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

Cancer-predicting transcriptomic and epigenetic signatures revealed for ulcerative colitis in patient-derived epithelial organoids.

Oncotarget · 2018
L1 65/100 PQI 91
Why this verdict

The main results reproduced, with only marginal, non-material deviations.

Reproduced on the brainbox compute brainarbeit.com
Scoring basis — itemised

Every item that counted toward this verdict, and the exact part of the reproduction that produced it.

Supporting (toward a concern)
Content-critical question only partially held
+2 pts
From: Q5 · Derivability / plausibility 🟡
Content-critical question only partially held
+2 pts
From: Q7 · Core claim 🟡
Content-critical question only partially held
+2 pts
From: Q8 · Severity of the miss (overall human judgment) 🟡
Minor / cosmetic deviation
+1 pts
From: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q6 · Severity of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +7
✓ What held up
  • Same input data as the authors
What did not (or only partly)
  • 🟡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
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
65/100
Reproducibility score
0.5 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 27% of all assessed papers rank 843 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

Described well enough for a PARTIAL 1:1. GSE102746 is RNA-seq only (20 samples, 10 UC + 10 normal organoids) and ships the authors' processed Cufflinks FPKM table; the listed repo (TrimGalore) is a third-party trimming tool that yields no comparable result on its own, so we reproduced the differential-expression call from the authors' shipped FPKM table using their own thresholds (FDR<0.05 & |log2FC|>1.5). RESULT: the data structure reproduces EXACTLY (35786 genes x 20 samples, groups verified) and the directional asymmetry reproduces ROBUSTLY (reported 84% up; we get 82-85% up). The absolute count 260 does NOT reproduce from the FPKM table with a standard Welch t-test (we get 9 strict / 40 nominal-p / 90 FC-only / 756 FDR-only) -- expected because the paper used Cuffdiff on BAM alignments, a different and less-conservative model than a t-test on FPKM; this is method divergence, NOT a fabrication signal (756 genes pass FDR<0.05 before the FC filter, so a count in the hundreds is plausible under Cuffdiff). NOT ATTEMPTED (the hard ~20%): re-running TopHat2+Cuffdiff from raw FASTQ (SRP115575) to hit 260 exactly; subset-B DE (1,048; sample membership unspecified); ChIP-seq H3K27ac (1,485 sites; raw data not in GSE102746). No fabrication indicated.

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 65
    assessed: 2026-06-14 ⛓ 0e89c82fa43a
✎ 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
2026-06-14
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-15
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: opus
Founding hypothesis

Do patient-derived primary colonic organoids from ulcerative colitis (UC) patients faithfully recapitulate primary UC tissue at histologic, transcriptomic, and epigenetic levels, and can such models reveal precancerous (colitis-associated cancer) molecular signatures already activated in UC?

