Pseudogenes in the mouse lineage: transcriptional activity and strain-specific history

Cristina Sisu, Paul Muir, Adam Frankish, Ian T. Fiddes, Mark Diekhans, David Thybert, Duncan T. Odom, Paul Flicek, Thomas Keane, Tim Hubbard, Jennifer Harrow, Mark GersteinbioRxiv 2018 · DOI

TL;DR

This paper presents a comprehensive genome-wide annotation of pseudogenes across 18 mouse strains and the reference genome, revealing their dynamic evolutionary history and transcriptional activity. This resource highlights multiple retrotranspositional bursts in mouse evolution and identifies strain-specific pseudogenes, offering new insights into genome remodeling and functional evolution.

What Problem Does This Paper Address?

Pseudogenes, once considered "junk DNA," are now recognized as valuable markers for understanding genome evolution and remodeling. However, a comprehensive and comparative analysis of pseudogenes across multiple mouse strains, particularly in conjunction with their transcriptional activity and evolutionary history, has been lacking. This paper aims to fill this gap by providing a detailed annotation and analysis of pseudogenes in the mouse lineage, leveraging the availability of multiple sequenced mouse strains and developmental transcriptional data. Understanding pseudogene evolution can shed light on the processes that shape genomes and contribute to phenotypic diversity.

Methods

The authors employed a multi-faceted approach for pseudogene annotation:

  • Manual Curation and Automatic Pipelines: They combined rigorous manual curation of over 10,000 pseudogenes with results from automatic annotation pipelines (PseudoPipe and RetroFinder). This hybrid approach aimed to achieve both high accuracy and comprehensive coverage.
  • Cross-Strain Annotation: A novel workflow was developed to identify pseudogenes in 18 different mouse strains by lifting over annotations from the reference genome and using conserved protein-coding genes as input.
  • Confidence Levels: Pseudogenes were assigned confidence levels (Level 1, 2, and 3) based on the supporting evidence from manual and automatic annotation, providing a measure of annotation reliability.
  • Unitary Pseudogene Identification: A specialized workflow was developed to identify unitary pseudogenes, which are formed when functional genes acquire disabling mutations and have functional orthologs in other species. This involved cross-strain and cross-species protein-coding inputs and rigorous filtering.
  • Pangenome Dataset: A pangenome pseudogene dataset was created by combining all unique pseudogene entries across the strains to study conservation and divergence.
  • Phylogenetic Analysis: Phylogenetic trees were constructed using both pseudogene and protein-coding gene sequences to infer evolutionary relationships and pseudogene ages.
  • Functional Analysis: Gene Ontology (GO) terms and Pfam families were used to characterize the biological functions associated with pseudogene parent genes and to identify strain-specific functional annotations.
  • Pseudogene Transcription Analysis: RNA-seq data was used to assess the transcriptional activity of pseudogenes across different tissues and strains.

Datasets

  • Mouse Reference Genome: C57BL/6J strain.
  • Mouse Reference Annotation: GENCODE VM12/Ensembl 87.
  • Human Reference Genome Annotation: GENCODE v25/Ensembl 87.
  • 18 Mouse Strains: Assemblies and strain-specific annotations were obtained from the Mouse Genome Project. This included laboratory strains, wild-derived strains, and two outgroup species (Mus Caroli and Mus Pahari).
  • Developmental Transcriptional Data: Embryogenesis RNA-seq time course data was used.
  • Adult Mouse Brain RNA-seq Data: Used for transcriptional analysis.
  • ENCODE Data: Used for adult mouse brain RNA-seq data.
  • MGI Database and International Mouse Phenotyping Consortium: Used for compiling lists of essential and non-essential genes.

Key Insights and Ideas

  • Pseudogenes as Evolutionary Chronicles: Pseudogenes are not just genomic relics but active recorders of genome remodeling, reflecting bursts of activity and selective pressures over evolutionary time.
  • Mouse vs. Human Pseudogene Evolution: While overall pseudogene repertoires are similar, the mouse lineage shows multiple distinct bursts of retrotranspositional activity, unlike the single major burst in human evolution.
  • Strain-Specific Pseudogenes Reflect Diversity: A significant fraction of pseudogenes are unique to individual mouse strains, suggesting their role in driving strain-specific phenotypes and adaptations.
  • Transcription is More Than a Remnant: Pseudogene transcription is not merely a leftover from their gene past; it exhibits tissue and strain specificity, hinting at potential exaptive functions or regulatory roles.
  • Highly Expressed Genes Seed Pseudogenes: Parent genes with high expression levels, particularly housekeeping genes, are more likely to give rise to processed pseudogenes, suggesting a link between gene activity and pseudogene formation.

