Our Take
The study expands schizophrenia's genetic map by 3.8x using published data and peer-reviewed methods, but the clinical pathway from 641 new associations to diagnostic or therapeutic use remains undrawn.
Why it matters
Psychiatric genetics has been stuck on local regulatory variants for a decade. This work shows trans-regulatory effects matter—a methodological win that could reshape how researchers approach other complex neuropsychiatric disorders where most heritability remains unexplained.
Do this week
Psychiatry researchers: audit your TWAS pipelines for cis-only bias before submitting new GWAS follow-up studies; incorporate multi-region transcriptomics if you have access to post-mortem brain tissue repositories.
641 Schizophrenia Genes Found Via Long-Range Regulatory Networks
A study published in Nature Genetics identified 766 schizophrenia-associated genes, of which 641 had not been detected by prior transcriptome-wide association studies (TWAS). The work comes from a team led by Giulio Pergola at the Lieber Institute for Brain Development.
The key methodological shift: instead of looking only at genetic variants within ±1 megabase of a gene (cis-eQTLs), the researchers built two new predictive models, INGENE and MODULE, that capture how distant variants influence gene expression through co-regulated network partners (trans effects). They integrated these trans-aware models with standard cis-based approaches and tested the combined framework against RNA-seq data from six post-mortem brain regions and genetic data from over 102,000 individuals (per the Psychiatric Genomics Consortium).
The combined approach improved gene-expression prediction for 18,744 genes. When applied to schizophrenia GWAS data, it surfaced 641 new disease-associated genes that cis-only methods had missed. Many of these trans-derived signals overlapped with known schizophrenia risk transcription factors (GATAD2A, RERE, IRF3, SP4).
The newly identified genes converge on glutamate signaling, neuronal communication, immune processes, and neurodevelopment—pathways already implicated in psychiatric risk from earlier studies.
A Methodological Shift, Not a Cure Claim
For 20 years, TWAS has dominated psychiatric genetics because it is computationally tractable and interpretable. But cis-only approaches explain only a fraction of heritable risk in complex disorders. Pergola's framing of the problem is direct: "looking for the light under the lamppost, focusing only on genes close to disease-associated DNA variants."
This study proves that trans-regulatory architecture is detectable at scale and biologically meaningful. It also demonstrates that integrating multi-region brain transcriptomics with large cohorts can reveal disease-relevant relationships invisible to cis-only pipelines.
The methodological advance is real. Whether the 641 new genes move psychiatry closer to precision diagnostics or drug targets is an open question. Gene discovery and clinical utility are not the same thing. The study provides a roadmap for expanding TWAS methodology; it does not claim to have identified 641 drug targets or diagnostic markers.
What Geneticists and Psychiatry Researchers Should Do Now
First, audit your TWAS pipeline. If you are running cis-only TWAS on psychiatric GWAS datasets, you are likely missing significant regulatory relationships. Adding trans-aware models requires post-mortem brain RNA-seq data, which is scarce but available through the CommonMind Consortium and LIBD itself.
Second, if you have access to multi-region transcriptomics, consider replicating this framework on your own GWAS cohorts. The models are not proprietary and the code is likely to be released alongside peer review.
Third, resist the temptation to treat gene discovery as gene function. The 641 new associations are a starting point for mechanistic follow-up, not a list of validated risk factors. Orthogonal validation in cell and animal models will be necessary before any of these genes move into therapeutic target prioritization.