A data-driven estimation of the ribosome drop-off rate in S. cerevisiae reveals a correlation with the genes length.
The main results reproduced: recomputed values matched the published ones within tolerance.
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
REPRODUCED (1:1) the paper's core pipeline-derived result: the global ribosome drop-off rate estimates for both GSE91068 conditions, D1 (Synthetic Defined control) and D3 (Methionine restricted), Table 2. Fresh re-run of the authors' own ribofilio tool end-to-end on the paper's own SRA data on «our HPC» SLURM («job», COMPLETED exit 0, 1:27:39): SRA prefetch/fasterq-dump -> trim_galore 0.6.7 (adapter CTGTAGGCACCATCAAT) -> bowtie2 2.3.5.1 vs Ensembl rel-98 R64-1-1 cDNA (6612 tx) -> samtools 1.11 -> bedtools bamtobed -> ribofilio.py -b 50. SRA read counts (D1FP 41,917,507; D1mRNA 51,165,369; D3FP 42,972,168; D3mRNA 34,390,503) reproduce the repo run-log exactly; D1FP alignment 32.98%. All 12 Table-2 claims match to every printed decimal EXCEPT R^2 in the 4th significant figure (D1 0.4912 vs 0.4907; D3 0.4107 vs 0.4104), consistent with minor aligner-version drift; reproduced D1 also matches the repo's shipped notebooks/D1.regression.log anchor (magnitude identical; anchor records the decay with a negative sign). Outputs bit-identical to the prior (requeued) run. NOT attempted: per-gene-length/per-GO subsets and the gene-length correlation (Table 3, the harder ~20%); other-accession datasets D2/D4/D5-D10 (out of scope).
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 98assessed: 2026-06-19 ⛓ 4bbe7b92773a
✎ 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
- 2026-06-22
- Rubric version
- not recorded
- Assessed by
- —
- 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: sonnetThe paper tests whether ribosome drop-off (premature ribosome detachment from the mRNA before the stop codon) occurs at a quantifiable, non-negligible 'basal' rate in S. cerevisiae under standard (non-stressing) growth conditions, and whether this rate is related to gene length or to stress conditions such as nutrient restriction.
- ★ Ribosome drop-off events occur at a significant rate in S. cerevisiae cultured in standard conditions finding
- ★ There is a correlation between ribosome drop-off rate and gene length, with longer genes showing lower drop-off rates finding
- ★ Ribofilio, a novel open-source Python tool embedded in a Snakemake pipeline, estimates ribosome drop-off rate from Ribo-seq data resource
- ★ The overall ribosome drop-off rate in S. cerevisiae is in the range of 4 × 10-4 events per codon, similar to previously reported E. coli estimates finding
- ★ The dynamics of ribosome drop-off in S. cerevisiae are significantly different from those in E. coli, reflecting the higher complexity of the eukaryotic genome and its expression finding
- A basal, sequence-independent ribosome drop-off rate produces a negative exponential distribution of ribosome density along the CDS when translation initiation is the rate-limiting step mechanism
- Ribofilio estimates the per-bin drop-off rate rb via weighted linear regression of log(Y) on bin position X, then converts it to a per-codon rate rc = 1-(1-rb)^(1/lc) method
- Ribofilio achieved 81% module test coverage during software validation resource
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| Ribosome profiling (Ribo-seq) | S. cerevisiae, log-phase growth, rich (YEPD) medium (datasets D5, D6) | none (standard/reference condition) | 3' end read counts per transcript bin, used to estimate ribosome drop-off rate | — |
| Ribosome profiling (Ribo-seq) | S. cerevisiae (datasets D7, D8) | amino acid starvation | 3' end read counts per transcript bin, ribosome drop-off rate | — |
| Ribosome profiling (Ribo-seq) | S. cerevisiae (datasets D1/D3) | methionine restriction vs Synthetic Defined (SD) control | 3' end read counts per transcript bin, ribosome drop-off rate | — |
| Ribosome profiling (Ribo-seq) | S. cerevisiae (datasets D2/D4) | glucose restriction (caloric restriction) vs SD control | 3' end read counts per transcript bin, ribosome drop-off rate | — |
| RNA-seq | S. cerevisiae, matched to each Ribo-seq sample | matched to corresponding Ribo-seq condition | mRNA read counts per bin, used to normalise Ribo-seq counts | — |
| Read trimming and quality filtering | Ribo-seq/RNA-seq reads from S. cerevisiae | none | adapter-trimmed reads ≥20 nucleotides | trim_galore v0.6.7 |