Core claims
  • UC patient-derived epithelial organoids histologically phenocopy primary UC tissue, while non-IBD organoids resemble healthy colonic epithelium. finding
  • Whole-transcriptome profiling shows upregulation of inflammatory, metabolism, cell adhesion, and cancer pathways in UC organoids relative to non-IBD organoids. finding
  • H3K27ac enhancer profiling reveals UC organoid enrichment for gastrointestinal/digestive cancer pathways and oncogenic markers including S100P. finding
  • LYZ and NPSR1 are identified as novel markers for GI cancer, enriched in UC organoids at both transcriptomic and epigenetic levels. finding
  • Immunolocalization shows increased LYZ, S100P, and NPSR1 protein levels in UC and colitis-associated cancer (CAC). finding
  • Patient-derived primary colonic organoids constitute a faithful human-derived model suitable for dissecting UC and CAC pathogenic mechanisms. resource
  • UC organoids and tissues already exhibit an oncogenic/precancerous signature validated at histologic, transcriptomic, and epigenetic levels. mechanism
  • Integration of RNA-Seq and H3K27ac ChIP-Seq enables identification of susceptibility loci that may serve as functional/mechanistic intervention targets. method
Experimental setups
Assay System Perturbation Readout Platform
Histology / H&E and Alcian Blue–PAS staining Patient-derived colonic epithelial organoids (UC and non-IBD) none (disease vs control comparison) Tissue architecture, mucus content, epithelial organization
Immunohistochemistry / Immunocytochemistry Patient-derived colonic organoids and matched primary tissues (UC, non-IBD, CAC) none Protein levels of MUC2, Ki-67, chromogranin A, CK19, vimentin, LYZ, S100P, NPSR1
Bulk RNA-Seq (whole-transcriptome) Patient-derived colonic epithelial organoids (10 UC, 10 non-IBD) none (UC vs non-IBD) Differential gene expression / pathway enrichment (MSigDB, GO, KEGG, Hallmark, Reactome)
ChIP-Seq for H3K27ac (active enhancers) Patient-derived colonic epithelial organoids (5 UC, 5 non-IBD; Ctd150 excluded) none (UC vs non-IBD) Genome-wide H3K27ac enrichment / enhancer-associated gene sets ENCODE normal colonic mucosa used as reference track
Short tandem repeat (STR) analysis Patient-derived organoid isolates none Verification of unique patient origin
Genomic Regions Enrichment of Annotations Tool (GREAT) ontology analysis H3K27ac ChIP-Seq peak regions from organoids none Gene ontology / disease term enrichment
Key results
  • 260 transcripts differentially expressed between UC and non-IBD organoids (full cohort) 219 (84%) up, 41 (16%) down
  • Subset B (4 UC vs 4 non-IBD) showed larger differential expression 1,048 transcripts; 864 (82%) up, 184 (18%) down
  • MUC2-positive goblet cells reduced in UC organoids two-fold decrease
  • H3K27ac region-specific sites identified comparing UC vs non-IBD organoids 1,485 sites; 97 within 5 kb of TSS, 894 located 50–500 kb away
  • GREAT ontology enrichment for gastrointestinal neoplasm p=2.85x10^-11
  • GREAT ontology enrichment for digestive system cancer p=7.08x10^-11
  • NPSR1 showed the largest log2 fold-change in the entire RNA-seq cohort and was enriched for H3K27ac
  • LYZ and S100P upregulated in RNA-Seq and enriched for H3K27ac in UC organoids
Key statistics
  • count 260 differentially expressed transcripts (FDR<0.05, log2FC>1.5) (Full cohort UC vs non-IBD RNA-Seq)
  • count 1,048 differentially expressed transcripts (Subset B (4 UC vs 4 non-IBD) RNA-Seq)
  • count 1,485 region-specific H3K27ac sites (ChIP-Seq UC vs non-IBD organoids (5 vs 5))
  • pvalue p=2.85x10^-11 (GREAT enrichment for gastrointestinal neoplasm)
  • pvalue p=7.08x10^-11 (GREAT enrichment for digestive system cancer)
  • fold_change two-fold decrease in MUC2-positive goblet cells (UC vs non-IBD organoids)
  • mean 42.3 ± 14.5 years (UC age); disease duration 9.9 ± 9.7 years (UC patient cohort characteristics)
  • mean 68.5 ± 7.6 years (Non-IBD patient age at presentation)

Statistical methods review

Model: opus

A 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 an observational, cross-sectional genomics study comparing patient-derived colonic organoids (and matched tissues) from ulcerative colitis (UC) versus non-IBD patients using histology, RNA-Seq, and H3K27ac ChIP-Seq. Differential expression and differential acetylation were assessed using fold-change and false-discovery-rate thresholds (log2 fold change > 1.5, FDR < 0.05, Benjamini-Hochberg), with results visualized as volcano/MA plots and principal-component analyses; pathway-level interpretation used gene-set enrichment against MSigDB and region-based ontology via GREAT. Histologic findings were reported descriptively with counts of replicates (e.g., n > 20, n ≥ 3).

