449,484 people, 136 positions, 599 genes — and the most interesting result is that the trait split in two. What one questionnaire measures as one thing turned out to be two genetically distinguishable things.
Neuroticism is the personality dimension that describes how readily someone experiences negative emotion: worry, self-consciousness, irritability, low mood. It is measured with a questionnaire, it is stable across adult life, and it is one of the strongest known predictors of later depression and anxiety.
It is also, as personality traits go, unusually well studied genetically — because a questionnaire is cheap and half a million people have filled one in.
A 2018 meta-analysis assembled 449,484 people and found 136 independent positions reaching genome-wide significance — 124 of them new at the time — implicating 599 genes.
Before it, sixteen loci were known. This is what happens to a heritable trait when the sample size crosses a few hundred thousand: the signal does not get stronger, it gets numerous.
Thirteen of those positions are on this page. Each one moves a questionnaire score by an amount too small to notice in a person.
The analysis did not stop at a list. It asked which tissues and cell types the implicated genes are active in, and got specific answers: dopaminergic neuroblasts (P = 3.49 × 10⁻⁸), medium spiny neurons (P = 4.23 × 10⁻⁸) and serotonergic neurons (P = 1.37 × 10⁻⁷).
The pathways that came up were neurogenesis (P = 4.43 × 10⁻⁹), axon part (P = 5.26 × 10⁻⁸) and — unexpectedly — behavioural response to cocaine (P = 1.84 × 10⁻⁷). That last one is not about drug use. It is a gene set defined by what those genes do in reward circuitry, and its appearance says the trait touches the machinery of reward as well as of mood.
Here is the part that changes how the trait should be thought about.
The genetic signal for neuroticism partly originates in two genetically distinguishable subclusters, which the authors label “depressed affect” and “worry”. They suggest distinct causal mechanisms.
Read that against how the trait is measured. A neuroticism score is a single number, produced by summing answers to a questionnaire that treats being low and being anxious as facets of one thing. The genetics says they are not quite one thing — they overlap, and they are separable.
That is a finding about the instrument as much as about the biology. When a measurement is a sum, the sum can hide that it is adding two different quantities together, and only something that can see beneath the score — like the genome — will notice.
Nothing about your personality. With 136 loci and 599 genes, each individual position contributes a fraction of a fraction, and no combination of the thirteen here produces a meaningful score for a person.
There is also no diagnosis here. Neuroticism is a normal dimension of personality, not a disorder — everybody has a position on it. It is a risk factor for depression and anxiety in the same statistical sense that cholesterol is a risk factor for heart disease: real at the level of populations, and not a verdict on anyone.
And personality is not fixed by its heritability. That a trait is partly inherited says nothing about whether it can change, and neuroticism scores demonstrably do change — with age, with circumstances, and with treatment for the conditions it predicts.
Source. Nagel et al. (Nat Genet 2018) conducted a genome-wide association meta-analysis of neuroticism in n = 449,484, identifying 136 independent genome-wide significant loci — 124 new at the time of analysis — implicating 599 genes. Functional follow-up showed enrichment in several brain regions and specific cell types: dopaminergic neuroblasts (P = 3.49 × 10⁻⁸), medium spiny neurons (P = 4.23 × 10⁻⁸) and serotonergic neurons (P = 1.37 × 10⁻⁷). Gene-set analyses implicated neurogenesis (P = 4.43 × 10⁻⁹), behavioural response to cocaine processes (P = 1.84 × 10⁻⁷) and axon part (P = 5.26 × 10⁻⁸). The authors report that the genetic signal partly originates in two genetically distinguishable subclusters, 'depressed affect' and 'worry', suggesting distinct causal mechanisms.
Positions listed here. Thirteen of the 136, recorded by the collector from this study. Nearest-gene labels are annotation and not causal claims; for most of these the effector gene is not established.
Measurement. Neuroticism is an item-sum from personality inventories, and the composition of that sum differs between cohorts. The subcluster result bears directly on this: a score constructed as a single dimension is aggregating at least two partly separable genetic factors, which constrains how a neuroticism polygenic score should be interpreted and compared across studies using different instruments.
Clinical use. None. Neuroticism is a normal-range personality dimension rather than a diagnosis, and no genotype is used in its assessment or in the assessment of the psychiatric conditions it predicts. Depression and anxiety are diagnosed clinically.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Neuroticism comes down to these specific, well-studied positions — not a diagnosis.
ENSG00000225421 · rs1356493
See detailed info →LINC01122 · rs2708146
See detailed info →No. It is a position on a normal dimension of personality that everybody has a position on. It predicts depression and anxiety statistically, in the way cholesterol predicts heart disease — real across a population, and not a verdict on any individual.
That the questionnaire score is a sum, and the sum was adding two partly separable things together — the authors call them depressed affect and worry, and report distinct causal mechanisms. It is a finding about the measuring instrument as much as about the biology, and it is why a single neuroticism score should be handled carefully.
Yes, and the two questions are unrelated. Heritability describes where variation in a population comes from; it says nothing about whether an individual can change. Neuroticism scores fall with age in most samples and change with circumstances and with treatment.
No. There are 136 known loci implicating 599 genes, each contributing a fraction of a fraction of a questionnaire score. Thirteen of them tell you nothing about a person, and neither would all 136.
Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.