Penetrance and pleiotropy of polygenic risk scores for schizophrenia in 90,000 patients across three healthcare systems

Zheutlin, A. B.; Dennis, J.; Restrepo, N.; Straub, P.; Ruderfer, D.; Castro, V. M.; Chen, C.-Y.; Kirchner, H. L.; Chabris, C. F.; Davis, L. K.; Smoller, J. W.

doi:10.1101/421164

bioRxiv · 10.1101/421164

Penetrance and pleiotropy of polygenic risk scores for schizophrenia in 90,000 patients across three healthcare systems

Zheutlin, A. B.·Dennis, J.·Restrepo, N.·Straub, P.·Ruderfer, D.·Castro, V. M.·Chen, C.-Y.·Kirchner, H. L.·Chabris, C. F.·Davis, L. K.·Smoller, J. W.

View original source ↗Export metadata ↓

Abstract

BACKGROUNDIndividuals at high risk schizophrenia may benefit from early intervention but few validated risk predictors are available. Genetic profiling is one approach to risk stratification that has been extensively validated in research cohorts, but its utility in clinical settings remains largely unexplored. Moreover, the broad health consequences of a high genetic risk of schizophrenia are poorly understood, despite being highly relevant to treatment decisions.\n\nMETHODSWe used electronic health records of 91,980 patients from three large healthcare systems to evaluate the penetrance and pleiotropy of genetic risk for schizophrenia. Polygenic risk scores (PRSs) for schizophrenia were calculated from meta-analysis summary statistics and tested for association with schizophrenia diagnostic codes and 1338 code-defined disease categories in a phenome-wide association study. Effect estimates were meta-analyzed across sites, and follow-up analyses evaluated the effect of a schizophrenia diagnosis.\n\nRESULTSPRSs were robustly associated with schizophrenia (odds ratio per standard deviation increase in PRS = 1.65 [95% confidence interval (CI), 1.5-1.8], p = 1.25 x 10-16) and patients in the highest risk decile of the PRS distribution had a four-fold increased odds of schizophrenia compared to those in the bottom decile (95% CI, 2.4-6.5, p = 4.43 x 10-8). PRSs were also associated with other psychiatric phenotypes, including anxiety disorders, bipolar disorder, depression, substance use disorders, personality disorders, and suicidal behavior. Non-psychiatric associations included heart palpitations, urinary syndromes, obesity, and nonspecific somatic symptoms. Most associations remained significant when conditioning on a diagnosis of schizophrenia, indicating genetic pleiotropy.\n\nCONCLUSIONSWe demonstrate that an available measure of genetic risk for schizophrenia is robustly associated with schizophrenia in healthcare settings and has pleiotropic effects on related psychiatric disorders as well as other medical symptoms and syndromes. Our results provide an initial indication of the opportunities and limitations that may arise with the future application of PRS testing in healthcare systems.

Source connections

Published article ↗

Explore related subjects

genetics Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Zheutlin, A. B., Dennis, J., Restrepo, N., Straub, P., Ruderfer, D., Castro, V. M., Chen, C.-Y., Kirchner, H. L., Chabris, C. F., Davis, L. K., Smoller, J. W.. 2018-09-18. Penetrance and pleiotropy of polygenic risk scores for schizophrenia in 90,000 patients across three healthcare systems. https://doi.org/10.1101/421164

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Cellular context shapes the transcriptomic impact of somatic variants

Somatic SNVs are commonly prioritized by recurrence or sequence annotation, leaving the molecular consequences of private variants unresolved. We developed PrismSNV and Ligo-Seq, a single-cell full-length transcriptome sequencing platform, to quantify transcriptomic displacement associated with within-cell SNV-state toggling. Across 26 samples from three cancer types, more than 60% of candidate SNVs showed significant cell-type-specific displacement with over 30% variation in magnitude. Sequence-based deleteriousness was poorly aligned with cellular displacement, and fewer than 3% of top-displacement SNVs across six tumor types were catalogued in ClinVar. Displacement was strongest in transcriptionally plastic, less differentiated states. Analysis of 655,349 CD4+ T-cell transcriptomes further showed that SNV-gene associations changed in magnitude and sometimes direction during activation. Lack of recurrence or clinical annotation does not imply transcriptomic inertness; variant impact is shaped by cellular context rather than sequence alone, establishing the variant-cell pair as a critical unit for interpreting somatic variation.

genetics↗

bioRxiv2026

Genetic and Epigenetic Landscape of Self-Identified Hispanics/Latino in All Of Us

Human genetic studies still underrepresent many populations, limiting discovery and interpretation of medically relevant variation. Extensive admixture in self-identified Hispanic/Latino populations increases the need for representative reference data and ancestry-aware interpretation of medically relevant variation. We used long-read sequencing to analyze genetic variation, ancestry, and DNA methylation in 1,490 self-identified Hispanic/Latino participants from the All of Us Research Program. We identified 10 million novel variants and more than 100 candidate phenotype associations, with fine-mapping prioritizing structural variants over nearby single-nucleotide variants in more than half. We linked 3,329 tandem repeats to DNA methylation, resolved widespread allele-specific methylation, and identified ancestry-linked regulation across paralogous genes. These results reveal previously inaccessible genetic and regulatory variation and show the value of long-read sequencing for diverse populations.

genetics↗

bioRxiv2026

The genetic architecture of human DNA methylome variation

Population DNA methylome analyses remain largely confined to reference-defined genomic space despite extensive human genetic diversity. A central challenge is therefore to define comparable methylation phenotypes across diverse genomes and map the genetic architecture of methylome variation. Here, using 460 haplotypes from the Human Pangenome Reference Consortium Release 2, we develop two complementary pangenome-based frameworks: one compares methylation within homologous regions while retaining haplotype-specific structural configuration and CpG content; the other represents CpG island (CGI) variation at homologous loci as sequence-resolved alleles. We identified 47,039 CGI alleles, including 8,040 absent from CHM13; 62% were attributable to structural variants and 31% to small variants. Through local genealogies, we reconstructed CGI gain, loss and turnover, revealing how present-day methylation differences align with CGI evolutionary trajectories. We found that genome-wide, methylation variability was buffered against genetic diversity across more than 90% of the genome, with exceptions at extreme nucleotide diversity and locally around structural variants. Our methylation framework enabled population-scale allele-aware methylation quantitative trait locus (mQTL) mapping into non-reference and structurally complex sequence, where mQTL-associated intervals were enriched. mQTLs linked methylation to transcriptomic phenotypes, population differentiation and complex traits, supporting 8% of population differentially methylated regions and colocalizing with more than half of genome-wide association study loci. Together, our study redefines human methylome analysis and our findings reveal a genome-wide architecture of genetic-epigenetic coupling, reshaping how methylation variation is interpreted across human development and disease.

genetics↗