Background Hedonic (e.g., happiness) and eudaimonic (e.g., meaning in life) well-being are negatively related to depressive symptoms. Genetic variants play a role in this association, reflected in substantial genetic correlations. We investigated the (genetic) overlap and differences between well-being and depressive symptoms. Methods We used results of Genome-Wide Association studies (GWAS) and applied GWAS-by-subtraction in the UK Biobank sample. Analyses were pre-registered. Results Subtracting GWAS summary statistics of depressive symptoms from those of happiness and meaning in life, we obtained GWASs of respectively "pure" happiness (neffective= 216,497) and "pure" meaning (neffective=102,300). For both, we identified one genome-wide significant SNP (rs1078141 and rs79520962, respectively). After the subtraction, SNP heritability reduced from 6.3% to 3.3% for pure happiness and from 6.2% to 4.2% for pure meaning. The genetic correlation between the well-being measures reduced from .78 to .65, indicating that only a part of the genetic overlap between happiness and meaning in life is due to overlap with depressive symptoms. Pure happiness and pure meaning became genetically unrelated to traits strongly associated with depressive symptoms, including tiredness, loneliness, and psychiatric disorders. For several other traits, including ADHD, income, educational attainment, smoking, and drinking alcohol, the genetic correlations of well-being versus pure well-being changed substantially. Conclusions GWAS-by-subtraction allowed us to investigate the genetic variance of well-being unrelated to depressive symptoms. Genetic correlations with different traits led to new insights about this unique part of well-being. The findings can have implications for interventions to increase well-being and/or decrease depressive symptoms.