Disome-seq reveals widespread ribosome collisions that promote cotranslational protein folding.
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 were directly comparable
- ✓No authors-side cause for any deviation
- ✓Reported values are derivable from the shared data
- ✓Any deviation was negligible
- ✓The central claim held under reproduction
- ✓Overall, the reproduction was clean
- 🟡A deviation arose in the data or preprocessing
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
Yeast Disome-seq (Qian lab). BRIEF links were both mis-resolved and were corrected: real code = github.com/mingming-cgz/Disome-seq (not jdstorey/qvalue, the cited FDR pkg); real data = GEO GSE158572 (not GSE145723, which is the human comparison set, PMID 32375038). Described well enough to reproduce: the authors ship their R pipeline + intermediate processed peak tables. Reproduced the core pipeline-derived claims 1:1 by re-running their documented method (Mantel-Haenszel pausing scores) in Python on «our HPC»: A-site codon pausing dominated by the three stop codons (OR 40-51, Fig 3b) = EXACT; P-site pausing >1 for exactly Pro/Gly/Asn/Cys/Lys (Fig 4) = EXACT; genome stop-codon proportions 0.475/0.296/0.229 = within-tol (counts within 0.4%, ORF-set diff); replicate reproducibility high (mono r=0.998, mRNA r=0.997) = within-tol. R1 '46% of genes collide' only partial: the gene universe (~5124->5263) reproduces but the exact collision fraction depends on the authors' FDR peak-calling threshold, not fully re-derivable from the deposited processed tables. NOT attempted: Fig 2d collision-vs-translation Spearman (normalization undefined to us) and Fig 4c helix-propensity correlation (helix_propensity.txt not shipped). Out of scope: 3-nt periodicity/footprint length (raw FASTQ), cross-species conservation, and all wet-lab/MS results. No integrity concerns: every reproduced value is derivable from the shipped data+code and matches in direction and magnitude.
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 74assessed: 2026-06-18 ⛓ dfb87edc0a47
✎ 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
- 2026-06-18
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-18no 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: opusA slowdown or pause of a leading ribosome generates collisions with the 5′-elongating ribosome, producing disomes; the authors hypothesize that disome-seq can reveal widespread endogenous ribosome collisions and that these regulated collisions promote cotranslational protein folding and protein homeostasis.
- ★ Disome-seq sequences mRNA fragments protected by two stacked (collided) ribosomes, detecting ribosome collisions at codon resolution. method
- ★ Ribosome collisions are widespread in unstressed fast-proliferating yeast cells, occurring in ~46% of translated genes. finding
- ★ Disome-seq detects translational pauses that are missed by traditional monosome-seq (ribo-seq). finding
- ★ Ribosomes preferentially collide at stop codons, indicating slow ribosome release. finding
- ★ Slow peptide-bond formation from proline, glycine, asparagine, and cysteine at the P-site, and slow exit of polylysine from the exit tunnel, induce ribosome collisions. mechanism
- ★ Positions of ribosome collisions are evolutionarily conserved between yeast and humans (d=0 at 19.4% of sites). finding
- ★ Endogenous disomes have a different cryo-EM conformation from the RQC-inducing di-ribosomes. finding
- ★ Collisions occur preferentially in gap regions between α-helices, and paused/collided ribosomes associate with specific chaperones to aid cotranslational folding. mechanism
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| disome-seq (sequencing of mRNA fragments protected by two stacked ribosomes) | Saccharomyces cerevisiae (yeast), exponentially dividing / mid-log in rich medium | none | footprint abundance/length and genomic location of ribosome collisions | high-throughput sequencing; RNase I digestion; polyacrylamide gel |
| disome-seq | yeast cells treated with 3-amino-1,2,4-triazole (3-AT) | drug (3-AT, histidine biosynthesis inhibitor inducing histidine-codon pauses) | footprint length distribution and pausing at histidine codons | high-throughput sequencing |
| monosome-seq (ribo-seq) | yeast cells (rich medium and 3-AT treated) | none / drug (3-AT) | ribosome footprint abundance, ribosome density, A-site pausing scores | high-throughput sequencing |
| mRNA-seq | yeast cells | none | transcript codon frequency background for pausing-score normalization | high-throughput sequencing |
| sucrose gradient ultracentrifugation | yeast ribosome-bound mRNA | RNase I digestion (varied concentration) | abundance of monosome vs disome particles via UV absorption (OD254) | — |
| cryo-electron microscopy | yeast endogenous disomes | none | structural conformation of collided disome | — |
| mass spectrometry | yeast paused/collided ribosomes | none | chaperones associated with paused/collided ribosomes | — |
| reporter gene validation | yeast | heterologous reporter constructs | validation of sequence features associated with collisions | — |
- – ~5.8% of ribosomes are trapped in disomes in fast-proliferating yeast cells (monosome:disome population ratio ~32.6:1). 5.8%; ratio 32.6:1
- – Ribosome collisions detected in 2361 of 5124 translated genes (46%); 24% (1156/4742) under stringent UMI criterion. 46%
- ▲ Genes with higher monosome (ribosome) density showed higher frequency of ribosome collisions. ρ=0.39
- – 58-nt disome footprints peak 45-nt upstream of histidine codons, 30-nt upstream of the monosome peak, fitting two collided in-frame ribosomes. 30-nt spacing
- ▲ All three stop codons show extremely high A-site pausing scores in disome-seq but not monosome-seq; disome reads accumulate at stop codons genome-wide.
- ▲ Proline, glycine, asparagine, cysteine, and lysine show significantly >1 P-site pausing scores in disome-seq.
