Scientists Just Found 824 New 'Personality Genes' — But the Real Shocker Is What They Ruled Out

By the numbers, this is the study personality genetics researchers have been waiting a decade for: a consortium of 46 cohorts, more than 1.1 million participants at peak sample size, and the discovery of 1,260 genetic variants tied to the Big Five traits — extraversion, agreeableness, conscientiousness, neuroticism and openness to experience. Of those variants, 824 had never been linked to personality before. The paper, from the Revived Genomics of Personality Consortium (ReGPC) and led by researchers including Ted Schwaba, Michel Nivard and Elliot Tucker-Drob, appeared in Nature on September 2, 2026.

The headline number is a leap in scale: for conscientiousness alone, the count of genome-wide significant loci jumped from 3 in the most recent prior work to 131. Neuroticism, the most-studied trait, went from 224 known loci to 706. But the more consequential finding in the paper isn't the variant count — it's what the researchers did with that data to stress-test whether the associations are real, or artifacts of environment, culture and family structure masquerading as genetics.

What they actually found

Genetics explain a modest — not small, not huge — slice of why people differ in personality. Common genetic variants account for 4.8% to 9.3% of the variance in each Big Five trait, rising to 9.3–13.3% when the researchers corrected for the fact that most personality questionnaires are imperfect measuring instruments. That's well below the 40–60% heritability estimates that come from classic twin studies, a gap the authors attribute to twin studies also capturing rare variants and genetic interactions that GWAS methods can't see. In plain terms: DNA matters, but it is nowhere close to deterministic, and the environment remains the dominant force shaping who you are.

Each individual genetic variant's effect is vanishingly small — the median effect size for an extraversion-linked SNP corresponds to scoring at roughly the 50.35th percentile versus the 50th. Personality is not a single-gene story; the researchers estimate an average of over 16,000 independent genetic variants contribute to each trait. This is a portrait of extreme genetic complexity, not a "personality gene."

The finding that will matter most to skeptics: it isn't just family privilege in disguise

This is the study's most important methodological achievement, and it's worth dwelling on. A recurring criticism of behavioral genetics is that GWAS results for traits like educational attainment or income are contaminated by "genetic nurture" — parents pass down both genes and an environment correlated with those genes, so population-level associations partly reflect family wealth, stability and social status rather than direct biological effects on the child.

The ReGPC team tested for exactly this using within-family designs: comparing genetic differences between twins, and comparing children's traits to their parents' untransmitted genetic variants (using parent-offspring trios from Iceland's deCODE genetics database). For educational attainment, a comparison trait, only about 75% of the population-level genetic prediction survived this test — a full quarter of the apparent "genetic" effect was actually confounding by family environment. For personality, 96% of the effect survived. The researchers found this consistency across two independent methods (twin-pair comparisons and parent-offspring trio analysis), which is unusual in the field and strengthens confidence that the personality signal is close to direct biological effect rather than a proxy for privilege.

There were small exceptions. For conscientiousness, untransmitted parental genes explained a small but statistically significant share of the prediction — about 8%. And for agreeableness and neuroticism, genetic effects estimated within families correlated at less than 1.0 with population-level effects, meaning some, but not most, of the population signal reflects factors beyond direct genetic transmission.

Genetics also barely predicts who you marry

The study also quantified assortative mating — the degree to which people pick partners who resemble them genetically. Using genetic data on spouses, siblings and cousins across the UK Biobank and deCODE cohorts, the researchers found only minimal genetic assortment for personality (correlations of just 0.01–0.04), a sharp contrast with educational attainment, where genetic assortment among spouses is roughly ten times stronger (0.37). People clearly do select partners with compatible personalities in some sense, but that selection appears to leave little genetic fingerprint — a finding relevant to debates about whether "positive assortative mating" is quietly increasing polarization in psychological traits across generations. It does not appear to be doing so, at least not yet, and not much.

