Transcriptome maps of general eukaryotic RNA degradation factors.
The main results reproduced, with only marginal, non-material deviations.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- ✓Same input data as the authors
- ✓Reported values are derivable from the shared data
- ✓Any deviation was negligible
- 🟡Reported values were only indirectly comparable
- 🟡A deviation arose in the data or preprocessing
- 🟡A deviation was attributed to the published material
- 🟡The central claim did not (fully) hold under reproduction
- 🟡Overall, the reproduction showed a material discrepancy
A 0–100 reproducibility-quality score from the per-question grades, shown as a z-score: standard deviations above (+) or below (−) the mean of comparable assessments.
▸Reproduction agent’s raw note
Described well enough to reproduce the entry-point results 1:1 from shipped data. GEO GSE128312 ships PROCESSED mockinbird crosslink-site tables (one per factor), so the heavy FASTQ->mockinbird preprocessing (authors' hard-coded reference paths) was correctly skipped (80/20) and we started from those tables + the public sacCer3 genome on «our HPC». REPRODUCED: (C1) every factor's verified-site table holds tens-to-hundreds of thousands of p<=0.005 cross-link sites - consistent with 'tens of thousands' as a floor; (C3) the Dcp2 AAAAU sequence preference - 100% of crosslinks at U (independent validation of PAR-CLIP T->C chemistry + coordinate indexing) and AAAA enrichment 5' of the U that brackets the reported ~5-fold. NOT REPRODUCED: (C4) decapping co-occupancy - our simplified +/-500bp overlap metric saturates and is not discriminative; the paper's actual metric is a Pearson correlation of co-occupancy PROFILES (co_occupancy.py + notebook), which we did not fully reimplement = the deliberately-skipped hard 20%, NOT a fabrication signal. NOT ATTEMPTED: replicate Spearman (0.94) needs per-replicate data only in SRA + a full mockinbird rerun; the 3193-mRNA metagene needs the authors' unshipped Steinmetz-TIF GFF. No fabrication detected; C1/C3 are directly derivable from shipped data and C3's perfect T-centering is independent evidence the published tables are genuine.
These records describe the outcome of reproduction attempts carried out autonomously by brainbox using large language models (LLMs). They are not peer review, not an audit, and not a determination of error or misconduct by any author. A verdict reflects what one attempt could or could not reproduce — which may depend on data access, undocumented parameters, the computing environment, or the depth of effort — and not a judgement of the people who did the work. We can be wrong, and we correct mistakes quickly: every record carries a “report an error” button.
Assessment versions
Every reproduction run is kept as an immutable version — anchored to the data as it stood, with a tamper-evident chain hash. A rerun (e.g. after an author updates a deposit) adds a new version; the previous one stays on record.
-
v1 current initial assessment Score 60assessed: 2026-06-14 ⛓ 0acaa11ed179
✎ I am an author of this paper
Updated or fixed a deposit, or is there an erratum? Ask us to re-run the metrics. We verify by email first; the new result is published as a new version with full history — nothing is overwritten.
Provenance — full disclosure
When this reproduction was carried out, which methodology version was used, and by whom — so the record can be audited and checked independently.
- Reproduced
- 2026-06-14
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-15no human curator yet
- Last updated
- 2026-08-05
Provisional, curator- or AI-assessed, and independently checkable. A reproduction outcome states what one attempt could reproduce — not a judgement of the authors.
Deep full-text extraction
Model: opusHow are RNA degradation pathways selected in a eukaryotic cell, and can the RNA sequence and class influence which degradation factors are recruited? The paper tests this by systematically mapping transcriptome-wide RNA-binding profiles of 30 general RNA degradation factors in S. cerevisiae.
