Genomic and transcriptomic plasticity in treatment-naive ovarian cancer.
The main result did not reproduce in this reproduction attempt. Where our recomputation produced values that differ from the published ones, those discrepancies are listed below. This is a single automated attempt — not peer review and not a finding of error or misconduct — and differences can also arise from data access, undocumented parameters or the computing environment. The verdict can be contested via “report an error”.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- ✓The central claim held under reproduction
- 🟡Could not use the authors’ exact input data
- 🟡Reported values were only indirectly comparable
- 🟡A deviation arose in the data or preprocessing
- 🟡A deviation was attributed to the published material
- 🟡Reported values were not (fully) derivable from the shared data
- 🟡The deviation was non-trivial in magnitude
- 🟡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
Of the paper's pipeline-derived results, the core intratumoral-plasticity claim (recurrent large-scale copy-number events present in a subset but not all sampled tumor regions of a patient) was independently reproduced from raw GEO GSE47633 SNP6 array intensities using a from-scratch LogR + DNAcopy CBS segmentation pipeline, after recovering an undocumented probe-ID-to-genomic-position mapping not provided in any GEO or Bioconductor metadata (cnvi IDs map to featureSetCNV.man_fsetid via a +1,000,000 numeric offset; rs IDs map directly via featureSet.dbsnp_rs_id), yielding 81,218 genome-wide mapped probes from the 300,695 in the raw file. Applying the paper's own >=0.97Mb size threshold plus an amplitude/recurrence filter to the resulting 2,784 CBS segments (28 tumor samples, 3 patients, tumor-normal pairing resolved from tissue-type metadata) found 17 subset-specific large CN events (9 in P1, 8 in P2) against the paper's reported 14 -- the same order of magnitude and a qualitative confirmation of the plasticity claim, graded partial given the different segmentation algorithm (CBS, not ASCAT) and absence of purity/ploidy correction. All results depending on the paper's custom '1-2-3-SV' structural-variant caller (SV breakpoints, kataegis, gene fusions) are a documented drop: the assigned GitHub repository has been permanently removed by the author with no located fork or mirror. Targeted-panel SNV calling and full-scale RNA-seq differential expression were also not attempted (panel data/pipeline not located; RNA-seq data downloaded this session covers only a tiny fraction of the 55 RNA-Seq runs in PRJEB4193, insufficient for the paper's per-patient pairwise multi-region DE design). PCR validation of SV breakpoints and RT-PCR validation of fusions are wet-lab work and out of scope regardless. A supplementary WGS alignment QC check surfaced and corrected a data-format assumption error: ERR303475 and ERR303476 are independent single-end WGS runs, not a paired-end mate pair as initially assumed, which had caused a failed bwa mem run -- documented here for any future session rather than silently discarded. GEO GSE47633 is fully accounted for (31/31 samples); ENA PRJEB4193 (178 total runs) was profiled via the ENA Portal API but only 2 WGS runs were downloaded and inspected, so its dataset completeness assessment is necessarily shallow. No self-invented deadlines caused any premature stop; all SLURM/time limits encountered were real cluster policies (no --mem flag permitted, «infra» command-chaining restrictions on the login node), not treated as negative scientific findings.
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.
✎ 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
- 2026-07-30
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-07-31no human curator yet
- Last updated
- 2026-07-31
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: opusThe study asks how extensive intra-tumor genomic and transcriptomic heterogeneity is in treatment-naïve advanced (stage IIIC/IV) epithelial ovarian cancer, and whether site-specific genomic rearrangements causally drive intra-tumor differences in gene expression.
