Comprehensive analyses of RNA-seq and genome-wide data point to enrichment of neuronal cell type subsets in neuropsychiatric disorders.
- Authors
- Olislagers, M; Rademaker, K; Adan, R A H; Lin, B D; Luykx, J J
- Year
- 2022
- Journal
- Molecular psychiatry
- PMID
- 34719691
- DOI
- 10.1038/s41380-021-01324-6
- PMCID
- PMC9054675
Neurological and psychiatric disorders, including substance use disorders, share a range of symptoms, which could be the result of shared genetic background. Many genetic loci have been identified for these disorders using genome-wide association studies, but conclusive evidence about cell types wherein these loci are active is lacking. We aimed to uncover implicated brain cell types in neuropsychiatric traits and to assess consistency in results across RNA datasets and methods. We therefore comprehensively employed cell type enrichment methods by integrating single-cell transcriptomic data from mouse brain regions with an unprecedented dataset of 42 human genome-wide association study results of neuropsychiatric, substance use and behavioral/quantitative brain-related traits (n = 12,544,007 individuals). Single-cell transcriptomic datasets from the Karolinska Institute and 10x Genomics were used. Cell type enrichment was determined using Linkage Disequilibrium Score Regression, Multi-marker Analysis of GenoMic Annotation, and Data-driven Expression Prioritized Integration for Complex Traits. We found the largest degree of consistency across methods for implication of pyramidal cells in schizophrenia and cognitive performance. For other phenotypes, such as bipolar disorder, two methods implicated the same cell types, i.e., medium spiny neurons and pyramidal cells. For autism spectrum disorders and anorexia nervosa, no consistency in implicated cell types was observed across methods. We found no evidence for astrocytes being consistently implicated in neuropsychiatric traits. In conclusion, we provide comprehensive evidence for a subset of neuronal cell types being consistently implicated in several, but not all psychiatric disorders, while non-neuronal cell types seem less implicated.
Overview of the approach of dataset integration as inputs for enrichment methods, in order to detect implicated brain cell types for various phenotypes.Two mouse brain transcriptomic datasets (10x Genomics, KI) have the data format Sg,c of cell type specificity for genes, which was calculated by dividing expression of gene g in cell type c by expression of g in all cell types of a given dataset. Custom cell type identification was performed for 10x Genomics (16 detected cell types), while existing annotation was re-used for KI (first level of 24 cell types and second level of 149 cells (sub-)types). The datasets were integrated with genome-wide association study (GWAS) data, and these were the input for cell type enrichment methods DEPICT, MAGMA and LDSC. External human and mouse brain transcriptomics data were used in cell type enrichment method FUMA, so that enriched cell types from any of the other three methods could be compared to FUMA-enriched cell types. Finally, LDSC was also used to estimate SNP-based heritability for each GWAS phenotype and to calculate genetic correlations across all phenotypes.
Cell type enrichment estimated by DEPICT, LDSC, and MAGMA top 10% mode in selected brain-related phenotypes.Cell type enrichment results are generated using KI data. Bars represent the mean strength of association (-log10(P)) of LDSC, DEPICT, and MAGMA top 10% mode. The red line indicates the Bonferroni threshold P < 0.05/(24*42). The red line is solid if any of the methods identified any cell type as significantly associated, and if none of the methods identified any of the cell types as significantly associated, the red line is dashed. A complete overview of cell type enrichment results using KI data, including MAGMA linear mode is available in the supplementary information (Fig. S18).