Where the genetics of personality touch health, behavior and where people live

The study's genetic-correlation and Mendelian randomization analyses (a technique that uses genetic variants as natural experiments to test causality) tie personality to a wide web of downstream outcomes:

• Conscientiousness is genetically linked to less substance use, more exercise, lower BMI, fewer hospital visits and healthier aging — and Mendelian randomization suggests some of this is causal, not just correlated: genetically-predicted conscientiousness reduced BMI and the likelihood of starting smoking.

• Neuroticism carries a broad genetic overlap with nearly all forms of psychiatric illness, especially internalizing disorders like depression and anxiety (genetic correlation of 0.73).

• Extraversion , somewhat unexpectedly, was linked via Mendelian randomization to increased COVID-19 infection risk — a plausible read is more social contact, more exposure.

• Openness to experience predicted a genetic tendency to move toward "cosmopolitan professional" urban neighborhoods later in life, while conscientiousness predicted movement toward affluent suburbs — a genetic fingerprint on the geography of where Americans and Britons choose to live, net of where they were born.

• Personality genetics also shape who shows up in research at all : people with genetic profiles linked to higher neuroticism were less likely to complete survey questionnaires and more likely to answer "I don't know," a quiet but important source of bias that could distort any study relying on self-report data.

The theory that didn't survive contact with the data

One of the paper's more pointed findings is negative: it undercuts a popular neurobiological theory that ties extraversion and openness to dopamine, and conscientiousness, neuroticism and agreeableness to serotonin. Using gene-expression data mapped onto specific brain cell types, the researchers found neither dopamine- nor serotonin-associated neurons were the most enriched cell type for any of the five traits — suggesting decades of textbook shorthand about "the extraversion neurotransmitter" oversimplifies a far messier biological picture involving broader neuronal gene networks in the prefrontal cortex.

The asterisk: this is still mostly a study of white Westerners

The study's generalizability claims are real but bounded. The consortium analyzed genomes with "European-like" ancestry (46 cohorts, the overwhelming majority of participants) alongside a smaller set with "African-like" ancestry (10 cohorts, roughly 43,000–48,000 participants per trait — under 5% of the total sample). Genetic prediction scores built from the European-ancestry data did carry over significantly to predict extraversion, neuroticism and openness among African-ancestry participants — a first for personality genetics — but the effect sizes were smaller, as expected whenever a genetic predictor trained on one ancestry group is applied to another. The paper is explicit that this is an early step, not proof of universal applicability, and it is confined to what the authors term "Western" cohorts across four geographic clusters (US, UK/Australia, continental Europe, Nordic countries) — leaving most of the globe, including Asia, Latin America and the Middle East, entirely outside the sampling frame.

Agreeableness was consistently the outlier trait throughout the paper: it had the lowest heritability, the least consistent genetic signal across countries and measurement instruments, and the fewest newly discovered variants relative to its prior literature. The authors suggest its genetic architecture may simply be more context-dependent than the other four traits.

Why this matters beyond the lab

For a story about DNA, this paper's real news value is almost sociological. It provides the cleanest evidence to date that personality — unlike education, income or cognitive test scores — is a trait where genetic associations are not mostly a mirror of family privilege. That distinction matters enormously for how policymakers, employers and insurers should (and shouldn't) think about using genetic information, and it complicates any simple narrative that "genes for success" traits are interchangeable with "genes for temperament."

At the same time, the study is a case study in appropriately modest claims. The authors are explicit, repeatedly, that individual genetic scores cannot predict any single person's future behavior, that even the largest effects found here are probabilistic and small, and that environment remains the dominant driver of personality variation. In an era where direct-to-consumer genetic testing and AI-powered "personality prediction" products are proliferating, that caveat — buried in paragraph after paragraph of an otherwise triumphant methods paper — may be the most important sentence in it.

Source: Schwaba, T., Clapp Sullivan, M.L., Akingbuwa, W.A. et al. "Robust inference and correlates from genetic associations with personality." Nature (2026). Published September 2, 2026. DOI: 10.1038/s41586-026-10992-9. Full summary statistics and code are publicly available via OSF.