- ★ Transcriptome-wide binding profiles of 30 general RNA degradation factors in S. cerevisiae reveal their distribution across different RNA classes. resource
- ★ Binding profiles are consistent with the canonical degradation pathway for closed-loop forming mRNAs after deadenylation. finding
- ★ Most degradation factors bind intact mRNAs, whereas decapping factors are recruited only for mRNA degradation, consistent with decapping being a rate-limiting step. mechanism
- ★ Decapping factors preferentially bind mRNAs with non-optimal codons, consistent with rapid degradation of inefficiently translated mRNAs. finding
- ★ The nuclear surveillance machinery (Nrd1/Nab3 and TRAMP4) targets aberrant nuclear RNAs and processes snoRNAs, while TRAMP5 is mainly involved in snoRNA degradation. mechanism
- ★ RNA degradation factors and individual subunits within complexes exhibit distinct transcript-class binding specificities. finding
- PAR-CLIP can be used to systematically generate high-confidence in vivo transcriptome maps for RNA degradation factors. method
- The Ccr4 subunit of the Ccr4/Not complex differs from other deadenylation factors and is strongly enriched at mRNA introns. finding
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| PAR-CLIP (photoactivatable ribonucleoside-enhanced crosslinking and immunoprecipitation) | S. cerevisiae (yeast) | TAP-tagged degradation factors | transcriptome-wide protein-RNA cross-link sites / RNA occupancy | TAP-tag immunoprecipitation; high-throughput sequencing |
| Western Blot | S. cerevisiae | TAP-tagged degradation factors | IP efficiency / protein detection | tag-specific antibody |
| Metagene occupancy profiling (PAR-CLIP derived) | S. cerevisiae snoRNA genes | none (TRAMP subunits Air1, Trf5, Mtr4, Air2, Trf4) | averaged binding occupancy along snoRNA gene body and 3' end (±50 nt) | — |
| CRAC comparison (external data reanalysis) | S. cerevisiae | none (Xrn1, Mtr4, Trf4, Ski2) | coverage/occupancy profiles around TSS and pA site (±700 nt) | — |
- – Biological replicate PAR-CLIP experiments for all 30 factors were highly reproducible. Spearman 0.87-1.00 (mean 0.94)
- – Most PAR-CLIP reads fall into the mRNA class, though many factors also bind ncRNAs, particularly rRNAs.
- ▲ Decapping factors and Xrn1 show strongest enrichment at SUTs and mRNAs (preferentially CDS and 3' UTR).
- ▲ Ccr4 is strongly enriched at mRNA introns, unlike other Ccr4/Not subunits.
- ▲ Rrp44 and Rrp4 preferentially bind short-lived nuclear CUTs and NUTs; Rrp6 binds rRNAs, snoRNAs, snRNAs, CUTs, NUTs.
- – TRAMP4 (Air2/Trf4) binds downstream of snoRNA 3' ends while TRAMP5 (Air1/Trf5) binds almost exclusively to the snoRNA gene body.
- – Tens of thousands of high-confidence cross-link sites obtained per factor.
- correlation Spearman 0.87 to 1.00 (mean 0.94) (reproducibility between two biological PAR-CLIP replicates per factor)
- pvalue p-values≤0.005 (threshold for verified factor-RNA cross-link sites)
- count 30 (general RNA degradation factors mapped)
- count 4928 (mRNA transcripts considered)
- count 77 (snoRNA genes used in metagene analysis)
- count 637 (CUTs transcript class)
- count 318 (SUTs transcript class)
- other 90 nt reduced to 50 nt then 10-12 nt (stepwise polyA tail shortening model (background, from literature))
Statistical methods review
Model: opusA neutral, descriptive read of the statistical approach — what was done, and (for shared learning, not as criticism) what could also have been done.
This is a genomics resource study that generated transcriptome-wide PAR-CLIP RNA-binding profiles for 30 RNA degradation factors in S. cerevisiae, with two independent biological replicates per factor. Reproducibility was assessed with Spearman correlations between replicates, cross-link sites were called at a stated significance threshold (p-values ≤ 0.005), and factor binding preferences across transcript classes were summarized using normalized log-enrichment scores and z-scores. Results are largely descriptive/comparative (heatmaps, metagene/occupancy profiles, enrichment scores) rather than reported via classical hypothesis-test p-values with effect sizes.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| Spearman rank correlation | agreement between two biological replicate PAR-CLIP experiments per factor (Figure 1—figure supplement 1A) | two biological replicates per factor; per-transcript points | not stated |
| PAR-CLIP cross-link site calling with significance threshold (p-value ≤ 0.005) | identification of high-confidence factor–RNA cross-link sites (Figure 1B) | — | not stated |
| Enrichment z-scores / normalized log-enrichment scores | binding preferences of factors across transcript classes/mRNA segments (Figure 2B) | transcript-class counts given (e.g., mRNA n=4928, rRNA n=24, tRNA n=299, snoRNA n=77, snRNA n=6, SUT n=318, CUT n=637, NUT n=298) | na |
| Coefficient of variation (SD/mean) | variability of enrichment per transcript class (Figure 2B top) | — | na |
-
Replicate agreement was summarized with Spearman correlation coefficients (range and mean reported).↳ Could also: One could additionally report concordance at the level of called sites (e.g., Irreproducible Discovery Rate, IDR) or Pearson correlation on transformed occupancies. — IDR is commonly used for CLIP/ChIP-type peak reproducibility and would quantify the fraction of reproducible binding sites in addition to overall rank agreement.