Key Results

  • Comprehensive Annotation: The study provides a detailed annotation of 20,397 pseudogenes in the mouse reference genome and a first draft annotation for 18 mouse strains.
  • Unitary Pseudogenes: 165 unitary pseudogenes were identified in mouse, and 303 in human, with new annotations significantly increasing the known counts.
  • Pseudogene Distribution: Approximately 80% of pseudogenes are processed, and 20% are duplicated, mirroring human patterns.
  • Strain-Specific Pseudogenes: On average, about a fifth of pseudogenes in each strain are unique.
  • Transcriptional Activity: Approximately 15% of pseudogenes are transcribed, with strain-specific transcription being more common than cross-strain conserved transcription.
  • Essential Gene Parents: Pseudogene parent genes are about three times more likely to be essential genes compared to regular protein-coding genes.
  • Retrotransposition Bursts: Mouse lineage shows continuous renewal of the pseudogene pool through successive bursts of retrotransposition, unlike the single burst in human.
  • Genome Remodeling: The conservation of pseudogene loci varies significantly between laboratory strains and outgroup species, indicating large-scale genomic rearrangements.

Limitations and Future Work

  • Annotation Snapshot: The current dataset is a snapshot of an ongoing annotation process, and future improvements in protein-coding annotation will refine pseudogene assignments.
  • Lower Bound Estimates: The number of strain-specific pseudogenes is presented as an upper bound, with efforts to estimate a lower bound by relaxing criteria.
  • Limited Tissue Data: Transcriptional analysis in mouse strains was limited to adult brain due to data availability.
  • Future Work: The authors suggest future work could involve investigating pan-strain and pan-tissue expression patterns as more RNA-seq data becomes available. Comparisons with primate lineages and human populations are also proposed, given the evolutionary distances. The developed pseudogene resource is available online at mouse.pseudogene.org for broader research community use.

Summary updated

Abstract

Pseudogenes are ideal markers of genome remodeling. In turn, the mouse is an ideal platform for studying them, particularly with the availability of developmental transcriptional data and the sequencing of 18 strains. Here, we present a comprehensive genome-wide annotation of the pseudogenes in the mouse reference genome and associated strains. We compiled this by combining manual curation of over 10,000 pseudogenes with results from automatic annotation pipelines. Also, by comparing the human and mouse, we annotated 165 unitary pseudogenes in mouse, and 303 unitaries in human. We make all our annotation available through mouse.pseudogene.org . The overall mouse pseudogene repertoire (in the reference and strains) is similar to human in terms of overall size, biotype distribution (~80% processed/~20% duplicated) and top family composition (with many GAPDH and ribosomal pseudogenes). However, notable differences arise in the pseudogene age distribution, with multiple retro-transpositional bursts in mouse evolutionary history and only one in human. Furthermore, in each strain about a fifth of the pseudogenes are unique, reflecting strain-specific functions and evolution. Additionally, we find that ~15% of the pseudogenes are transcribed, a fraction similar to that for human, and that pseudogene transcription exhibits greater tissue and strain specificity compared to protein-coding genes. Finally, we show that highly transcribed parent genes tend to give rise to processed pseudogenes.
BibTeX
@misc{sisu2018pseudogenes,
  author = {Cristina Sisu and Paul Muir and Adam Frankish and Ian T. Fiddes and Mark Diekhans and David Thybert and Duncan T. Odom and Paul Flicek and Thomas Keane and Tim Hubbard and Jennifer Harrow and Mark Gerstein},
  title = {{Pseudogenes in the mouse lineage: transcriptional activity and strain-specific history}},
  year = {2018},
  doi = {10.1101/386656},
  howpublished = {bioRxiv},
}

Loading related content...

Figure