Corpus 1,272 assessed · 1,173 scored · 643 reproduced ≥75 · 168 flagged ·∅ 74.1/100
← New search

Transcriptome-wide analyses of piRNA binding sites suggest distinct mechanisms regulate piRNA binding and silencing in C. elegans.

RNA · 2023
L1 88/100 3/4
Why this verdict

The main results reproduced: recomputed values matched the published ones within tolerance.

Reproduced on the brainbox compute brainarbeit.com
✓ What held up
  • No relevant deviation in data/preprocessing
  • No authors-side cause for any deviation
  • Any deviation was negligible
What did not (or only partly)
  • 🟡Could not use the authors’ exact input data
  • 🟡Reported values were only indirectly comparable
  • 🟡Reported values were not (fully) derivable from the shared data
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
88/100
Reproducibility score
0.8 SD above mean
vs. all fields · 1173 studies
🎯 Scores higher than 74% of all assessed papers rank 276 of 1173 scored

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 (partial, healthy). RDT (github RyanCCJ/RDT @b8df63c) is the authors' shipped analysis toolkit; it ships the processed per-binding-site/per-read CSVs (data1..5) plus reference PNGs and gene lists that regenerate the paper's read-distribution figures. On «our HPC»/«infra» («job», conda env on compute nodes): 4 reported numbers reproduce EXACTLY -- gene-set sizes Germline 27792 / CSR-1 15821 / WAGO 3644 (line counts of shipped lists) and WS275 transcriptome N=43040 (region-table rows). 4 of 5 figures reproduce within-tol by running RDT on its shipped CSVs and overlaying onto the shipped references: Fig 2-A (miRNA metagene), 2-C (codon), 3-B (predicted-piRNA-site density boxplots), and 4-B-1 (WAGO-1 fold-change; the three M.W.W. p-values reproduce to ~3 significant figures, e.g. 3.318e-289 vs reported 3.324e-289). The plotted curves/boxes are visually IDENTICAL to the references -- the only difference is the auto-title (no --title passed). Fig 5-E-1 is honestly NOT regenerable from shipped data: it overlays two WAGO-9 HRDE-IP datasets (WT1_HRDEIP, CSR_KO1_HRDEIP) absent from shipped data1..5 (uncheckable). Critical env note: repo pins statannot==0.2.3 which needs seaborn<0.12 (sns.categorical._BoxPlotter, removed in seaborn>=0.12); with a modern seaborn the density/fold_change plots crash -- pin seaborn 0.11.2. NOT attempted: re-deriving the CSVs from raw SRA through the external CLASH-analyst/RNAup/pirScan/bowtie upstream (out of scope, external tool). No fabrication signal: every reproduced value/figure is derivable from the shipped data+code.

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.

  1. v1 current initial assessment Score 88
    assessed: 2026-06-19 ⛓ 3d64398bb50f
✎ 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.

Reason for the rerun

We email you a confirmation link first. The rerun is an objective re-measurement — it cannot change the verdict in your favour, only ask us to look again.

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
2026-06-30
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-19
no 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: sonnet
Founding hypothesis

Since C. elegans piRNAs trigger gene silencing only at a select subset of their sequence-predicted targeting sites, the paper asks what cellular mechanisms regulate which predicted piRNA targets are actually bound and silenced in vivo, by comparing in silico predicted piRNA targeting sites to experimentally determined in vivo piRNA binding sites.