- – 19.4% of yeast collision sites occur at the same (orthologous) site in human, more than random expectation. 19.4%
- – Most abundant disome footprints are 58 and 59-nt under 3-AT, with 58-nt in-frame, versus 28-nt in-frame monosome footprints. 58/59-nt vs 28-nt
- correlation ρ = 0.39, P < 2.2 × 10−16 (Spearman correlation between monosome density and ribosome collision frequency)
- count 2361/5124 (46%) (translated genes with detected ribosome collisions)
- count 1156/4742 (24%) (genes with collisions under stringent ≥3 UMI criterion)
- other 5.8% (fraction of ribosomes trapped in disomes)
- other 32.6:1 (population ratio of monosomes to disomes)
- other 19.4% (d=0), P < 0.0001 permutation test (yeast collision sites conserved at orthologous human positions)
- count 589,461 monosome footprints; disome down-sampled to 18,082 (footprints used to define translated genes / matched ratio)
- count 3527 disome genes vs 5230 monosome genes (genes in aggregated footprint density profiles over CDS)
Statistical methods review
Model: sonnetA neutral, descriptive read of the statistical approach — what was done, and (for shared learning, not as criticism) what could also have been done.
This computational genomics study used high-throughput sequencing (disome-seq and monosome-seq) in yeast to characterize ribosome collisions at codon resolution. Associations between sequence features and collision propensity were quantified using Spearman rank correlations, Mantel-Haenszel pooled odds ratios, and permutation-based significance testing. Results were primarily reported as enrichment scores, percentages, and correlation coefficients across genome-wide codon positions, with two biological replicates used throughout.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| Spearman's rank correlation | Correlation between gene-level ribosome (monosome) density and disome collision frequency (Fig. 2d) | N = 4143 genes | not stated |
| Permutation test (10,000 randomizations) | Assessing evolutionary conservation of ribosome collision positions between yeast and human (Fig. 2h) | null | not stated |
| Mantel-Haenszel test (common odds ratio) | A-site and P-site pausing score estimation for each of 64 codons / 20 amino acids across all genes (Fig. 3b, 4a) | null | not stated |
-
The A-site and P-site pausing scores were computed as Mantel-Haenszel common odds ratios pooled across all genes, with the Mantel-Haenszel test providing significance↳ Could also: A mixed-effects logistic regression or negative binomial regression (e.g., via DESeq2 or edgeR on count data) could also model collision enrichment per codon while explicitly accounting for gene-level random effects and overdispersion in count data — Mixed-effects or count-based models directly model the hierarchical structure (codons nested within genes) and the discreteness/overdispersion of sequencing counts, which can improve calibration of p-values and provide confidence intervals on effect sizes
-
Evolutionary conservation of collision sites was assessed by a permutation test (randomizing collision positions 10,000 times within each gene)↳ Could also: A bootstrap confidence interval on the conserved fraction, or a binomial test against the null expectation derived from gene lengths, could also quantify the same conservation signal — Bootstrap CIs would additionally convey the precision of the 19.4% conserved-fraction estimate, making the magnitude of conservation easier to compare across future studies
-
Correlation between ribosome density and collision frequency was assessed with Spearman's rank correlation↳ Could also: Partial Spearman or partial Pearson correlation controlling for gene length and transcript abundance (mRNA-seq RPKM) could also be used — Gene length and expression level are potential confounders of both monosome density and collision rate; partial correlation would isolate the association of interest from these covariates
-
The study used two biological replicates for both disome-seq and monosome-seq↳ Could also: Three or more biological replicates per condition are commonly used in genomic sequencing studies, which enables formal variance estimation and improves the statistical power of differential analysis tools (e.g., DESeq2, edgeR) — With n=2 replicates, dispersion estimates are unstable; additional replicates would allow more reliable identification of collision sites with modest effect sizes and enable formal reproducibility quantification
-
Pausing scores were estimated for all 64 codons and 20 amino acids simultaneously without a described multiple-testing correction↳ Could also: A Benjamini-Hochberg false discovery rate (FDR) correction applied across the family of 64 codon tests (or 20 amino acid tests) could also be used to control the expected proportion of false positives — When scoring all codons simultaneously, the probability of at least one spurious significant result increases with the number of tests; FDR correction is a standard approach in genomics for communicating how many discoveries are expected to be false
-
Overlap between monosome-seq and disome-seq pausing sites was assessed by categorizing sites as above or below the per-gene mean footprint abundance↳ Could also: A receiver-operating-characteristic (ROC) analysis or a quantitative correlation of per-codon footprint densities between the two methods could also characterize the degree of concordance — A binary above/below-mean threshold can be sensitive to the choice of cutoff; a continuous comparison (e.g., rank correlation of per-site abundances) would provide a threshold-free assessment of concordance between the two sequencing modalities
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
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.
The scientific core reproduces 1:1: A-site pausing is dominated by the three stop codons (TGA 50.9/TAG 41.5/TAA 39.7, Fig 3b) and the exact P-site set Pro/Gly/Asn/Cys/Lys is significantly >1 (Fig 4), both re-derived independently via the authors' Mantel–Haenszel odds ratios. The only factual deviation (R6 stop-codon counts within 0.4%, identical proportions) is an input-side annotation-version effect, not an authors' or methodology defect. Unreproduced items (R1 collision fraction, R2 Fig 2d, R5 Fig 4c) are honest coverage gaps caused by undeposited FDR thresholds/tables, and the BRIEF links were mis-resolved and had to be corrected first. No integrity concerns — every reproduced value is derivable from the shipped data and code.
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.