-
High-confidence cross-link sites were defined using a fixed per-site p-value threshold (≤0.005).↳ Could also: A false discovery rate framework (e.g., Benjamini-Hochberg) applied across the many sites tested could also be used to set the threshold. — An explicit FDR controls the expected proportion of false positives across the large family of simultaneously tested positions, complementing a single nominal p-value cutoff.
-
Factor binding preferences across transcript classes were conveyed with normalized enrichment scores/z-scores in a heatmap.↳ Could also: Accompanying point estimates with uncertainty intervals (e.g., bootstrap confidence intervals over replicates or sites) could also be provided. — Interval estimates would communicate the precision of each enrichment value, which is helpful for classes with very small n (e.g., snRNA n=6).
-
Per-class variability was summarized with a coefficient of variation.↳ Could also: Reporting SD or a 95% CI alongside the mean could also convey spread. — For small-n classes, showing both the central value and an explicit interval often makes the magnitude and reliability of differences easier to interpret.
-
Comparisons between factors and methods (e.g., PAR-CLIP vs. CRAC) were presented descriptively via averaged occupancy/metagene profiles.↳ Could also: A formal statistical comparison of profiles (e.g., per-position confidence bands or a model-based test of profile differences) could also be applied. — Adding uncertainty bands or a test would let readers gauge whether visible profile differences exceed sampling noise.
Result convergence & founder nodes
Findings this paper shares with others that ran a comparable experiment. A node’s strength is how many independent papers report it (replication breadth) — not how often it is cited, so a heavily-replicated but under-cited founder still stands out.
-
CCR4 is uniquely enriched at mRNA introns compared to all other Ccr4-Not deadenylase complex subunits in S. cerevisiae.other yeast up 2019×1papers★ This paper is the founder (earliest)
-
RNA degradation factors predominantly bind mRNA substrates transcriptome-wide, with many factors also targeting rRNAs and other ncRNAs in S. cerevisiae.other yeast 2019×1papers★ This paper is the founder (earliest)
-
RRP6 preferentially binds rRNAs, snoRNAs, snRNAs, CUTs, and NUTs, while RRP44 and RRP4 primarily target short-lived CUTs and NUTs in S. cerevisiae.other yeast up 2019×1papers★ This paper is the founder (earliest)
-
TRAMP4 (Air2/Trf4) binds downstream of snoRNA 3' ends while TRAMP5 (Air1/Trf5) binds primarily within the snoRNA gene body in S. cerevisiae.other yeast 2019×1papers★ This paper is the founder (earliest)
-
XRN1 and decapping factors show the strongest PAR-CLIP enrichment at SUTs and mRNAs, with binding concentrated in CDS and 3' UTR regions in S. cerevisiae.other yeast up 2019×1papers★ This paper is the founder (earliest)
Citation network
Where this publication sits in the reproducibility-weighted citation graph — what it is built on, and what is built on it. Citation data from OpenAlex.
No assessed neighbours yet — the network grows as more papers are assessed.
Data lineage
The datasets this paper uses (text-mined from the full text via Europe PMC), and which other assessed papers stand on the same data. A shared dataset is a factual link — not a judgement.
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-31135339
Paper: Sohrabi-Jahromi et al. 2019, Transcriptome maps of general eukaryotic
RNA degradation factors, eLife 8:e47040. PMID 31135339 / PMC6570525.
Code: https://github.com/soedinglab/Degradation_scripts (commit f3310951e0a2df1463da4ca8bd647af3073f5490)
Data: GEO GSE128312 → GSE128312_RAW.tar (128.9 MB), 30 files
GSM36712xx_<Factor>.table.txt.gz — one mockinbird crosslink-site table per
factor with columns seqid position transitions coverage score strand occupancy p_value. Raw FASTQ in SRA SRP188442 (not used).