Core claims
  • C. elegans piRNAs preferentially bind the coding regions (CDS) of target mRNAs in vivo, rather than 3' UTRs. finding
  • Sequence-based predicted piRNA targeting sites are instead enriched at the 3' end/3' UTR of germline mRNAs, opposite to the observed in vivo binding pattern. finding
  • piRNA CDS binding preference drives preferential production of secondary WAGO 22G-RNA silencing small RNAs at the CDS. finding
  • CSR-1 protects germline mRNAs from piRNA silencing through two distinct mechanisms: inhibiting piRNA binding across the entire CSR-1-targeted transcript, and inhibiting secondary WAGO 22G-RNA production locally at CSR-1-bound sites. mechanism
  • The CDS binding preference of piRNAs is not explained by CSR-1 activity, since CDS enrichment of piRNA binding occurs for both CSR-1 (germline-expressed) and WAGO (germline-silenced) targets. finding
  • In vivo miRNA (ALG-1) binding sites are enriched at the 3' UTR of mRNAs, confirming the CLASH/iCLIP-based analysis pipeline reproduces known small RNA binding biology. finding
  • Germline-enriched miRNAs also remain 3' UTR-enriched even when restricted to germline mRNAs, showing the CDS preference is piRNA-specific and not a general germline small RNA feature. finding
  • CLASH (crosslinking, ligation and sequencing of hybrids) analysis of PRG-1 is used as the method to identify direct in vivo piRNA-mRNA binding sites transcriptome-wide. method
Experimental setups
Assay System Perturbation Readout Platform
CLASH (crosslinking, ligation, sequencing of hybrids) reanalysis of published PRG-1 data C. elegans germline none (wild type, reanalysis of published data) in vivo piRNA binding sites/events on mRNAs
Bioinformatic prediction of piRNA targeting sites C. elegans germline mRNAs none predicted piRNA targeting site locations and density across 5'UTR/CDS/3'UTR
iCLIP reanalysis of ALG-1 (miRNA Argonaute) data C. elegans none in vivo miRNA binding site distribution across mRNA regions
Small RNA sequencing (WAGO-9/HRDE-1 and WAGO-1 22G-RNAs) C. elegans, WAGO-targeted and CSR-1-targeted germline mRNAs none density of WAGO 22G-RNAs across mRNA regions
Small RNA sequencing, WAGO-1 22G-RNA comparison C. elegans prg-1 (PIWI) mutant vs wild type ratio/reduction of WAGO-1 22G-RNA levels across mRNA regions
Small RNA sequencing (CSR-1 22G-RNAs) C. elegans, CSR-1 and WAGO target mRNAs none distribution/density of CSR-1 22G-RNAs across gene body
CLASH reanalysis of PRG-1 binding C. elegans, CSR-1 and WAGO target mRNAs CSR-1 depletion vs wild type piRNA binding events (normalized to piRNA-tRNA hybrid reads) across mRNA regions
Key results
  • In vivo piRNA binding density is highest in the CDS of germline mRNAs, peaking near the 5' end and declining sharply after the stop codon.
  • Predicted piRNA targeting sites are enriched at the 3' end, with highest density in the 3' UTR, opposite to observed binding.
  • In vivo miRNA (lin-4/ALG-1) binding sites are located in 3' UTRs, with binding density rising sharply just downstream of the stop codon.
  • WAGO-9 and WAGO-1 22G-RNAs are significantly enriched in the CDS for both WAGO targets and CSR-1 targets.
  • In prg-1 mutants, WAGO-1 22G-RNA levels are reduced more strongly in the CDS than in 5' or 3' UTRs, for both WAGO and CSR-1 targets.
  • CSR-1 22G-RNAs are higher at the 3' end of CSR-1 targets, while CSR-1 22G-RNAs on WAGO targets are more evenly distributed with enrichment at the CDS.
  • Restricting miRNA analysis to germline-enriched miRNAs and germline mRNAs still shows 3' UTR/3' end enrichment, unlike piRNAs.
  • Synthetic piRNAs targeting the CDS trigger robust gene silencing in reporter assays, whereas those targeting 5' or 3' UTRs do not, corroborating the transcriptome-wide CDS preference.
Key statistics
  • other Mann-Whitney U-test (Statistical test used to assess significance of density differences between mRNA regions (5' UTR, CDS, 3' UTR) for miRNA binding sites, predicted piRNA targeting sites, in vivo piRNA binding sites, and WAGO/CSR-1 22G-RNA densities across multiple figures; exact P-values not given in the extracted text.)
  • other 200 nt window centered at start or stop codon (Window size used for metagene analyses of binding site distribution around start and stop codons for both miRNA and piRNA binding data.)

