Multivariate Analyses and Genomic Imaging of the Well-being Spectrum

Several phenotypes related to well-being (e.g., life satisfaction, positive affect, neuroticism, and depressive symptoms), are genetically highly correlated (| rg | > .75) suggesting one underlying genetic liability between... [ view full abstract ]

Several phenotypes related to well-being (e.g., life satisfaction, positive affect, neuroticism, and depressive symptoms), are genetically highly correlated (| r_g | > .75) suggesting one underlying genetic liability between these traits. We performed two multivariate genome-wide meta-analyses (multivariate GWAMA’s) of these four phenotypes to increase the power to identify associated genetic variants. The dependence between effect sizes induced by sample overlap, was estimated from GWAS summary statistics obtained from univariate GWAMA using LD score regression. Both multivariate GWAMAs aggregated the effect of a single SNP across multiple traits based on prior knowledge of the genetic correlation between these traits. In a similar fashion, we proceeded to aggregate the effect across multiple SNPs based on prior knowledge that these SNPs collectively influence the expression level of a gene transcript or the methylation level at a CpG site. Biological annotation revealed significant enrichment of SNP effects in regions of the genome characterized by active histone mark in several cortical and subcortical brain tissues. In order to accurately pinpoint brain regions in which genes relevant for well-being are actively expressed we computed differential brain expression. To interpret the results we mapped the results to a coordinate system common to multiple other neuroimaging techniques.