The Alleles Your Parents Kept Still Predict Your Schooling

Start with the strangest result in this field, because everything else makes more sense afterwards.

Every parent passes on half of their DNA. The other half stays behind — it exists in the parent and never reaches the child. In 2018, a team writing in Science built a genetic score out of that leftover half for 21,637 people, and asked whether it predicted anything about them.

It did. The DNA that was not passed on still predicted how far the child went in school.

That should be impossible if we are talking about genes doing something inside a person. Those variants are not in the child. They cannot be affecting their brain, because they are not there. The only route left is the parents: the same DNA shaped the adults who built the home, chose the neighbourhood, filled or did not fill the shelves, sat down or did not sit down at the kitchen table in the evening. The researchers named it genetic nurture, and it is one of the cleanest demonstrations anyone has produced that a "genetic effect" and "an effect happening in your body" are not the same thing.

What the big studies actually found

Educational attainment — years of schooling completed — is one of the most heavily studied traits in human genetics, for an unglamorous reason: it is recorded for nearly everyone, so the samples get enormous.

In 2018, Nature Genetics published an analysis of about 1.1 million people that found 1,271 independent variants associated with it. In 2022, the same group went to about three million people and found 3,952.

Notice what happened between those two numbers. Tripling the sample tripled the count of variants. That is the signature of a trait where thousands of positions each nudge the outcome by an amount too small to detect until the sample is vast — the opposite of a single gene with a visible effect. There is no gene for finishing university, in the way there is a gene for whether milk sugar can be digested in adulthood.

The 12 to 16 per cent, and why it is smaller than it sounds

Add all 3,952 up into one score — a polygenic index — and it explains 12 to 16 per cent of the variation in how far people went in school.

Two things about that figure.

First, it is real, and by the standards of behavioural genetics it is large. Second, it is a statement about a population, not about a person. Twelve per cent of the variation across millions of people leaves the overwhelming majority of any individual's outcome to everything else: the school, the money, the illness, the teacher who noticed, the year that went wrong.

And then the same 2022 paper does something unusual and admirable. It compares the score's effect between families with its effect within a family — siblings, who share a home and half their DNA. Comparing siblings strips out everything the family has in common. When the authors control for the parents' own scores, the direct effect of a person's own DNA turns out to be roughly half the size of the headline association.

So of a number that was already smaller than people expect, about half is not the person's DNA acting on them at all. It is the same genetic nurture the Science paper found, showing up again from a different direction.

The single genes that were not there

If you have read about genetics and studying before, you have probably met a specific gene with a memorable story attached. A "warrior gene". A "worrier gene". A dopamine receptor that supposedly decided whether a child could focus.

Almost all of it came from a period when studies could afford a few hundred participants, could test a handful of genes chosen by guesswork, and could publish whichever result came out interesting. In 2019, a group in the American Journal of Psychiatry took the eighteen candidate genes that had been studied most often for depression — a closely related literature — and tested them properly in samples of up to 443,264 people.

None of them held up. Not the main effects, not the gene-by-environment interactions, not on any of the several ways of defining the outcome.

This matters here because those are still the results people meet first. They are in old news stories, in supplement marketing, and in direct-to-consumer reports. A page that quietly ignores them leaves the reader believing something the field abandoned a decade ago.

What the variants actually point at

The 2018 analysis noted that the associated positions cluster in genes involved in brain development and communication between neurons — which is both reassuring and much less useful than it sounds. It is the biological equivalent of learning that a city's traffic depends on its roads. True, unsurprising, and not a route to a prediction about one driver.

On this site the same pattern shows up in miniature. Our learning and focus topic collects the variants that share published papers with studying, attention and memory, and one position keeps arriving first: rs6265 in BDNF, a gene that makes a protein involved in how neurons grow and connect. It also turns up under sleep, under ADHD, under stress, under exercise. That is not a sign that BDNF explains all of them. It is a sign that a gene at the centre of how neurons work will be mentioned in almost any paper about the brain — and a reminder that being mentioned often is not the same as explaining much.

So what is actually true

• Differences in how far people get in school are partly heritable. That has been known from twin and family studies for decades and the molecular work agrees with it.
• The genetic part is spread across thousands of positions, none of which means anything on its own.
• The best available score explains 12 to 16 per cent at the population level, and about half of even that is the family rather than the person.
• The alleles your parents kept still predict your schooling, which is the most direct evidence anyone has that this is not a story about your biology alone.
• No test is worth taking, and no child should be sorted by one.

The honest version is more interesting than the headline version, and it is also kinder. "Partly heritable, mostly not, and half of the heritable part is the home you grew up in" is a description of a life, not a verdict on one.

Sources: educational attainment GWAS — Lee et al., Nature Genetics 2018, PMID 30038396; Okbay et al., Nature Genetics 2022, PMID 35361970. Genetic nurture — Kong et al., Science 2018, PMID 29371463. Candidate genes — Border et al., American Journal of Psychiatry 2019, PMID 30845820. This post is educational and is not medical or educational advice.