power. This information paints a pretty clear picture that combining these two datasets was really not a wise idea. And we didn't do that,” Thorpe said. He explained that combining databases can hide effects, leading researchers to the wrong conclusions — or even canceling out each other.
Sanchez-Roige says the researchers have some insight into what caused the differences in the results. First, there was an apples-and-oranges aspect to the surveys. For example, the 23andMe survey asked, “How many 5-ounce (cup-size) cups of caffeinated coffee do you drink per day?” Compare that to the UK Biobank's “How many cups of coffee do you drink per day? (include decaf coffee).”
Aside from serving size and the caffeinated/decaf split, the surveys made no adjustments for different ways coffee is served. “We know that in the U.K., they generally prefer instant coffee, whereas in the U.S., ground coffee is more preferred,” Thorpe said.
“And then there's the Frappuccino,” Sanchez-Roig said, citing the American trend of drinking coffee laden with sugary additives. Palmer mentioned tea in the context of other caffeinated beverages and specifically the UK Biobank, none of which were included in the GWAS, which addressed only coffee. Palmer said the GWAS shows that the relationship between genotype and phenotype is more distinct than the relationship between coffee and tea.
“Genetics affects many things. For example, it affects how tall you might be,” he said. “And things like that will probably turn out very similarly whether you live in the US or the UK. But coffee is a decision that people make.”
Coffee comes in many forms, from instant to Frappuccino, and is consumed amid different cultural norms from place to place, Sanchez-Roige explained. A person with a given genotype may have a very different phenotype if they live in the U.K. versus the U.S.
“And that's what the data are telling us,” he said. “Because unlike height, where your behavior really has nothing to do with it, your behavior and the choices you're making in your environment come into play in different ways. So the interaction between genotype and environment complicates the picture.”
The colleagues stressed the need for further investigation to understand the relationship between genetics and environment, focusing not only on coffee/caffeine intake but also on other substance use issues.
In addition to the researchers mentioned above, co-authors of the UC San Diego paper are: Benjamin K. Pham, John J. Meredith, Mariella V. Jennings, Natasia S. Courchesne-Kraak and Sevim B. Bianchi, all from the Department of Psychiatry. Other co-authors are Pierre Fontanillas of 23andMe, Inc.; Laura Villar-Ribo of Barcelona Autónoma de Universitat, Spain; Julian Mutz of King's College London, UK; Sarah L. Elson and Gibran Y. Khokhar of the University of Guelph, Canada; Abdel Abdellaoui of the University of Amsterdam, the Netherlands; Lee K. Davis of Vanderbilt University Medical Center; and the 23andMe research team.
Mariella V. Jennings, Sevim B. Bianchi, and Sandra Sanchez-Roige are supported by funds from the California Tobacco-Related Diseases Research Program (TRDRP; grant numbers T29KT0526 and T32IR5226). Sevim B. Bianchi and Abraham Palmer were also supported by P50DA037844. BKP, Julian Mutz, and Sandra Sanchez-Roige are supported by NIH/NIDA DP1DA054394. Hayley H. A. Thorpe is funded through a Natural Sciences and Engineering Research Council PGS-D scholarship and a Canadian Institutes of Health Research (CIHR) fellowship. Gibran Y. Khokhar is supported by a CIHR Canada Research Chair in Translational Neuropsychopharmacology. Lee K. Davis is supported by R01 MH113362. Natassia S. Courchesne-Crack is funded through an Interdisciplinary Research Fellowship in NeuroAIDS (grant number R25MH081482). Julian Mutz is funded by the National Institute for Health and Care Research (NIHR) Maudsley Biomedical Research Centre, South London and Maudsley NHS Foundation Trust and King's College London.
The datasets used for the PheWAS and LabWAS analyses described were obtained from Vanderbilt University Medical Center's BioVU, which is supported by multiple sources: institutional funding, private agencies, and federal grants. These include the NIH-funded Shared Instrumentation grant S10RR025141; and CTSA grants UL1TR002243, UL1TR000445, and UL1RR024975. Genomic data is also supported by investigator-led projects, including U01HG004798, R01NS032830, RC2GM092618, P50GM115305, U01HG006378, U19HL065962, R01HD074711; and additional funding sources listed at https://victr.vumc.org/biovu-funding/. PheWAS and LabWAS analyses used CTSA (SD, Vanderbilt resources). This project was supported by the National Center for Research Resources, Grant UL1 RR024975-01, and is now in the National Center for Advancing Translational Sciences, Grant UL1 TR000445-06.
Hayley H. A. 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 is on the Scientific Advisory Board of Vivid Genomics, for which he receives stock options. The remaining authors have nothing to disclose.