Rare Mutations Highlight 8 New Schizophrenia Risk Genes

The largest meta-analysis of whole-genome exome sequencing in schizophrenia to date has pushed the boundaries of the disease’s “genetic map.” The researchers combined new data from 4,650 patients and 5,719 controls with previously published data sets, bringing the sample to 28,898 cases, 103,041 controls, and 3,444 proband-parent trios. The result is two risk genes at the level of strict exome significance (STAG1 and ZNF136) and six more at the FDR < 5% level. The work reinforces the role of disrupted chromatin organization and offers specific candidates for models and target biology.

An important detail is that they did not simply increase the statistics, but showed the convergence of rare and common variants: for STAG1 and KLC1 in the same loci there are "finely mapped" associations according to GWAS, and for STAG1 this adds up to an "allelic series" - from common weak alleles to rare but strong damaging mutations. This increases the chance that the mechanisms seen in the models of rare variants will be relevant to a wide clinic.

Background

Schizophrenia is one of the most “genetically complex” mental illnesses: its heritability is estimated at 60-80%, with contributions coming from both thousands of common alleles of small effect (the GWAS map already includes hundreds of loci) and rare but “stronger” mutations in the coding regions of the genome. Modern large GWAS show that signals are especially concentrated in genes that work in excitatory and inhibitory neurons and are associated with synaptic transmission, that is, in the “wiring” of brain networks. It is against this background that rare, damaging variants are interesting as “mechanical anchors”: they are less likely, but better at highlighting vulnerable biological pathways.

In recent years, the SCHEMA consortium has collected and combined exome data and for the first time confidently demonstrated that rare "breaking" variants (premature stop codons, high-harm missense) in a number of genes significantly increase the risk of schizophrenia. At that time, at a strict significance level, it was possible to "catch" about a dozen genes and outline important intersections with other neurodevelopmental disorders (autism, epilepsy, mental retardation) - another argument that these conditions have a common biological architecture. But even such meta-analyses ran into statistical power: to confidently add new genes, tens of thousands of exomes and a combination of case-control with trio (search for de novo mutations) are needed.

This gap is what the current paper in Nature Communications is closing: the authors expand the exome sample to ~29,000 cases, >100,000 controls, and 3,400 trios, combining new and published data to squeeze out the rare mutation signal at the exome significance level and test for convergence with a map of common alleles (GWAS). This coupling of rare and common variants is key to prioritizing biology: if a locus is confirmed from both sides, the probability that it is indeed the causal gene/pathway increases dramatically.

In theory, this yields two practical dividends. First, precise models (iPSC neurons, CRISPR) for specific risk genes - from regulators of chromatin/transcription organization to participants in synaptic transmission and axonal transport. Second, biological stratification of future clinical trials: subgroups of patients with rare “anchor” mutations may respond differently to drugs affecting inhibitory transmission, synaptic plasticity, or gene regulation. But for this logic to work, the map of rare variants needs to become denser - which is why the next “leap” in exome volume and integration with GWAS is critical.

What exactly did they find?

• Exome significance (Bonferroni):
  STAG1 (PTV + missense MPC > 2; P = 3.7 × 10⁻⁸) is a component of the cohesin complex, a key to the spatial architecture of the genome (TADs, transcription regulation);
  ZNF136 (PTV; P = 6.2 × 10⁻⁷) is a KRAB zinc-finger repressor, its functional study is poor.
• New genes at FDR < 5%:
  SLC6A1 (GAT-1, GABA transporter; association via missense),
  KLC1 (kinesin light chain; missense),
  PCLO (Piccolo, active synapse zone),
  ZMYND11 (H3.3K36me3 tag reader, transcriptional regulation),
  BSCL2 (seipin, EP biology),
  CGREF1 (cell growth regulator).
• Cross-over with other disorders: enrichment of rare coding variants in STAG1, SLC6A1, ZMYND11, CGREF1 has been observed in neurodevelopmental and psychiatric conditions, further suggesting a common genetic architecture.

Why is this important? First, the "chromatin" line has strengthened: STAG1 directly indicates the vulnerability of the genome topology (cohesin, TAD boundaries), which is consistent with previous signals for variants that disrupt the structural organization of DNA. Second, SLC6A1 is a clear bridge to GABAergic dysfunction: harmful missense mutations in the GABA transporter are logically associated with changes in inhibitory transmission. Third, PCLO and KLC1 add components of the synaptic zone and axonal transport to the picture - levels where the delicate "logistics" of signals are easily disrupted.

How it was done - and why to trust

• New cohort + meta-analysis: public exome data were added to fresh 4,650/5,719, gene-wise analysis of rare coding variants (PTV, missense with MPC thresholds) was applied, case-control and de novo signals from the trio were meta-analyzed separately. Exome significance threshold was 1.63 × 10⁻⁶ (30,674 tests).
• Artifact control: sequencing coverage analysis, sensitive checks for "synonymous" singletons in controls/cases - resulting in effects for rare deleterious variants appearing conservatively estimated rather than false positives.
• Convergence of data layers: rare coding + common alleles (GWAS fine-map) + association with CNV loci (e.g. NRXN1) - classic "triangulation" increasing confidence in causality.

What does this add to the old SCHEMA picture?

• Before this work, ~12 genes had exome significance; the authors confirmed and “increased” two of the FDR candidates (STAG1, ZNF136) to a strict threshold and “added” six more at the FDR < 5% level. In other words, the exome map expanded and became more accurate.

Practical meaning - on the horizon of several years

• Models and target screening:
  • STAG1/KLC1 as “dual” candidates (rare+common variants) - first priority for cellular/animal models;
  • SLC6A1 - a natural entry point for studying GABAergic pharmacology in patient subgroups.
• Functional experiments:
  • reading chromatin footprints, CRISPR editing of alleles, analysis of TAD boundaries in the corresponding neuronal types and developmental stages; • testing the effect of SLC6A1/KLC1
  missense mutations on transport and transporters in neurons.
• Clinical perspective:
  • not about “a test tomorrow in the clinic”, but about stratification and biological subgroups in future trials;
  • possible link between genetic profile and response to drugs that affect inhibitory transmission or chromatin regulation.

Restrictions

• Exome is coding regions; does not catch rare regulatory variants in non-coding regions (WGS will come here).
• Most of the functional conclusions are inference from gene annotations; ZNF136 has almost no mechanisms - "wet" biology lies ahead.
• The effects of rare mutations are large but rare; they do not "explain" the entire disease but rather mark vulnerable pathways.

What will science ask next?

• Whole genome sequencing (WGS) to search for rare non-coding variants that disrupt TAD boundaries and enhancer-promoter contacts.
• Functional validation of novel risk genes (especially ZNF136, CGREF1, BSCL2 ) in human iPSC-derived neurons.
• Combining omnics: exome + single cell transcriptome + epigenome in the developing brain - to catch "when and where" a mutation strikes.

Conclusion

Rare "breakthrough" mutations continue to reveal vulnerable mechanisms of schizophrenia, from chromatin architecture (STAG1) to GABAergic transmission (SLC6A1). Convergence with data from common variants makes these genes prime candidates for functional biology and future patient stratification.

Source: Chick SL et al. Whole-exome sequencing analysis identifies risk genes for schizophrenia. Nature Communications, 2 August 2025. https://doi.org/10.1038/s41467-025-62429-y