- ★ Treatment-naïve epithelial ovarian cancers show extensive intra-tumor heterogeneity of genomic rearrangements, with the most substantial differences occurring between omentum/peritoneum metastases and ovarian tumor sites. finding
- ★ Intra-tumor gene expression differences are caused by site-specific genomic alterations, including formation of in-frame fusion genes. finding
- ★ Lesion-specific breakpoints affect cancer genes including NF1, CDKN2A, and FANCD2. finding
- ★ Intra-tumor variability spans multiple mutational hallmarks: lesion-specific kataegis coinciding with genomic breakpoints, rearrangement class distributions, and coding mutations (including differing Ti/Tv ratios between branches). finding
- ★ Two independent TP53 missense mutations (P278L and I195N) arose in different tumor locations of one patient (ovary tumors vs. omentum/peritoneum metastases). finding
- ★ Key cancer pathways (WNT, integrin, chemokine, Hedgehog signaling) are up-regulated in only subsets of tumor samples from the same patient, and biopsies from one patient can fall into different established ovarian cancer expression subtypes (e.g. C1/C2 vs C4). finding
- ★ Combined topographic mapping of somatic breakpoints with transcriptional profiling of multiple physically separated biopsies per patient enables multilevel analysis of tumor evolution. method
- Simultaneous clustering of discordant mate-pair reads across all biopsies of a patient allows sensitive genotyping of low-frequency breakpoints (down to a single discordant read pair after a robust call elsewhere). method
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| Whole-genome mate-pair sequencing (~3 kb insert) | Tumor biopsies and whole blood from three treatment-naïve advanced epithelial ovarian cancer patients (34 samples: 27 tumor, 7 reference) | none (observational, treatment-naïve) | Somatic genomic rearrangement breakpoints and rearrangement classes (deletions, tandem duplications, inversions, interchromosomal) | — |
| Targeted sequencing of coding sequences of 2099 cancer genes | Each tumor biopsy and matching normal tissue from the three patients | none | Somatic single-nucleotide mutations and their per-sample frequencies | — |
| PCR-based resequencing validation of mutations | All tumor and matching normal tissue samples | none | Mutation validation and refined mutation frequencies at >1000× coverage | MiSeq |
| PCR validation of breakpoint junctions | Tumor biopsy DNA from the three patients | none | Confirmation of breakpoint calls (specificity) | — |
| SNP-array genotyping and copy number analysis | Tumor biopsies per patient | none | B-allele frequencies, copy number variation, tumor purity (ASCAT) | — |
| RNA sequencing (transcriptional profiling) | Each tumor biopsy from the three patients | none | Differentially expressed genes, ovarian cancer subtype signature concordance, mutant allele frequencies in transcripts, fusion gene expression | — |
| Histopathological examination | Tumor biopsy tissue sections (serous adenocarcinoma in patients 1 and 3; carcinosarcoma in patient 2) | none | Histology and tumor content percentage per sampling site | — |
| Pathway/network enrichment analysis of top 5% most significantly differentially expressed genes | RNA-seq profiles of tumor biopsies per patient | none | Branch-specific signaling pathway activation (WNT, integrin, chemokine, Hedgehog) | Cytoscape |
- – Between 120 and 369 somatic genomic rearrangements were detected across primary and metastatic tumor samples of the three patients 120–369 rearrangements
- – In patient 1 only 2 of 369 somatic breakpoints were shared by all samples, while patient 3 showed the vast majority shared across five to seven of seven tumor samples 2/369 (patient 1) vs 34/120 shared in all nine samples (patient 2)
- – Hierarchical clustering of breakpoints split patients 1 and 2 into two clusters separating omentum/peritoneum biopsies from ovary (and, for patient 2, pelvis) biopsies; patient 3 was homogeneous
- – Branch-specific shift in rearrangement classes in patient 1: omental/peritoneal metastases had higher tandem duplication and inversion fractions and fewer deletions and interchromosomal events than ovarian tumors, despite shared BRCA status deletions 40% overall in patient 1
- – Two distinct driver TP53 missense mutations at distinct tumor locations in patient 1; only right ovary samples contained both, with I195N at low DNA frequency I195N 1%–9% vs P278L 33%–77%
- – Kataegis in patient 2: 12 of 17 mutations were in FANCD2, confined to samples p2.VI-1 and p2.VI-2, within a 1.2 kb window at TpCpX trinucleotides and coinciding with a breakpoint unique to those samples 12 mutations within 1.2 kb
- – Patient 1 carried 63 somatic single-nucleotide mutations versus 17 per patient in patients 2 and 3; 19 mutations were private to a single ovary tumor sample in patient 1 while none of the four omentum/peritoneum metastases carried private mutations 63 vs 17 mutations; 19 private
- – Biopsies of patient 1 split by expression subtype along the same branches: omentum/peritoneum samples showed C1 (high stromal) overlapping C2 (high immune), ovary samples fell into C4 (low stromal); patient 2 showed the C5 mesenchymal signature; 1000–1300 differentially expressed genes per sample 1000–1300 DE genes per sample
- count 95 out of 121 tested breakpoints confirmed (>78% specificity) (PCR validation of mate-pair breakpoint calls)
- count 120–369 somatic genomic rearrangements (Range across the three patients' tumor samples)
- count 2/369 breakpoints shared between all samples (Patient 1, eight tumor samples)
- count 34/120 breakpoints shared between all nine tumor samples (Patient 2)
- count 63 somatic single-nucleotide mutations in patient 1; 17 mutations per patient in patients 2 and 3 (Coding screen of 2099 cancer genes)
- other I195N detected at 1%–9% versus 33%–77% for P278L (TP53 mutation frequencies in patient 1 right ovary samples)
- other median physical genomic coverage ∼50×; genotyping of a heterozygous breakpoint in ≥5% and 14% of tumor cells at 90% and 30% tumor percentage (Mate-pair sequencing sensitivity)
