A Genome-Wide Association Analysis of Happiness: Consistent Genetic Effects Across the Lifespan and Across Genetic Ancestries in Multiple Cohorts

Abstract

We present a genome-wide association study of a general happiness measure in 118,851 participants from the UK Biobank. Using BOLT-LMM, we identify 3 significant loci with a heritability estimate of 0.8%. Linkage disequilibrium score regression was performed on the 'big five' personality traits finding significant associations with lower neuroticism and higher extraversion and conscientiousness. Using a novel approach, we construct LDpred-inf polygenic risk scores in the Adolescent Brain Cognitive Development (ABCD) cohort and the Add Health cohort. We detected nominally significant associations with several well-being measures in ABCD and significant correlations with a happiness measure in Add Health. Additionally, we tested for associations with several brain regions in a white British subsample of UK Biobank finding significant associations with several brain structure and integrity phenotypes. We demonstrated a genetic basis for general happiness level and brain structure that appears to remain consistent throughout the lifespan and across multiple ancestral backgrounds.

Response bias is genetically biased: Another argument for Kagan’s philippic against questionnaires in developmental psychology.

“…GWAS did teach developmental researchers some important lessons to overcome the replication crisis, with admirable features like striving for large consortia, preregistration, built-in replications and meta-analyses, and stringent corrections for multiple testing. But if GWAS rely on a few self-report questions to measure complex phenotypes, for example, “happiness” (Ward et al, 2022), a substantial genetic crud factor is almost guaranteed, and the resulting PGS for happiness might turn out to be in part an indicator of the acquiescence response bias. Large numbers of respondents and SNPs might not compensate for crudely measured phenotypes with a high risk of response bias.…”

“…Paradoxically, this is critically important for large genome wide association studies in search of polygenic scores as well. If GWAS rely on a few selfreport questions to measure for example 'happiness' (Ward et al, 2022) a substantial genetic crud factor is almost guaranteed, and the resulting PGS for happiness might turn out to be a valid indicator of the acquiescence response bias. Large numbers of respondents and SNPs might not compensate for crudely measured phenotypes with high risk of response bias.…”

Epidemiologic Questionnaire (EPI-Q) – a scalable, app-based health survey linked to electronic health record and genotype data

“…The study of life meaning, along with other psychosocial measures such as life purpose, life satisfaction, and happiness, has been increasing [ 15 , 16 ]. This includes studies that employ genetic analyses [ 17 , 18 ]. Our sample consisted of 2,433 participants, primarily of European ancestry (as inferred from available genotype data), who answered at least 1 question in the life meaning module.…”