Statistical methods review

Model: sonnet

A 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 compares transcriptome-wide sequencing-derived signals (predicted piRNA targeting sites, in vivo piRNA binding sites from CLASH data, miRNA binding sites from iCLIP data, and WAGO/CSR-1 22G-RNA read densities) across mRNA regions (5' UTR, CDS, 3' UTR) and across genotypes (wild type vs. prg-1 or CSR-1-depleted mutants). Results are shown as metagene traces (mean plus/minus one standard error) and as region-wise density comparisons (medians shown as blue lines), with pairwise statistical comparisons performed using the Mann-Whitney U-test. The manuscript text does not include a separate detailed statistical methods section describing sample sizes, replicate numbers, or software versions.

Replicationunclear GroupsmRNA regions (5' UTR, CDS, 3' UTR) and genotypes (wild type vs. prg-1 mutant; presence vs. absence of CSR-1) Pairingunclear Randomization/blindingnot stated Dispersionmixed Confidence intervalsno
Statistical tests used
Test Applied to n Assumptions
Mann–Whitney U-test Comparison of miRNA binding site density between mRNA regions (Fig. 2C) not stated
Mann–Whitney U-test Comparison of predicted piRNA targeting site density between mRNA regions (Fig. 3B) not stated
Mann–Whitney U-test Comparison of experimentally identified piRNA binding site density between mRNA regions (Fig. 3D) not stated
Mann–Whitney U-test Comparison of WAGO-9 and WAGO-1 22G-RNA density between mRNA regions, for WAGO and CSR-1 targets (Fig. 4A) not stated
Mann–Whitney U-test Comparison of WAGO-1 22G-RNA fold change (prg-1 mutant/WT) between mRNA regions (Fig. 4B) not stated
Approaches that could also have been used
  • Densities of binding sites/reads across three mRNA regions (5' UTR, CDS, 3' UTR) were compared using pairwise Mann–Whitney U-tests.
    Could also: A Kruskal-Wallis test across all three regions followed by a post-hoc test with multiplicity correction (e.g., Dunn's test with Benjamini-Hochberg or Bonferroni adjustment) — This would jointly test for an overall regional difference before pairwise comparisons and would explicitly control the family-wise error rate across the multiple region-to-region comparisons performed throughout the figures.
  • Spread in metagene traces is depicted as mean plus/minus one standard error (SE).
    Could also: Reporting a 95% confidence interval or standard deviation alongside the mean — A confidence interval directly communicates the precision of the estimated mean count at each position, while standard deviation conveys the underlying variability among individual transcripts, which can complement the SE-based depiction.
  • Read-density and binding-site comparisons between regions use a nonparametric rank-based test (Mann–Whitney U).
    Could also: A generalized linear model (e.g., negative binomial regression) modeling counts as a function of region and genotype — Count-based small RNA/CLASH read data are often overdispersed; a negative binomial or Poisson-based regression framework can model region and genotype effects jointly, account for covariates such as transcript length or expression level, and yield effect-size estimates with confidence intervals in addition to significance tests.
  • Comparisons of WAGO 22G-RNA levels between wild type and prg-1 mutant animals are summarized as fold-change ratios with density compared by Mann–Whitney U-test.
    Could also: A paired or repeated-measures analysis (if replicate samples exist) such as a paired Wilcoxon signed-rank test, or a dedicated differential-expression tool (e.g., DESeq2/edgeR) for small RNA count data — If the wild type and mutant measurements derive from matched transcripts or replicate samples, a paired approach or a count-based differential-expression model would account for gene-specific baseline levels and provide shrinkage-based fold-change estimates with associated confidence intervals.
  • Sample size, replicate structure (biological vs. technical), and randomization/blinding procedures are not detailed in the text.
    Could also: Explicit reporting of the number of biological replicates, transcripts, or reads underlying each comparison, along with a power or precision statement — Stating the n underlying each density/metagene comparison would let readers assess the precision of the reported medians and P-values independently.
  • Statistical significance across many region-by-region and genotype-by-genotype comparisons is reported without an explicit multiplicity adjustment.
    Could also: A pre-specified correction such as Benjamini-Hochberg FDR applied across the full set of regional/genotype comparisons in each figure panel — Applying a family-wise or FDR-based correction across the collection of Mann-Whitney comparisons shown in Figures 2-5 would control the overall false-positive rate when many comparisons are presented together.