- count 1000–1300 differentially expressed genes per sample versus all other samples of the same patient (RNA sequencing per patient)
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 study is a descriptive genomic and transcriptomic case series of three treatment-naïve, advanced-stage epithelial ovarian cancer patients, using whole-genome mate-pair sequencing, SNP-array genotyping, targeted resequencing, and RNA sequencing across multiple tumor biopsies per patient. Findings were reported mainly through breakpoint/mutation counts, percentages, unsupervised hierarchical clustering of breakpoints, SNP allele frequencies and expression profiles, and subtype concordance scoring, rather than through classical inferential hypothesis testing with p-values.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| PCR-based validation (reported as a specificity/confirmation rate) | validation of somatic breakpoint calls from mate-pair sequencing | 121 breakpoints tested (95 confirmed) | not stated |
| Unsupervised hierarchical clustering | breakpoint junctions per patient, SNP-array allele frequencies, and RNA-seq expression differences across biopsies | per-patient biopsy sets (e.g., patient 1: 8 samples; patient 2: 9 samples; patient 3: 7 samples) | not stated |
| Pairwise differential expression comparison (each sample vs. all other samples of the same patient) | RNA sequencing across tumor biopsies | 27 tumor biopsies across 3 patients | not stated |
| Transition/Transversion (Ti/Tv) ratio comparison | comparison of ovarian-site samples vs. omentum/peritoneum metastasis samples in patient 1 | — | not stated |
| Percentage concordance scoring against reference expression signatures | classification of tumor biopsies into six previously described ovarian cancer expression subtypes (Tothill et al.) | 1500 marker genes | not stated |
-
Differentially expressed genes were identified by comparing each sample to all other samples from the same patient, without a stated statistical test or multiple-testing correction.↳ Could also: A formal RNA-seq differential expression framework (e.g., DESeq2 or edgeR using a Wald or likelihood-ratio test) combined with a false-discovery-rate correction such as Benjamini-Hochberg — This would attach explicit p-values and FDR-adjusted q-values to each gene, giving readers a direct sense of the statistical confidence behind the reported differentially expressed genes across the many pairwise sample comparisons.
-
Relationships among biopsies (breakpoints, SNP allele frequencies, RNA expression) were assessed through unsupervised hierarchical clustering and visual inspection of heat maps.↳ Could also: Clustering supported by quantitative robustness measures, such as bootstrap resampling, silhouette scores, or a model-based tumor phylogenetics approach (e.g., maximum-likelihood or Bayesian phylogenetic inference) — This could complement the visual branching patterns with a numeric estimate of how strongly the inferred sample groupings are supported by the data.
-
PCR-based validation of predicted breakpoints was summarized as a single specificity percentage (95/121, >78%).↳ Could also: Reporting the same estimate together with a confidence interval for the proportion (e.g., a Clopper-Pearson exact interval) — This would convey the precision of the specificity estimate given the number of breakpoints tested, which is useful information alongside a point estimate based on a moderate sample size.
-
The difference in Transition/Transversion (Ti/Tv) ratio between the two branches of patient 1 was described qualitatively as suggestive of distinct mutational forces.↳ Could also: A formal statistical comparison of the transition/transversion counts between branches, such as a chi-square or Fisher's exact test — This would let readers evaluate whether the observed difference in Ti/Tv ratio between branches is larger than would be expected from sampling variation alone.
-
Conclusions about intra-tumor heterogeneity are drawn from detailed within-patient comparisons across biopsies in a cohort of three patients.↳ Could also: A mixed-effects or hierarchical statistical model treating patient as a random effect, applicable if the design were extended to a larger cohort — This could explicitly separate within-patient (biopsy-level) heterogeneity from between-patient variability, which can help generalize heterogeneity findings beyond individual cases.
-
Tumor subtype assignment was based on percentage concordance with reference gene expression signatures from a prior study.↳ Could also: A formal classifier producing class probabilities or discriminant scores (e.g., nearest-centroid or machine-learning classification with permutation-based significance testing) — This could provide a statistical confidence measure for each subtype call, complementing the descriptive concordance percentages currently reported.
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.
What deviates: the one executable claim reproduced qualitatively but not exactly — 17 large (>=0.97Mb) subset-specific copy-number events (9 in P1, 8 in P2) against the paper's reported 14, derived from a fully independent rebuild off raw GSE47633 SNP6 intensities. Whose side: the ~21% count gap is predominantly ours (DNAcopy CBS substituted for ASCAT, no purity/ploidy correction, only 81,218/300,695 = 27.0% of probes mappable, self-invented |logR|>=0.3 and recurrence filters), compounded by an authors'-side underspecification — the amplitude and 'subset-specific' criteria behind '14' are nowhere stated, so the exact number is not derivable from the deposit. The real reproducibility defect is availability, not validity: the paper's own SV caller at github.com/Vityay/1-2-3-SV has been deleted (404), which drops SV breakpoints, kataegis and gene fusions outright, and the targeted-panel and full multi-region RNA-seq DE inputs were never located — 5 of 8 claims unattemptable. Severity: moderate. The central plasticity conclusion is independently confirmed with no flip in direction or significance, so this is a solid reproduction with explainable deviations over a badly shrunken claim surface, not a discrepancy against the science.
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.