Diverse phenotypic measurements of wellbeing: Heritability, temporal stability and the variance explained by polygenic scores.

Wellbeing, a key aspect of mental health, is moderately heritable with varying estimates reported from independent studies employing a variety of instruments. Recent genome-wide association studies (GWAS) have enabled the construction of polygenic scores (PGS) for wellbeing, providing the opportunity for direct comparisons of the variance explained by PGS for different instruments commonly employed in the field. Nine wellbeing measurements (multi-item and single-item), two personality domains (NEO-FFI neuroticism and extraversion), plus the depression domain of the DASS-42 were drawn from a larger self-report battery applied to the TWIN-E study-an Australian longitudinal twin cohort (N = 1660). Heritability was estimated using univariate twin modeling and 12-month test-retest reliability was estimated using intra-class correlation. PGS were constructed using wellbeing GWAS summary-statistics from Baselmans et al. (Nat Genet. 2019), and the variance explained estimated using linear models. Last, a GWAS was performed using COMPAS-W, a quantitative composite wellbeing measure, to explore its utility in genomic studies. Heritability estimates ranged from 23% to 47% across instruments, and multi-item measures showed higher heritability and test-retest reliability than single-item measures. The variance explained by PGS was ~0.5% to 1.5%, with considerable variation between measures, and within each measure over 12 months. Five loci with suggestive association (p < 1 × 10 ) were identified from this initial COMPAS-W wellbeing GWAS. This work highlights the variability across measures currently employed in wellbeing research, with multi-item and composite measures favored over single-item measures. While wellbeing PGS are useful in a research setting, they explain little of the phenotypic variance, highlighting gaps for improved gene discovery.