What was reproduced

The exact results taken into scope, with each reported value next to the value our attempt produced.

scope.md — pmid-36737102

Paper: Wu WS, Brown JS, et al. "Transcriptome-wide analyses of piRNA binding sites suggest distinct mechanisms regulate piRNA binding and silencing in C. elegans." RNA (2023). PMID 36737102 / PMC10158993 / DOI 10.1261/rna.079441.122.

Code (P16, third-party-but-authors' own): https://github.com/RyanCCJ/RDT — "RNA-seq Read Distribution Toolkit" (RDT), pinned commit b8df63c9220755afc8142a62450e975215dd5875 (2023-10-17, default branch HEAD).

How the analysis is layered

The published analysis has two layers, and the repo ships BOTH plus the processed intermediates:

  1. Upstream pipeline (bioinfo/piRNA_project/code/, tool1–tool4):

    • Input: CLASH/iCLIP hybrids (from external "CLASH analyst"), RNAup energies, pirScan scores, 22G-RNA bowtie mappings — these RAW-derived inputs are NOT shipped (the *_output/ dirs contain only ignore.txt placeholders).
    • Produces the per-binding-site CSVs that feed the figures.
    • To run this from scratch needs: CLASH analyst (separate tool), RNAup (ViennaRNA), pirScan, bowtie/bowtie2, raw SRA reads. Hard / partially external.
  2. Downstream read-distribution + plotting (rdt-analyze.py, rdt-plot.py, examples/):

    • Ships the processed per-site/per-read CSVs examples/data/data1..5.csv (16–65 MB) that directly regenerate the paper's figures, plus reference PNGs in examples/fig/.
    • Deterministic pandas/seaborn/matplotlib. Easy, fully self-contained.

Shipped example data → figure map (from examples/data/README.md + wiki)

file sample (SRA) type level figure
data1 AGO/ALG iCLIP (SRR3882949) RNAup site 2-A (metagene), miRNA
data2 PRG-1 CLASH WT1 (SRR6512652) pirScan site piRNA distribution
data3 WAGO-1 IP WT (SRR8482951) 22G read_count 22G
data4 WAGO-1 IP WT, miRNA-normalized 22G read_count fold-change WT arm
data5 WAGO-1 IP prg-1 mutant, miRNA-norm 22G read_count fold-change mutant arm

Reference figures shipped: examples/fig/{2-A,2-C,3-B,4-B-1,5-E-1}.png.

IN SCOPE (pipeline-derived, attempted)

  • S1. Gene-set sizes (Germline 27,792 / CSR-1 15,821 / WAGO 3,644): consistency check of shipped gene_list/* vs reported N. DONE offline — all EXACT.
  • S2. Reference transcriptome N (WS275, 43,040 mRNAs): row count of shipped region table. (verify on «infra»)
  • S3. Figure regeneration via RDT downstream (Fig 2-A, 2-C, 3-B, 4-B-1, 5-E-1) from shipped data1..5.csv — metagene / density / codon / fold_change. Compare regenerated PNG to shipped reference PNG. Primary reproduction target.
  • S4 (stretch). Re-derive one dataN.csv from raw reads (e.g. data1 from SRR3882949) through RNAup/CLASH-analyst — upstream layer. Attempt only after S3.

