Toddler temperament linked to common DNA variations in study of 78,894 children

Two regions stood out when checked against gene expression in 1,433 post-mortem adult cortex samples. A chromosome 12 signal for third-year activity shared an underlying variant with RHEBL1 expression at an estimated 93% probability; the version tied to lower toddler activity also links to more schooling years and higher cognitive test scores. A chromosome 1 signal tied to lower emotionality shared a variant with MR1 expression at 99% probability. MR1, never before linked to behavior, produces a protein that presents microbial byproducts to immune cells — fitting, the authors note, with evidence tying immunity to brain development.

How does toddler temperament connect to later outcomes?

Genome-wide correlations showed widespread overlap with neurodevelopmental, psychiatric, cognitive and personality traits:

• Higher activity: moderately strong positive overlap with ADHD, strong link to earlier walking, moderate negative link to educational attainment and cognitive performance.
• Higher sociability: moderately linked to lower likelihood of autism and higher adult extraversion.
• Higher shyness: lower extraversion, lower ADHD likelihood, slightly later walking.
• Higher emotionality: modest overlap with autism, neuroticism, depression and schizophrenia.

Follow-up modeling estimated 86% to 97% of shyness-linked variants overlap with those for ADHD, extraversion and age at walking — though many act in opposite directions across traits. Notably, shyness showed no significant genetic overlap with anxiety disorder, despite observational studies linking early shyness to later anxiety.

Does the finding hold beyond European populations?

Adding 5,618 to 5,653 Japanese-ancestry children from the Tohoku Medical Megabank cohort and 595 to 843 children from South Africa's Drakenstein Child Health Study, the chromosome 12 variant near RHEBL1 remained associated with third-year activity. Multi-ancestry polygenic scores performed similarly to European-trained scores in European children.

Polygenic scores — sums of many weighted variants — predicted parent-rated temperament in held-out cohorts, but explained only 0.06% to 1.9% of behavioral differences. Fraternal-twin comparisons suggested family environments did not heavily distort the signals.

What are the limits?

Plenty. The 10 regions were not replicated in an independent sample, and the cross-age analyses were added beyond the preregistered plan. Gene-expression evidence came from adult, not toddler, brain tissue. The common-variant estimates fall below twin studies' 30% to 60%, possibly because parent ratings reflect parental bias or exaggerate differences between fraternal twins. More than 90% of participants had European ancestry. The authors call for larger multi-ancestry samples, laboratory studies of RHEBL1 and MR1, and research on how environments interact with these genetic associations.

These links are probabilistic, not destiny — no score locks in a child's personality or diagnoses their future.