📄 genetic and genomic medicine

Unraveling the Genetic Overlap Between Parkinson's Disease and Schizophrenia Through Genome-wide Association and Cell-Type Specific Transcriptomic Analysis

This study integrates genome-wide association and single-cell transcriptomic analyses to identify shared genetic loci between Parkinson's disease and schizophrenia, revealing that the RAI1 gene links both disorders through mitochondrial dysfunction while highlighting distinct genetic pathways associated with non-motor symptoms.

Original authors: Sun, W., Dehestani, M., Braun, A., Karmali, N., Wurster, I., Roeben, B., Kemmner, R., Brockmann, K., Sharma, M., Mitjans, M., Streit, F., Gasser, T., Bansal, V.

Published 2026-06-16

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Sun, W., Dehestani, M., Braun, A., Karmali, N., Wurster, I., Roeben, B., Kemmner, R., Brockmann, K., Sharma, M., Mitjans, M., Streit, F., Gasser, T., Bansal, V.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine the human brain as a massive, bustling city. In this city, two very different "disasters" can happen: Parkinson's Disease (PD) and Schizophrenia (SZ).

Parkinson's is like a power outage in the city's transportation system. The "dopamine trains" stop running, causing movement to slow down, shake, or freeze.
Schizophrenia is like a chaotic storm in the city's communication network. The "dopamine signals" get too loud or scrambled, leading to confusion, hallucinations, and disorganized thoughts.

For a long time, scientists thought these were two completely separate disasters with different causes. But this new study suggests that, surprisingly, they might share some of the same "blueprints" for how the city is built.

Here is a simple breakdown of what the researchers did and what they found:

1. The Great Genetic Detective Work

The researchers acted like detectives looking at two massive libraries of genetic instructions (DNA) from thousands of people: one library for people with Parkinson's and one for people with Schizophrenia.

They used a clever trick called the "Proxy-Phenotype Method." Think of it like this: If you are trying to find a specific type of brick used in a castle (Parkinson's), but you don't have enough blueprints of the castle yet, you look at the blueprints of a nearby fortress (Schizophrenia) that might use similar bricks. By cross-referencing the two, they hoped to find hidden connections.

2. The Three New "Shared Bricks"

After sifting through millions of genetic variations, they found 14 shared genetic spots (called SNPs) that appeared in both groups.

11 of these were already known to be linked to Parkinson's.
3 were brand new discoveries.

The most interesting of these new discoveries is a genetic spot called RAI1. Imagine RAI1 as a master switch in the city's power plant. The study found that this switch has a specific setting (a "missense variant," or a tiny typo in the code) that affects both the transportation system (Parkinson's) and the communication network (Schizophrenia).

3. What Happens Inside the Cells?

To understand why this switch matters, the researchers looked at tiny, living brain cells grown in a lab from Parkinson's patients. They treated these cells like a "test kitchen" to see what happens when the RAI1 switch is flipped.

The Power Failure: They found that when the RAI1 switch is in the "risk" position, the cells' mitochondria (the tiny batteries that power the cells) start to malfunction. It's like the city's power generators are sputtering.
The Backup Plan: Because the batteries are weak, the cells try to compensate by switching to a different, less efficient fuel source (glycolysis). It's like a city switching from electric trains to horse-drawn carriages to keep moving when the power grid fails.
The Serotonin Connection: They also noticed changes in how the cells handle serotonin (a chemical related to mood and sleep), which might explain why people with Parkinson's often experience depression or sleep issues, similar to symptoms seen in Schizophrenia.

4. The "Opposite Effects" Mystery

One of the most fascinating findings was that some of these shared genetic spots act in opposite directions.

Some genes make you more likely to get Parkinson's but less likely to get Schizophrenia (and vice versa).
The researchers found that the genes acting in this "opposite" way were uniquely linked to blood vessels in the brain.
The Analogy: Imagine the city's plumbing. If the plumbing is faulty, it doesn't just cause a leak (Parkinson's); it might also cause pressure issues that mess up the city's noise levels (Schizophrenia). The study suggests that these "opposite" genes might be the reason why some Parkinson's patients have specific non-motor symptoms (like mood or thinking issues) that feel different from the standard movement problems.

