TL;DR
This study presents a comprehensive analysis of pseudogene expression in human and macaque brains, revealing that many pseudogenes are actively transcribed and show tissue-specific patterns, particularly in the brain. These findings challenge the view of pseudogenes as mere "genomic fossils" and suggest they may play functional roles in brain development and psychiatric disorders.
What Problem Does This Paper Address?
While gene expression in the brain has been extensively studied, the expression patterns and potential functions of pseudogenes—DNA sequences similar to protein-coding genes but with disabling mutations—remain largely unexplored. Pseudogenes pose a technical challenge for analysis due to their high sequence similarity to their parent genes, making it difficult to accurately quantify their expression using standard RNA-seq pipelines. This lack of comprehensive data hinders our understanding of their potential roles in biological processes, including brain development and disease.
Methods
The researchers developed a novel computational framework to accurately quantify pseudogene expression from RNA-seq data. This framework involved:
- Pseudogene Annotation: Generating high-confidence pseudogene annotations for both human and macaque genomes by mapping existing annotations and integrating data from tools like PseudoPipe.
- Exon Filtering: Extracting and retaining only those pseudogene exons located in uniquely mappable genomic regions with high sequence uniqueness to minimize alignment ambiguity. Short exons were also excluded to enhance specificity.
- Expression Quantification: Applying this framework to large-scale RNA-seq datasets from human and macaque brains, as well as various non-brain tissues.
- Comparative Analysis: Comparing pseudogene expression patterns with those of their parent protein-coding genes and other protein-coding genes across tissues and brain regions.
- Spatiotemporal Analysis: Investigating the dynamic expression patterns of pseudogenes during brain development and aging.
- Co-expression Network Analysis (WGCNA): Constructing co-expression modules of pseudogenes and protein-coding genes to infer potential biological functions.
- Differential Expression Analysis: Identifying pseudogenes with altered expression in psychiatric disorders (ASD, BD, SCZ) and investigating their overlap with known disease-associated genes.
Datasets
The study utilized several large-scale RNA-seq datasets:
- Human Brain and Non-Brain Tissues: Data from the PsychENCODE Consortium and the ENCODE EN-TEx project.
- Macaque Brain Tissues: Data from various brain regions across the lifespan.
- Psychiatric Disorder Datasets: Human brain samples from individuals diagnosed with Autism Spectrum Disorder (ASD), Bipolar Disorder (BD), and Schizophrenia (SCZ), along with neurotypical controls.
Key Insights and Ideas
- Pseudogenes are Actively Transcribed in the Brain: Contrary to the "junk DNA" hypothesis, a significant proportion of pseudogenes are actively transcribed, especially in the brain.
- Brain-Specific Expression Patterns: Many pseudogenes exhibit tissue-specific expression, with a notable enrichment in brain tissues compared to other tissues. This suggests a potential functional role in brain biology.
- Complex Spatiotemporal Dynamics: Pseudogenes display more complex and varied temporal expression trajectories in different brain subregions during development and aging compared to the more uniform patterns of protein-coding genes.
- Functional Relevance in Disease: A subset of pseudogenes shows differential expression in psychiatric disorders, and some of these overlap with genes linked to disease susceptibility (eQTLs and GWAS-implicated genes), suggesting their involvement in pathogenesis.
- Co-expression Networks Reveal Functional Links: Co-expression analysis with protein-coding genes suggests that pseudogenes may participate in similar biological processes, particularly those related to neural activity and development.
Key Results
- 4,067 human pseudogenes were reliably quantified for expression.
- 56.8% of pseudogenes showed elevated expression in brain samples compared to their parent protein-coding genes (27.5%) and total protein-coding genes (29.2%).
- 49.3% of actively transcribed pseudogenes were broadly expressed across diverse brain regions, compared to 40.1% of total protein-coding genes.
- Spatiotemporal analysis revealed distinct patterns: 46.0% of temporally dynamic (TD) pseudogenes in humans were region-specific, compared to only 17% of TD protein-coding genes.
- In psychiatric disorders: Pseudogenes exhibited significantly greater expression fold changes than protein-coding genes. A considerable proportion (22.1% upregulated, 26.5% downregulated) of differentially expressed pseudogenes were shared across ASD, BD, and SCZ.
- Overlap with disease risk: Several SCZ-associated pseudogenes showed elevated brain expression and region-specificity, and overlap was found between differentially expressed pseudogenes and H-MAGMA-derived risk loci.
Limitations and Future Work
- Short-read RNA-seq limitations: Despite the developed framework, short-read RNA-seq can still lead to some bias in expression quantification due to high sequence similarity between pseudogenes and their parent genes. The exclusion of certain exonic regions might also contribute to underestimation of gene expression.
- Long-read sequencing: The authors suggest that the increasing availability of long-read RNA-seq datasets could offer greater sensitivity in detecting expression changes, particularly for pseudogenes.
- Complete genome assemblies: Advances in complete genome assemblies, such as for macaques, will facilitate more accurate genome annotation and transcriptomic analyses.
- Developmental datasets: Future studies can leverage developmental Genotype-Tissue Expression (dGTEx) projects to gain deeper insights into the regulatory mechanisms of pseudogene expression during development.
- Mechanistic roles: Further investigation is needed to elucidate the precise mechanistic roles of pseudogenes in brain function and psychiatric disorders, moving beyond transcriptional artifacts.
Summary updated