SU74 - COMMON GENETIC VARIANTS ASSOCIATED WITH PERSONALITY DIMENSIONS IN THE HEIDELBERG COHORT STUDY OF THE ELDERLY (HeiDE): AN UPDATE

Abstract

The HeiDE study is an ongoing longitudinal investigation that started in the 1990s and, at baseline, assessed an array of personality tests in 5,114 individuals. Principal components factor analysis was used to identify five latent personality dimensions (“The Heidelberg Five”), interpreted as emotional lability (ELAB), lack of behavioral control (LBCN), type-A-behavior (TYAB), locus of control over disease (LOCC), and psychoticism (PSYC). At follow-up, a subset of responding participants (n=2,734) were genotyped using Illumina PsychChip arrays. We conducted five initial GWAS, analyzing common genetic variants underlying the previously identified orthogonal personality dimensions with factor scores as phenotypes. For ELAB, we observed a locus that was genome-wide significant (rs79136259; p=8.2×10-9). Recently, genome-wide data from a second sample of the HeiDE study became available and we have now jointly analyzed both samples and combined results using meta-analysis (n=2,387 and n=881; post-QC). In both samples, we imputed common variants (MAF≥0.01) using the 1000 Genomes Phase 3 reference panel. Data were analyzed using PLINK 1.07 (http://zzz.bwh.harvard.edu/plink/) with sex, age and the first four ancestry principal components as covariates. Fixed-effects meta-analysis was conducted using METAL (http://genome.sph.umich.edu/wiki/METAL). SNP-based heritability estimates and genetic correlations were calculated using GCTA (http://cnsgenomics.com/software/gcta/index.html), jointly on all available individuals. The association of SNP rs79136259 with ELAB, discovered in the initial HeiDE sample, did not replicate in the second sample. Association strength in the fixed-effects meta-analysis of both samples decreased to a nominal level (p=1.3×10-4). For TYAB, an InDel variant on chromosome 8 (rs58535027, p=1.1×10-8) that was not significant in either sample alone reached genome-wide significance in the meta-analysis. No genome-wide significant associations were found in the meta-analyses for LBCN, LOCC or PSYC. SNP-based heritability estimates of the joint genotype data are: 28.8% (ELAB; p=0.014), 27.3% (LBCN; p=0.019), 8.4% (TYAB; p=0.262), 8.4% (LOCC; p=0.269) and 23.6% (PSYC; p=0.028). Genetic correlations between The Heidelberg Five will be presented at the meeting. The HeiDE cohort represents a unique opportunity to study the association of personality, genetics, and longitudinally defined phenotypes. Using an extended sample, we did not find evidence for the previously reported association of rs79136259 with ELAB but found evidence for a genetic variant influencing TYAB. Significant SNP-based heritability estimates of The Heidelberg Five demonstrate biological validity of some latent personality dimensions.