5. The Big Picture

The study concludes that while Parkinson's and Schizophrenia look very different on the outside, they share some deep, hidden genetic roots.

Shared Mechanisms: Both diseases seem to involve problems with mitochondrial power (energy) and inflammation (the city's immune response).
New Clues: The discovery of the RAI1 gene suggests that fixing how this gene works could potentially help with the energy problems in the brain cells of both disorders.
Non-Motor Symptoms: The study highlights that the "opposite" genetic effects might be the key to understanding why some Parkinson's patients struggle with non-movement issues like mood and sleep.

In short: This research didn't find a cure, but it found a new map. It shows that two different diseases might be caused by the same broken "blueprints" in the brain's power plant and plumbing, offering scientists new places to look for future treatments.

Technical Summary: Unraveling the Genetic Overlap Between Parkinson's Disease and Schizophrenia

Problem Statement
Parkinson's disease (PD) and schizophrenia (SZ) are clinically distinct disorders characterized by dopaminergic dysfunction, yet they exhibit overlapping symptoms (e.g., psychosis in PD, parkinsonism in SZ) and potential shared neurobiological mechanisms. While previous studies have identified individual shared risk variants, the extent of their shared genetic architecture and the specific convergent molecular pathways remain incompletely understood. Existing genetic correlation analyses have yielded mixed results, with some suggesting no global correlation, potentially due to the presence of shared variants with opposing effect directions (discordant effects) that cancel out in standard correlation metrics. This study aims to identify novel shared genetic loci between PD and SZ using large-scale, non-overlapping Genome-Wide Association Study (GWAS) summary statistics and to elucidate the functional relevance of these loci through single-cell transcriptomic analysis.

Methodology
The study employed a multi-stage analytical framework integrating statistical genetics with functional genomics:

GWAS Data Integration and Quality Control: The authors utilized summary statistics from large-scale GWAS of European ancestry: PD (56,306 cases, 1.4M controls; including 23andMe data) and SZ (55,193 cases, 74,132 controls). A custom meta-analysis was performed for SZ to ensure non-overlapping cohorts with PD. Strict quality control (QC) using MungeSumstats filtered for biallelic SNPs, corrected allele flipping, and removed strand-ambiguous variants, resulting in ~4.07 million overlapping SNPs.
Proxy-Phenotype Method (PPM) with Bayesian Framework: To enhance statistical power for discovering shared variants, the authors applied a two-stage PPM approach.
- Stage 1: SZ-associated SNPs ( $p < 1 \times 10^{-5}$ ) were identified and clumped ( $r^2 > 0.1$ ) to select independent candidates.
- Stage 2: These candidates were tested for association with PD. A strict Bonferroni-corrected threshold ( $p < 5.92 \times 10^{-5}$ ) was applied.
- Validation: A bivariate Bayesian regression model (brms package) was used to condition on effect sizes and standard errors, ranking SNPs by combined absolute effect sizes.
Novelty Assessment: Identified shared SNPs were cross-referenced with previously reported PD GWAS hits (90 loci) and other cross-disorder studies using LD tools (LDlink, LDassoc) to exclude variants in linkage disequilibrium (LD) with known signals.
Transcriptomic Analysis (scRNA-seq): The functional impact of the top shared missense variant, RAI1 rs11649804, was investigated using single-cell RNA sequencing data from iPSC-derived dopaminergic neurons (DNs) and immature DNs (IDNs) from the FOUNDIN-PD consortium. Genotypes were extracted using cellsnp-lite. Differential gene expression (DGE) analysis was performed between risk (CC) and non-risk (AA/CA) genotypes in idiopathic PD (iPD) and GBA1-PD subgroups.
Functional Enrichment: Gene Ontology (GO) enrichment analysis was conducted on differentially expressed genes (DEGs) using EnrichR.
Polygenic Risk Score (PRS) and Cell-Type Enrichment:
- PRS models were constructed in the Tuebingen Parkinson cohort (TUEPAC) using overlapping SNPs stratified by effect direction (concordant vs. discordant) to predict clinical outcomes (age at onset, non-motor symptoms).
- MAGMA gene-set enrichment analysis was performed to identify cell types enriched for concordant and discordant SNP sets using post-mortem brain single-nuclei data.

