The clearest published answer to "how many genes?" for any psychiatric condition: about 8,300 of them, together explaining at least a third of the liability. Which is also why no single one of the fourteen on this page tells you anything.
Solid lines are connections this site curates. Dashed lines mean the two ends share a research paper — worth knowing, and not a claim that one explains the other.
Schizophrenia is a psychiatric illness of altered perception and thought — hallucinations, delusions, disorganised thinking — usually beginning in late adolescence or early adulthood. It affects roughly 1 in 100 people over a lifetime and has been recognised as a distinct condition for over a century, without a biological marker to define it.
It is also strongly heritable, which for decades produced an obvious question and no answer: which genes?
A 2013 study assembled a Swedish national sample of 5,001 cases and 6,243 controls, meta-analysed it with earlier work covering 8,832 cases and 12,067 controls, then tested the leading signals again in 7,413 cases, 19,762 controls and 581 parent-offspring trios.
Twenty-two positions passed genome-wide significance — thirteen of them new. But the finding that mattered most was not any of them individually. It was the estimate the authors derived from the data as a whole:
That is the shape of the thing. Not a gene for schizophrenia; eight thousand positions, each nudging.
The genes near these positions were not a random draw. Several converge on neuronal calcium signalling — most visibly rs1006737 in CACNA1C, which encodes a subunit of a voltage-gated calcium channel.
Calcium entry is how a neuron converts electrical activity into a lasting change, so a set of hits landing there is a biological statement rather than a statistical one. CACNA1C is also one of the clearest points of overlap with bipolar disorder, which is a finding about psychiatric categories as much as about genetics.
These two facts are both true, and the tension between them is where most misreading happens.
Heritability of 32% from common variants means inheritance matters at the level of a population. Eight thousand contributing positions means no individual result is informative: each one moves risk by a fraction of a percent against a lifetime risk of about 1 percent.
Fourteen of those eight thousand are on this page. Reading them tells you nothing about yourself.
Schizophrenia is diagnosed clinically, from symptoms and their course over time. No genotype is used in diagnosis, in predicting onset, or in choosing an antipsychotic. Polygenic scores separate groups and do not classify people.
If this is a live question for you or someone close to you, the thing that changes an outcome is a psychiatric assessment — and early treatment, which is one of the few things in this illness that reliably matters.
The source. Ripke et al. (Nat Genet 2013) conducted a multi-stage GWAS beginning with a Swedish national sample of 5,001 cases and 6,243 controls, meta-analysed with previous schizophrenia GWAS totalling 8,832 cases and 12,067 controls, followed by replication of SNPs in 168 genomic regions in independent samples of 7,413 cases, 19,762 controls and 581 parent-offspring trios. Twenty-two loci reached genome-wide significance, 13 of them new and one previously implicated in bipolar disorder. Earlier signals on this page derive from the 2009 ISC/SGENE reports and subsequent meta-analyses.
Polygenicity. The authors estimated that approximately 8,300 independent, mostly common SNPs (95% credible interval 6,300–10,200) contribute to liability, collectively accounting for at least 32% of variance. This is the quantitative statement that the architecture is extremely polygenic; individual odds ratios at genome-wide significant loci are correspondingly small.
Calcium signalling. Examination of candidate genes at the associated loci implicated neuronal calcium signalling, with CACNA1C (rs1006737) the most-cited example. CACNA1C is among the loci shared with bipolar disorder, and the cross-disorder overlap is itself a replicated finding rather than an artefact of diagnostic imprecision.
MHC. rs13194053 lies in the major histocompatibility complex, the region with the strongest and least tractable schizophrenia signal. Extended linkage disequilibrium across the MHC makes attribution to a gene difficult, and interpretation of any single marker there should be correspondingly cautious.
Clinical use. None of these variants is used diagnostically or prognostically, and none informs antipsychotic selection. Polygenic risk scores for schizophrenia discriminate at the group level and are not clinically actionable for an individual.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Schizophrenia comes down to these specific, well-studied positions — not a diagnosis. 56 positions are linked to this page; the ones this page's own text discusses are shown first.
Because the inheritance is spread across roughly 8,300 positions. Heritability is a statement about how variation in a population is distributed; prediction is a statement about one person. With eight thousand contributors each moving risk by a fraction of a percent, the first can be large while the second stays useless.
It encodes a subunit of a voltage-gated calcium channel. Calcium entry is how a neuron turns electrical activity into a lasting change, so hits landing there are a biological claim rather than a statistical coincidence. It is also one of the clearest overlaps between schizophrenia and bipolar disorder.
There is no test to be given. No genotype is used in diagnosis or in predicting onset, and a polygenic score would not change what anybody does. What does matter is knowing that risk in a first-degree relative is elevated, and that early assessment and treatment are among the few things that reliably improve outcomes.
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.