Cell-Type Classification of Psychiatric Disorders: New Study Reveals Cellular Basis of Schizophrenia and Other Complex Brain Disorders

Academic Background

Psychiatric disorders, such as schizophrenia, depression, and bipolar disorder, are significant public health issues worldwide. These disorders are typically caused by a combination of multiple genetic and environmental factors, with limited treatment options available. Although genome-wide association studies (GWAS) have identified thousands of genetic loci associated with psychiatric disorders, the physiological significance of these loci remains unclear. In recent years, advances in single-cell RNA sequencing (scRNA-seq) and single-nucleus RNA sequencing (snRNA-seq) technologies have enabled researchers to analyze gene expression at the single-cell level, providing deeper insights into the cellular basis of psychiatric disorders. However, integrating GWAS data with single-cell transcriptomic data to determine which cell types are closely related to the onset of psychiatric disorders remains a significant challenge.

This study aims to systematically analyze the cellular basis of psychiatric disorders like schizophrenia by combining GWAS and snRNA-seq data and to construct a cell-type-based classification system that offers new directions for drug development and personalized treatment.

Paper Source

This paper was co-authored by Laramie E. Duncan, Tayden Li, Madeleine Salem, and other researchers from Stanford University and several other institutions, and was published in Nature Neuroscience in February 2025.

Research Workflow

1. Data Sources and Preprocessing

The study utilized two main data sources: - GWAS Data: Derived from the latest GWAS datasets on phenotypes such as schizophrenia, alcohol consumption, sleep duration, multiple sclerosis, and Alzheimer’s disease. Sample sizes ranged from tens of thousands to millions, covering millions of genetic variants. - snRNA-seq Data: From Siletti et al.’s study, which performed single-nucleus RNA sequencing on 3,369,219 nuclei from 105 human brain regions and clustered them into 461 cell types.

During the preprocessing stage, the researchers applied a logarithmic transformation to the single-cell expression data, calculated the average expression levels of each gene across different cell types, and further computed the “specificity” scores, representing the proportion of a gene’s expression in a specific cell type relative to its total expression across all cell types.

2. Cell Type-Phenotype Association Analysis

The researchers used MAGMA software to perform gene-level analysis on GWAS data and tested the associations between each cell type and phenotype using a linear regression model. The specific steps included: - Gene-Level Analysis: Mapping single nucleotide polymorphisms (SNPs) from GWAS to genes and calculating the association P-values for each gene. - Gene Property Analysis: Using a linear regression model to test the relationship between gene specificity scores and gene association P-values while adjusting for covariates such as gene size and gene density. - Conditional Analysis: Determining relatively independent significant cell types through stepwise selection.

3. Comparative Phenotype Analysis

To validate the effectiveness of the method, the researchers also analyzed phenotypes such as alcohol consumption, sleep duration, multiple sclerosis, and Alzheimer’s disease. These phenotypes were chosen because they have well-established cell type associations and sufficiently powered GWAS data.

Key Results

1. Cell Type Associations in Schizophrenia

Among the 461 cell types, the researchers identified 109 cell types significantly associated with schizophrenia, including 10 relatively independent significant cell types. The most significant cell type was somatostatin (SST) interneurons in the cortex (p = 4.3 × 10^-17), followed by PAX6 interneurons widely distributed across the cortex and excitatory neurons primarily located in the retrosplenial cortex. Additionally, inhibitory neurons in the amygdala and excitatory neurons in the hippocampus were also found to be significantly associated with schizophrenia.

2. Cell Type Associations in Other Phenotypes

• Alcohol Consumption: The most significant cell type was D2 medium spiny neurons (p = 1.3 × 10^-9).
• Sleep Duration: The most significant cell type was D1 medium spiny neurons (p = 2.8 × 10^-9). Additionally, cell types in the pons and medulla related to sleep regulation were identified.
• Multiple Sclerosis: The most significant cell type was T cells (p = 6.0 × 10^-20), followed by B cells and natural killer cells.
• Alzheimer’s Disease: The most significant cell type was microglia (p = 2.4 × 10^-7).

3. Statistical Power and Method Robustness

By analyzing GWAS data with varying sample sizes, the researchers found that as sample size increased, the number of identified cell types gradually increased but stabilized after reaching a certain point. Moreover, simulation experiments involving random permutation of gene labels demonstrated the robustness of the MAGMA method in controlling false positives.

Research Conclusions

This study systematically identified cell types associated with complex brain disorders like schizophrenia by combining GWAS and snRNA-seq data and constructed a cell-type-based classification system. This system not only provides new perspectives for understanding the etiology of psychiatric disorders but also offers potential targets for drug development and personalized treatment. For instance, the finding that SST interneurons are closely related to schizophrenia provides a basis for developing treatment strategies targeting these cell types.

Research Highlights

1. Systematic Analysis: This study is the first to combine GWAS data with comprehensive single-cell transcriptomic data to systematically analyze the cellular basis of complex brain disorders like schizophrenia.
2. Multi-Phenotype Validation: The validity and robustness of the method were validated by analyzing phenotypes such as alcohol consumption, sleep duration, multiple sclerosis, and Alzheimer’s disease.
3. Cell Type Classification System: The researchers proposed a cell-type-based classification system, offering a new framework for the categorization and treatment of psychiatric disorders.

Significance and Value

This study not only deepens our understanding of the causes of psychiatric disorders but also provides new directions for future drug development. By identifying specific cell types associated with diseases, researchers can provide a basis for developing more precise treatment strategies. Furthermore, the methods used in this study can be applied to research on other complex diseases, offering new tools for understanding the cellular basis of diseases.