OUT OF SCOPE (not pipeline-derivable here / wet-lab / external tool)

  • Wet-lab: PRG-1 CLASH library prep, IP experiments, mutant strains.
  • "CLASH analyst" hybrid-calling itself is an external closed-ish tool; not the authors' shipped code. We treat its OUTPUT (shipped CSVs) as the reproduction substrate (valid per brief P16) rather than re-deriving it.
  • Biological interpretation / model claims (CSR-1 "local vs transcript-wide inhibition" mechanism) — these are conclusions drawn from the distributions, not separately computable numbers.

Datasets relied upon (for profiling)

SRR3882949 (iCLIP, RU focus), SRR6512652/SRR6512654 (PRG-1 CLASH), SRR6512642/43 (mRNA-seq), SRR12318132 (CSR-1 22G), SRR8482951/49 (WAGO-1 22G), SRR12318140/44 (WAGO-9 22G), SRR3356499 (Ribo-seq). Primary profiled accession = SRR3882949; the shipped data1.csv is its processed binding-site table.

Compute plan

  • All on «our HPC»/«infra». Clone RDT + build a small conda env (python, numpy,
Figures / tables: Fig 2AFig 3BFig 2CFig 5EFig 4BFig 2
geneset_germline
Reported
27792
Reproduced
27792
exact
geneset_csr1
Reported
15821
Reproduced
15821
exact
geneset_wago
Reported
3644
Reproduced
3644
exact
transcriptome_n
Reported
43040
Reproduced
43040
exact
fig2A_mirna_metagene
Reported
Fig 2-A miRNA (RNAup) metagene
Reproduced
metagene/RNAup/data1 curve visually identical to examples/fig/2-A.png (NCC 0.78; only title differs)
within tolerance
fig2C_codon_dist
Reported
Fig 2-C codon (start/stop) distribution
Reproduced
codon/RNAup/data1 both panels identical to examples/fig/2-C.png (NCC 0.60)
within tolerance
fig3B_density_regions
Reported
Fig 3-B predicted piRNA targeting-site density (5pUTR/CDS/3pUTR)
Reproduced
density/pirScan/data2 boxplots + p-annotations identical to examples/fig/3-B.png (NCC 0.75)
within tolerance
fig4B_foldchange
Reported
Fig 4-B-1 log2(prg-1/WT) WAGO-1 22G; M.W.W. p=3.324e-289 / 5.043e-47 / 7.378e-115
Reproduced
foldchange data4/data5 restricted to WAGO-1 targets; M.W.W. p=3.318e-289 / 5.039e-47 / 7.359e-115 (match to ~3 sig figs)
within tolerance
fig5E_metagene_custom
Reported
Fig 5-E-1 WAGO-9 metagene comparison (WT1_HRDEIP vs CSR_KO1_HRDEIP)
Reproduced
NOT regenerable: the two WAGO-9 HRDE-IP input datasets are not among shipped data1..5
m.public.grade.uncheckable
piRNA_cds_pref
Reported
piRNAs preferentially recognize the coding region (model claim)
Reproduced
figure-level distributions reproduced; high-level model claim is an interpretation across figures+wet-lab, not one computable value
partial

Assessments & scoring basis

Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.

🤖 AI curator · claude (ai-curator room) · v1.0 L1 88/100

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.

🟡1. Data identity
🟡2. Endpoint comparability
🟢3. Location of the main deviation
🟢4. Cause of the deviation
🟡5. Derivability / plausibility
🟢6. Severity of the deviation
🟡7. Core claim
🟡8. Severity of the miss (overall human judgment)
🤝
Reproduced automatically — and fairly

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-4-8

Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.

101.8 k
tokens (I/O) · 5 M incl. cache
11 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.