Overview of enriched cell types of 42 common-variant psychiatric, neurologic, and behavioral/quantitative GWAS results in the KI dataset.ADHD; attention deficit hyperactivity disorder, ALS; amyotrophic lateral sclerosis, BMI; body mass index. Analyses from LDSC, DEPICT, and MAGMA top 10% mode, referred to as “methods” in the graph, show enrichment in MSNs and pyramidal cells (CA1) and pyramidal cells (SS) across brain-related phenotypes. The largest degree of consistency was found in SCZ and cognitive performance. Phenotypes and cell types are grouped by hierarchal clustering. Shades of pink are proportional to the mean strength of association (−log10(P)) of all methods. The color of the frames refers to the number of methods that identified a given cell type as significant in a given phenotype, after Bonferroni correction (P < 0.05/(24*42)). Gray frames: one method (intelligence, excessive daytime sleepiness, ADHD, drinks per week, ever smoked, chronotype, overall sleep duration, short sleep duration, MDD). Black frames: two methods (cross-disorders, educational attainment, BIP). Red frames: all three methods (human height, cognitive performance, SCZ).
No entities extracted from this document yet.
No uploaded files.
In this knowledge base
| Title | Year | PMID |
|---|---|---|
| Genome-wide analyses identify 30 loci associated with obsessive-compulsive disorder. | 2025 | 40360802 |
External
| Title | Authors | Journal | Year | Link |
|---|---|---|---|---|
| Detection of pleiotropic genetic factors and critical brain cell types linking insomnia with psychiatric disorders. | Xue B et al. | — | 2026 | → |
| Chromatin interaction maps of human arterioles reveal mechanisms for the genetic regulation of blood pressure. | Liu Y et al. | — | 2025 | → |
| Genome-wide analyses identify 30 loci associated with obsessive-compulsive disorder. | Strom NI et al. | — | 2025 | → |
| Integrative multi-omics data from early development to identify the genes and cell types underlying attention-deficit/hyperactivity disorder. | Jiao S et al. | — | 2025 | → |
| Decoding frontotemporal and cell-type-specific vulnerabilities to neuropsychiatric disorders and psychoactive drugs. | Ji J et al. | — | 2024 | → |
| Exploring PDE5A upregulation in bipolar disorder: insights from single-nucleus RNA sequencing of human basal ganglia. | Bai Z et al. | — | 2024 | → |
| Genome-wide analyses identify 30 loci associated with obsessive-compulsive disorder | Strom NI et al. | — | 2024 | — |
| scQA: A dual-perspective cell type identification model for single cell transcriptome data. | Li D et al. | — | 2024 | → |
| Single-cell multiomics analysis reveals cell/tissue-specific associations in bipolar disorder. | Wei W et al. | — | 2024 | → |
| An approach for prioritizing candidate genes from RNA-seq using preclinical cocaine self-administration datasets as a test case. | Vannan A et al. | — | 2023 | → |
| Associations between brain gene expression perturbations implicated by COVID-19 and psychiatric disorders. | Gao Y et al. | — | 2023 | → |
| Classification of Inflammation of Unknown Origin patients based on RNA-seq and SomaScan data. | Papagiannopoulos OD et al. | — | 2023 | → |
| Formulating treatment of major psychiatric disorders: algorithm targets the dominantly affected brain cell-types. | Fessel J | — | 2023 | → |
| Genetic Insights of Schizophrenia via Single Cell RNA-Sequencing Analyses. | Wu Y et al. | — | 2023 | → |
| Identification of brain cell types underlying genetic association with word reading and correlated traits. | Price KM et al. | — | 2023 | → |
| Single-Cell RNA Sequencing and Its Applications in the Study of Psychiatric Disorders. | Antunes ASLM et al. | — | 2023 | → |
| Unraveling Psychiatric Disorders through Neural Single-Cell Transcriptomics Approaches. | Chehimi SN et al. | — | 2023 | → |
| Disruption of the gene regulatory programme in neurodevelopmental disorders. | Griffin A et al. | — | 2022 | → |
| Editorial: Cross-disorder Genetics in Neuropsychiatry. | Van Assche E et al. | — | 2022 | → |
| Genetic identification of tissues and cell types underlying attention-deficit/hyperactivity disorder. | Wei WQ et al. | — | 2022 | → |