| Read alignment to transcriptome and BED conversion | S. cerevisiae reads vs Saccharomyces_cerevisiae.R64-1-1.cdna.all.fa | none | mapped reads, primary alignment percentage, BED-format alignments | Bowtie2 v2.4.5; bedtools bamtobed |
- – Drop-off rate per codon (rc) across the eight datasets is broadly in the range of 4 × 10-4 events per codon rc 0.0003-0.0006
- ▼ Longer genes show lower ribosome drop-off rate
- – t-test rejects the null hypothesis rb=0 for all eight datasets p<0.00001 (D1-D4); p=0.0038/0.0029/0.0016/0.0020 (D5-D8)
- – Goodness-of-fit (R2) of the exponential decay model of ribosome density varies substantially across datasets R2 0.0341-0.6887
- – Ribofilio pipeline achieved 81% test coverage in module testing 81%
- other rc range 0.0003-0.0006 events per codon (drop-off rate per codon across datasets D1-D8 (Table 2))
- pvalue <0.00001 (t-test for rb=0, datasets D1-D4)
- pvalue 0.0038 (D5), 0.0029 (D6), 0.0016 (D7), 0.0020 (D8) (t-test for rb=0, datasets D5-D8 (rich medium / starvation))
- other R2 = 0.4907 (D1), 0.6887 (D2), 0.4104 (D3), 0.4993 (D4), 0.0341 (D5), 0.0462 (D6), 0.5336 (D7), 0.4188 (D8) (coefficient of determination for exponential fit of ribosome density per dataset)
- other rb per bin: D1=0.0051±0.0003, D5=0.0057±0.0003, D6=0.0053±0.0003 (drop-off rate per bin (bin size=50) with SE, Table 2)
- other 4 × 10-4 events per codon (previously reported in vivo E. coli drop-off rate (ref. 22), used as comparison)
- count 81% (Ribofilio module test coverage)
- other bin size = 50 nucleotides (Ribofilio parameter used to compute Table 2 drop-off rates)
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.
The paper introduces Ribofilio, a Python pipeline that estimates the ribosome drop-off rate per bin (r_b) from Ribo-seq data by fitting a weighted log-linear regression to binned, RNA-seq-normalised ribosome density profiles across eight publicly available S. cerevisiae datasets. Statistical significance of each estimated r_b is assessed by a t-test against the null hypothesis r_b = 0, with results reported per dataset as point estimates, standard errors, 95% confidence intervals, RMSE, and R². The authors further describe a qualitative correlation between gene length and the estimated drop-off rate.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| Weighted linear regression (log-linear model): log(Y) regressed on bin number X, slope estimated as r_b | Estimation of drop-off rate per bin for each of eight Ribo-seq datasets (D1–D8), Table 2 | N bins per dataset (bin size fixed at 50 nt; exact N per dataset not stated in text) | not stated |
| t-test (one-sample, null hypothesis r_b = 0) derived from regression slope SE with N−2 degrees of freedom | Statistical significance of estimated drop-off rate per dataset, Table 2 p-value column | N−2 degrees of freedom; N = number of bins per dataset (not given explicitly) | not stated |
-
The exponential model is fitted by log-transforming Y and applying weighted linear regression to log(Y) vs bin number X↳ Could also: Nonlinear least squares (NLS) fitting of the untransformed exponential model Y = A·exp(−r_b·X) directly on the original scale — Log-linearisation changes the error structure: residuals that are homoscedastic on the log scale become heteroscedastic on the original scale, and vice versa; NLS on the original scale allows weights that more directly reflect measurement uncertainty in ribosome read counts, and is a standard alternative when the functional form is known
-
95% confidence intervals for r_b are derived analytically from the regression slope SE and a t critical value (Equations 2–5), assuming normally distributed residuals↳ Could also: Non-parametric bootstrap confidence intervals for r_b obtained by resampling bins or gene subsets — Bootstrap CIs make no distributional assumption about residuals; they can better capture uncertainty when residuals are skewed or when N (number of bins) is small, and serve as a useful robustness check alongside analytic CIs
-
Eight independent t-tests are performed (one per dataset) with no adjustment for the family of comparisons↳ Could also: Benjamini-Hochberg FDR correction applied across all eight p-values — When multiple hypothesis tests are conducted on related datasets, FDR correction controls the expected proportion of false discoveries; it is a standard step when interpreting a set of related tests jointly and would not change any of the very small p-values reported here but would make the multiple-testing stance explicit