Pipeline structure
- Preprocessing (mockinbird,
preprocess.yaml) — FASTQ → UMI/adapter trim → Bowtie map → dedup → T→C transition calling → per-site p-values. Heavy; needs authors' reference genome/index at hard-coded paths«path». OUT OF SCOPE (the shipped.tablefiles ARE this stage's output — we start here). - Postprocessing (mockinbird,
postprocess.yaml) — metagene CenterPlots, heatmaps, k-mer FASTA generation. Needs transcript-annotation GFFs (Steinmetz TIFs, intron, CDS) at authors' local paths — not shipped. - Figure notebooks (this repo) — consume
.tablefiles (+ some need the postprocess GFF/FASTA intermediates or external CRAC datasets).
In scope (reproducible from the shipped .table files; compute on «our HPC»)
| id | result | paper loc | script | inputs |
|---|---|---|---|---|
| C1 | "tens of thousands of verified factor-RNA cross-link sites with p-values≤0.005" | Results | (count) | shipped tables only |
| C3 | Dcp2 binds "AAAAU ... enriched ... ~2^2.3 = 5-fold" | Fig 7C / Fig 4–S1 | kmer_counting.ipynb | shipped table + public sacCer3 genome |
| C4 | "All factors of the decapping complex show very high co-occupancy and co-localization" | Fig 5 / text | co_occupancy.py | shipped tables only |
Out of scope / not attempted (the hard ~20%) — why
- Replicate Spearman 0.87–1.00, mean 0.94 (Fig 1–S1,
replicate_similarity.ipynb): needs per-replicate.tablefiles. GEO ships one merged table per factor → per-replicate data only via SRA + full mockinbird rerun. Skipped. - Metagene on 3193 mRNAs (Fig 3,
heatmap_metageneplots.ipynb): needs the authors' Steinmetz-TIF transcript GFF (steinmetz_transcripts_tifs_clean_sorted.gff, not shipped) for the 1500–5000 nt transcript set + mockinbird CenterPlot postprocessing. - Fig 2 transcript-class enrichment, Fig 6 codon/halflife regression: need external annotation/half-life/codon-optimality tables not in GSE128312.
- Full 74-factor co-occupancy matrix (Fig 5): 44 of 74 factors are external (Baejen/Schulz datasets). We reproduce the 30-degradation-factor sub-matrix.
Pipeline used for in-scope results: mockinbird PAR-CLIP (upstream, taken as given via shipped tables) + the repo's own Python analysis (reimplemented faithfully).
Assessments & scoring basis
Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.
An automated assessment. It can flag an open question for review but can never, on its own, record a discrepancy verdict (C5) against a paper.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
C1 (tens-to-hundreds of thousands of p<=0.005 sites) and C3 (Dcp2 AAAAU motif, 100% crosslinks at U, fold bracketing the reported ~5x) reproduce directly from the shipped mockinbird tables, with deviations only at the rounding / background-model level — and center_T=1.0 is strong internal evidence the published tables are genuine. The only mismatch is C4 decapping co-occupancy, which failed because we used a simplified ±500bp overlap proxy rather than the paper's Pearson correlation of co-occupancy profiles — a deliberately-skipped 20% on our side, not an authors' or data defect. R1/M1 were not attempted because per-replicate (SRA) and the Steinmetz-TIF GFF data were not shipped. Net: a solid partial reproduction with explainable, our-method-side gaps and no fabrication signal.
Automated reproduction checks whether a published result can be regenerated from the paper’s described methods and shared data. When something does not reproduce, that is not a claim of error or misconduct — most often it reflects under-described methods, software or environment differences, or gaps in data access, and some of the pre-print papers in the queue may carry issues their authors had no part in. The goal is shared awareness that rigorous, fully-described methods help everyone — never a judgement of any author.
Are you an author? We would genuinely like to hear from you — to clarify the record, add data or code, re-run the pipeline after an accession update, and publish your response right next to the assessment. Everything here is open and auditable.
🚩 Report an error in this record
Spotted something wrong — a verdict you’d contest, a data or value error, or a private detail that slipped through? Tell us, with a short justification. Authors and readers are equally welcome to write in; we review every report.
Prefer email, or the form below not working? Contact us at support@doesitreproduce.com.
Reproduction footprint
claude-opus-4-8Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.