Key Contributions and Results

Identification of Shared Loci: The study identified 14 shared SNPs (5 concordant, 9 discordant effect directions). After filtering for LD with known loci, three novel shared SNPs were confirmed: rs2240921 (ITIH3), rs11649804 (RAI1), and rs9607782 (EP300-AS1).
Focus on RAI1: The missense variant rs11649804 in the Retinoic Acid Induced 1 (RAI1) gene was highlighted as a key shared signal.
- Expression Patterns: In iPD-derived DNs, the risk allele (C) was associated with increased RAI1 expression. Conversely, in GBA1-PD-derived IDNs, the risk allele was associated with decreased RAI1 expression.
- Pathway Enrichment: GO analysis of DEGs linked to RAI1 genotypes revealed significant enrichment in mitochondrial dysfunction pathways (e.g., aerobic electron transport, oxidative phosphorylation) and glycolysis. In GBA1-PD, upregulated genes (e.g., ENO1, PGK1, PKM) suggested a compensatory glycolytic response to mitochondrial deficits.
- Serotonergic Link: The study noted RAI1 expression distinguished GBA1-PD specifically in IDNs (which express serotonergic markers like TPH1), suggesting a potential shared mechanism involving serotonergic dysregulation.
Clinical Relevance of Effect Direction:
- PRS analysis in the TUEPAC cohort revealed that the discordant PRS (variants with opposing effects in PD and SZ) was significantly associated with non-motor symptom burden (NMSQ), whereas concordant PRS was not.
- Neither PRS significantly predicted age at onset after stratification, though both retained negative coefficients.
Cell-Type Specificity:
- Concordant variants were enriched in oligodendrocytes (ODCs), oligodendrocyte precursor cells (OPCs), and microglia.
- Discordant variants were enriched in ODCs, microglia, and uniquely in vascular cells. This suggests that the discordant genetic architecture may contribute to vascular components of PD non-motor symptoms.
Validation with Larger GWAS: Comparison with the 2025 PD GWAS (unavailable at the start of the study) showed that the association signal at the RAI1 locus strengthened, with rs12941356 reaching genome-wide significance, supporting the utility of the PPM approach for future discovery.

Significance and Claims
The authors claim that this study provides novel insights into the shared genetic architecture of PD and SZ, moving beyond global genetic correlations to identify specific loci with concordant and discordant effects. Key claims include:

Convergent Mechanisms: Functional characterization of RAI1 suggests that mitochondrial dysfunction represents a convergent biological mechanism underlying both neurodegenerative and psychiatric pathologies.
Non-Motor Heterogeneity: Variants with opposing cross-disorder effects (discordant SNPs) appear to capture biological pathways specifically contributing to non-motor symptom heterogeneity in PD, distinct from the broader genetic burden of the disease.
Cell-Type Specificity: The study identifies distinct cellular mechanisms for shared genetic liability, with discordant variants uniquely implicating vascular cells, potentially linking to vascular parkinsonism.
Methodological Utility: The integration of PPM with Bayesian frameworks and single-cell transcriptomics successfully identified novel shared loci that were not apparent in previous global correlation analyses, offering a refined strategy for cross-disorder genetic discovery.

The paper concludes that these findings advance the understanding of the molecular interplay between PD and SZ and suggest that targeting shared molecular pathways, particularly those involving mitochondrial function and vascular biology, may offer new therapeutic avenues. The authors maintain a modest tone, acknowledging limitations such as the reliance on bioinformatic approaches without functional validation, the restriction to European ancestry, and the need for larger cohorts to validate clinical associations.