From genes to beans: genetic correlations with habitual coffee intake reveal geneculture interactions across uk- and us-based cohorts

Caffeinated coffee is one of the most consumed beverages; its habitual consumption is heritable but is also influenced by environmental factors such as cultural norms surrounding the consumption of caffeinated beverages. While some epidemiological studies suggest that habitual coffee intake may have health benefits, others indicate that it increases the risk of certain cancers, metabolic disorders, and substance use. Previous genetic studies have investigated the biological basis of habitual coffee intake and its relationship with select health outcomes, but no studies have examined if these associations are subject to cultural or cohort influences. We conducted a genome-wide association study (GWAS) of coffee intake in 23 and Me research participants of European ancestry (N=130,135). Secondary analyses were performed using Multimarker Analysis of Genomic Annotation (MAGMA), Hi-C-coupled-MAGMA, and Summary-MultiXcan to explore associations between genes, transcriptomics, and tissues with coffee intake. To investigate genetic associations between health and coffee intake in two culturally unique cohorts, summary statistics of coffee intake from 23 and Me and the UK Biobank (UKB; N=334,659) were used to calculate genetic correlations with 316 health, psychiatric, and anthropomorphic traits using Linkage Disequilibrium Score Regression (LDSC). Polygenic scores (PGS) of coffee intake were calculated in a hospital-based cohort (BioVU; N=72,225) to conduct phenome- and lab-wide association analyses across 2,137 medical phenotypes and biomarkers, respectively. We identified seven loci, all of which replicated associations from prior coffee GWAS. Gene-based analyses supported transcriptional regulation of coffee intake genes within the nervous and digestive systems. Comparisons across top loci (p < 5E-08) between 23 and Me and UKB cohorts revealed that only 52% of the 29 significant variants shared the same direction of effect on coffee intake. Comparisons across 23 and Me and UKB summary statistics of habitual coffee intake revealed predominantly discrepant or inconsistent genetic associations - 83% of correlations with 23 and Me summary statistics indicated greater health risk by coffee intake versus 54% in the UKB. Elevated risk for substance use and obesity were consistent in both cohorts. We expanded previous GWAS of coffee intake to 23 and Me research participants based in the US, replicating previous associations and uncovering biological relationships with tissues and gene sets important for digestive and nervous system function. Aside from consistent positive associations with substance use and obesity, we observed significant differences in genetic associations between two large cohorts of European ancestry, suggesting that genetic predisposition for coffee intake and associations with other health traits may be influenced by cultural differences. Our study provides a cautionary perspective on combining large cohort datasets from populations of similar ancestries that are exposed to different environments.