-
The correlation between gene length and drop-off rate is reported as a qualitative descriptive finding without a stated formal test or effect-size measure↳ Could also: Spearman rank correlation coefficient (ρ) with a permutation-based p-value, or Pearson correlation on log-transformed gene lengths, reported with a 95% CI — A formal correlation coefficient with confidence interval and p-value would quantify the strength and precision of the length–drop-off relationship; Spearman rank correlation is robust to the non-normality and heavy right tail typical of gene-length distributions
-
Model adequacy is assessed descriptively using R² and RMSE of the weighted log-linear regression↳ Could also: Formal information-criterion comparison (AIC or BIC) between the exponential decay model and nested alternatives such as a constant (no drop-off) or linear model — AIC/BIC penalise model complexity and provide a principled, quantitative basis for concluding that the exponential model captures the data better than simpler alternatives, complementing the descriptive R² value
-
A single fixed bin size of 50 nt is used for all reported analyses (Table 2 header)↳ Could also: Sensitivity analysis across a range of bin sizes (e.g. 25, 50, 75, 100 nt) with r_b estimates compared across choices — Binning granularity affects the number of data points N in the regression and the smoothing of read-count variability; reporting how r_b estimates vary with bin size demonstrates robustness of the key parameter to this methodological choice
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — PMID 38638702 (ribofilio, ribosome drop-off rate in S. cerevisiae)
- Paper: Awad et al., NAR Genom Bioinform 2024; DOI 10.1093/nargab/lqae036
- Code: https://github.com/SherineAwad/ribofilio (BSD-3, authors' own tool)
- Data: GEO GSE91068 (Zou et al.) — the RU's assigned dataset.
The tool & pipeline (from README + pipeline/Snakefile)
ribofilio estimates the ribosome drop-off rate from Ribo-seq footprints. It bins
each transcript into l-nt bins, counts footprint 3'-ends per bin (optionally
normalized by RNA-seq), and regresses log(Y) on bin number by weighted linear
regression: Y(X)=A·e^(−r·X). Output regression.log columns: Dropoff (per bin),
Dropoff per codon, RMSE, Rsquare, SE, Margin Error, tscore, pvalue, No.of Bins.
Full Snakemake pipeline per dataset:
trim_galore --gzip --retain_unpaired --trim1 -a CTGTAGGCACCATCAAT(adapter from config_D1D3.yaml)bowtie2-build yeast.fa yeast;bowtie2 -x yeast -U trimmed.fq.gz -S out.sam -p4(yeast.fa = Ensembl release-98 S. cerevisiae R64-1-1 cdna.all)samtools view -Sb;bedtools bamtobed→ BEDpython src/ribofilio.py -t yeast.fa -f FP.bed -r mRNA.bed -b 50 -v 1 -o PREFIXPinned tool versions (pipeline/env/*.yaml): bowtie2=2.3.5.1, samtools=1.11, trim_galore (paper text says trim_galore 0.6.7, bowtie2 2.4.5 — minor version drift noted).
GSE91068 sample → dataset mapping (from data/Makefile)
- D1 (Synthetic Defined, control): FP = SRR5090936, mRNA = SRR5090934
- D3 (Methionine restricted, treatment): FP = SRR5090937, mRNA = SRR5090935
IN SCOPE (pipeline-derived, attempted)
- D1 global drop-off rate — primary anchor. The repo ships the exact expected
output
notebooks/D1.regression.log, which equals paper Table 2 D1 (r_c = 0.0003/codon, R² = 0.4907, RMSE = 0.0143, 95% margin 0.0011, 295 bins). - D3 global drop-off rate — same GSE91068, treatment condition (secondary).
OUT OF SCOPE (not attempted, with reason)
- D2/D4 (GSE134152), D5–D10 (GSE13750): other accessions, not this RU's dataset.
- Per-gene-length and per-GO subset drop-off rates (Table 3, gene-length correlation): many sub-analyses; deferred as the hard last ~20% (80/20 rule). The global D1/D3 values exercise the same core estimator end-to-end.
- Wet-lab / external biological interpretation: out of scope by definition.
Reproduction strategy
Run the authors' own Snakemake pipeline steps (faithful third-party-of-record =
authors' tool on the paper's data) for D1 and D3 on «our HPC» SLURM, compare the
produced regression.log to the shipped notebooks/D1.regression.log (D1, exact
target) and to paper Table 2 / reported D3 value.
Assessments & scoring basis
Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.
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.