Personality’s DNA emerges across 1,260 variants — none decides who we are

The question is not whether there is a “personality gene,” but how thousands of small genetic differences combine with experience to produce relatively stable patterns of feeling, thought, and action. An international consortium has expanded that inquiry: its analysis identified 1,260 lead variants associated with the Big Five traits—extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience. Of those, 824 lay in regions not previously linked to the corresponding trait.

Led by Ted Schwaba and published in Nature on September 2, the study combined 46 cohorts. Depending on the trait, genome-wide association analyses included 611,037 to 1.14 million people whose genomes were classified as European-like or African-like. Participants provided DNA and completed personality instruments; the team compared roughly 10 million single-letter DNA variants with their scores. The statistical threshold was P below 5 × 10⁻⁸, appropriate when millions of comparisons are made at once.

The size of each effect rules out a deterministic reading. Among lead variants for extraversion, the corrected median effect was 0.009 standard deviations per allele—roughly the difference between the 50th and 50.35th percentiles. At the population level, common variants jointly explained from 4.8% of measured variation in agreeableness, with a 0.2-percentage-point standard error, to 9.3% in extraversion, with a 0.3-point standard error. After adjusting for typical questionnaire reliability, the estimated range rose to 9.3%–13.3%. The remainder includes environment, rare variants, interactions, and measurement error.

Robustness was tested rather than assumed. When the European-like cohorts were split in half, genetic signals had a mean correlation of 0.94. Similarity remained high, though not perfect, across four Western geographic clusters, with a mean genetic correlation of 0.86 and mean standard error of 0.15; across age groups, 0.80 and 0.18; and across five measurement instruments, 0.85 and 0.07. Agreeableness was the least consistent trait across cohorts, suggesting that cultural or measurement differences may matter more for it.

To better separate direct inheritance from conditions shared within families, the researchers repeated part of the analysis in up to 50,725 relatives. Within-family results resembled those from the broader population and changed less than is common in studies of education or income. This reduces some forms of confounding but does not remove them: confidence intervals were wider, a few traits diverged slightly, and even a direct genetic effect can operate through experiences a person encounters or seeks.

The study also found genetic overlap between personality traits and outcomes involving health, education, work, habits, and residential mobility. Some Mendelian-randomization analyses suggested potentially causal relationships, but the authors call for independent confirmation. The sample remains dominated by people with European-like ancestry and Western settings, leaving cultural and ancestral generalization incomplete. The genuine advance is methodological: personality looks distributed and probabilistic, useful for studying population pathways rather than for sentencing an individual biography.