TL;DR
This paper describes a community-driven initiative to create a standardized catalog of human translated open reading frames (ORFs) identified by Ribo-seq. By consolidating data from multiple studies and establishing a consensus nomenclature, this effort aims to bridge the gap between novel Ribo-seq discoveries and their integration into major biological databases, paving the way for deeper functional and clinical research.
What Problem Does This Paper Address?
Ribosome profiling (Ribo-seq) has revolutionized our understanding of the human genome's translational output, revealing thousands of novel translated open reading frames (ORFs), many of which reside in previously unannotated regions like long non-coding RNAs (lncRNAs) or untranslated regions (UTRs) of protein-coding genes. However, these "Ribo-seq ORFs" have largely remained outside of standard gene annotation databases (like GENCODE, Ensembl, UniProt, HGNC). This exclusion stems from uncertainties regarding their experimental reproducibility, physiological roles, and the lack of standardized annotation approaches. This creates a significant bottleneck, hindering research into the function and potential medical implications of these newly discovered translated elements.
Methods
The authors describe a multi-step process for creating a consolidated catalog of human Ribo-seq ORFs.
- ORF Retrieval and Mapping: They collected data from seven key Ribo-seq studies, filtering for ORFs found in lncRNAs, alternative protein-coding frames, and UTRs. They applied minimum length cutoffs (16 amino acids) and excluded near-cognate ORFs (those starting with non-AUG codons) for initial filtering to maximize replicability. ORF coordinates were mapped to the GRCh38/hg38 human genome assembly and Ensembl Release v.101. Redundant ORFs were clustered using a method similar to UniRef90, selecting the longest ORF variant.
- ORF Classification and Transcript Assignment: ORFs were classified into six biotypes based on their host transcript and genomic location relative to annotated protein-coding sequences (uORFs, uoORFs, dORFs, doORFs, intORFs, lncRNA-ORFs). When an ORF mapped to multiple transcript isoforms, the one with the highest APPRIS score was chosen.
- Multiple-species Alignment and PhyloCSF: For replicated ORFs (detected in more than one study), multiple-genome alignments were generated for 120 mammals. PhyloCSF, a tool for detecting protein evolution, was used to assess conservation and purifying selection. This was compared against annotated short coding sequences (sCDSs).
- Analysis of Proteomics Datasets: The authors searched for experimental evidence of protein production by re-analyzing 16 published mass spectrometry (MS) datasets. Peptides uniquely mapping to Ribo-seq ORFs and not to annotated protein-coding regions were considered as MS evidence.
Datasets
The study utilized data from seven published human Ribo-seq studies. Additionally, they incorporated multiple-genome alignments for 120 mammals and analyzed 16 published mass spectrometry (MS) datasets. The primary output of their work is a consolidated catalog of Ribo-seq ORFs, released as supplementary files.
Key Insights and Ideas
- The "Annotation Gap": Ribo-seq has uncovered a vast translational landscape, but this information is largely absent from standard gene annotations, creating a significant disconnect for researchers.
- Community-Driven Consensus: Recognizing the complexity and evolving nature of Ribo-seq ORF research, the authors emphasize a community-led approach involving major annotation projects (GENCODE, UniProt, HGNC) to establish standardized nomenclature and annotation strategies.
- Replicability as a Proxy for Robustness: The study highlights the importance of identifying Ribo-seq ORFs that are consistently detected across multiple independent studies as a measure of robustness, rather than relying solely on single-study findings.
- Diverse Roles of Ribo-seq ORFs: The paper underscores that Ribo-seq ORFs are not just "noise" but can represent functional proteins, regulatory elements, or even disease-associated peptides, necessitating a nuanced approach to annotation.
- Challenges in Evolutionary Interpretation: While evolutionary conservation (PhyloCSF) is a key indicator for canonical protein-coding genes, it is less straightforward for Ribo-seq ORFs, many of which are lineage-specific or may have regulatory functions not directly tied to protein sequence conservation.
Key Results
- The study generated a Phase I catalog of 7,264 unique Ribo-seq ORFs from seven human studies, after applying filtering and consolidation steps.
- A subset of 3,085 "replicated" Ribo-seq ORFs were identified, meaning they were detected in at least two independent studies.
- Ribo-seq ORFs are significantly shorter than annotated coding sequences (aCDSs).
- Only a small fraction (2.4%) of replicated Ribo-seq ORFs showed positive PhyloCSF scores, indicating limited evidence of purifying selection compared to annotated short CDSs (48%).
- 10 Ribo-seq ORFs were manually annotated as protein-coding by GENCODE based on this effort, with an additional 15 previously annotated through integrated analyses.
- Evidence for protein production was found for 424 replicated Ribo-seq ORFs through analysis of mass spectrometry data.
Limitations and Future Work
- Reproducibility Variability: The authors acknowledge that Ribo-seq ORF identification can vary significantly between studies due to differences in experimental protocols and computational pipelines.
- Limited Evolutionary Signal: Many Ribo-seq ORFs lack strong evolutionary conservation, making it challenging to infer function based solely on sequence evolution.
- Functional Ambiguity: The precise biological roles of many Ribo-seq ORFs (e.g., regulatory vs. protein-coding) remain unclear, posing a challenge for annotation.
- Near-Cognate ORFs: While near-cognate ORFs were identified, they were largely excluded from the initial Phase I catalog due to limited replicability across studies, suggesting a need for future investigation.
- Future Work: The paper outlines a Phase II that aims for deeper integration of Ribo-seq ORFs into the GENCODE gene annotation and the broader Ensembl database. This will involve addressing complex issues like database structure, developing standardized biotypes, and further evaluating the role of MS data. The authors also call for continued community engagement to refine ORF annotation and explore their implications for human health and disease.
Summary updated