Replicationbiological Sample sizeCohort sizes given as counts (10 UC, 10 non-IBD; ChIP-Seq 5 vs 5; Subset B 4 vs 4; histology n > 20 or n ≥ 3); no formal power/sample-size calculation described GroupsUC vs non-IBD organoids/tissues (plus CAC for immunostaining) Pairingunpaired Randomization/blindingnot stated DispersionSD Exact p-valuesyes Effect sizesyes Confidence intervalsno Multiplicity correctionBenjamini-Hochberg FDR
Statistical tests used
Test Applied to n Assumptions
Differential gene-expression analysis with FDR/Benjamini-Hochberg thresholding (RNA-Seq), reported as log2 fold change and FDR Volcano plots and gene lists for full cohort (260 transcripts) and Subset B (1,048 transcripts), UC vs non-IBD organoids (Figure 2A, 2E) full cohort 10 UC vs 10 non-IBD; Subset B 4 UC (Ctd111,139,153,155) vs 4 non-IBD (NL141,143,148,156) not stated
Differential H3K27ac enrichment analysis with FDR thresholding (ChIP-Seq), visualized as MA/volcano plot H3K27ac UC vs non-IBD organoids (Figure 3A), FDR < 0.05 5 non-IBD and 5 UC organoid isolates (Ctd150 excluded) not stated
Gene-set enrichment analysis (MSigDB) Pathway enrichment of up/downregulated genes, full cohort and Subset B (Figure 2C,2D,2G,2H) and ChIP-seq peaks (Figure 3B) na
Principal-component analysis Clustering of UC vs non-IBD organoid expression profiles (Figure 2B, 2F) na
Region-based ontology enrichment (GREAT), reported with p-values H3K27ac-enriched regions; gastrointestinal neoplasm (p=2.85x10^-11) and digestive system cancer (p=7.08x10^-11) (Supplementary Figure 2) not stated
Approaches that could also have been used
  • Differentially expressed genes were defined using fixed thresholds (log2 fold change > 1.5 and FDR < 0.05).
    Could also: Reporting shrunken effect-size estimates and/or formal significance from a count-based model (e.g., DESeq2/edgeR/limma-voom) alongside the thresholds. — Model-based effect-size shrinkage and explicit per-gene statistics can convey both magnitude and uncertainty, which is helpful with modest per-group sample sizes.
  • A 'Subset B' of four UC isolates was selected based on PCA clustering for a focused differential-expression analysis.
    Could also: Presenting both the full-cohort and subset analyses with the selection rule pre-specified, or modeling heterogeneity directly (e.g., including covariates or a mixed model) rather than subsetting. — Showing how results depend on grouping choices, and modeling variability rather than excluding it, can make the heterogeneity of UC easier to interpret.
  • Sample sizes were described as counts without an accompanying power or sample-size calculation.
    Could also: Reporting an a priori or post hoc power consideration, or confidence intervals around key effect estimates. — An explicit power statement or interval estimates communicates the precision achievable with the available organoid cohorts.
  • Patient characteristics were summarized as mean ± SD.
    Could also: Adding median and IQR or range, and 95% confidence intervals for group means. — For small cohorts, medians/IQR and intervals can complement SD by conveying both central tendency and the range of spread.
  • Group differences in organoid features (e.g., MUC2-positive goblet cells described as a two-fold decrease) were presented descriptively.
    Could also: Pairing the descriptive comparison with a formal nonparametric test (e.g., Mann-Whitney U) and reporting the n and exact p-value. — An accompanying inferential statistic with stated n would quantify the comparison alongside the descriptive observation.
  • Multiplicity was controlled via Benjamini-Hochberg FDR for the genome-wide screens.
    Could also: Also stating the correction approach used within the pathway/enrichment analyses (MSigDB, GREAT). — Documenting how multiple comparisons are handled at the pathway level, in addition to the gene level, clarifies the full multiplicity scope.
Software: MSigDB (Molecular Signatures Database) gene-set enrichment · GREAT (region-based ontology tool)

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.

Citations
38
Impact: medium
Foundation confidence
None of its references are in our reproducibility record yet — its foundation cannot be assessed.
Topics

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.

GSE102746 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet

What was reproduced

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

Scope — pmid-29983891

Paper: Sarvestani et al. 2018, Oncotarget — "Cancer-predicting transcriptomic and epigenetic signatures revealed for ulcerative colitis in patient-derived epithelial organoids." DOI 10.18632/oncotarget.25617.

Pipelines named in Methods

analysis pipeline data shipped?
RNA-seq trimming/QC TrimGalore (the listed repo) — (preprocessing only)
RNA-seq alignment + quantification Tuxedo (TopHat/Cufflinks), GRCh38 yes — FPKM table on GEO
RNA-seq differential expression Tuxedo / Cuffdiff, FDR<0.05 & |log2FC|>1.5 Cuffdiff output NOT shipped (FPKM is)
ChIP-seq (H3K27ac) FastQC + TrimGalore + Bowtie2 (GRCh19) + MACS2 + DESeq2 raw ChIP data NOT in GSE102746
qRT-PCR / IHC validation wet-lab n/a

In scope (attempted)

  • C3 Processed data structure: shipped FPKM table, 20 samples (10 UC + 10 normal). Directly verifiable.
  • C1 Full-cohort DE transcript count (reported 260). Re-derived from the FPKM table.
  • C2 Direction of DE (reported 84% up). Re-derived from the FPKM table.

Out of scope (not attempted — 80/20 and data limits)

  • C4 Subset-B DE (1,048 transcripts): the subset's sample membership is not specified in the paper or GEO, so the comparison is not pinnable.
  • C5 ChIP-seq H3K27ac (1,485 sites): the raw ChIP-seq data is not part of GSE102746 and no separate accession is given; cannot obtain input.
  • Re-running TopHat2+Cuffdiff from raw FASTQ (SRP115575) to reproduce 260 exactly: the hard ~20% — intentionally skipped per the 80/20 rule.

Reproduction approach

The listed repo (TrimGalore) is a third-party trimming tool that produces no comparable result value on its own. The clear, low-hanging pipeline-derived result is the DE call. We therefore re-derived differential expression from the authors' own shipped Cufflinks FPKM table using the authors' own stated thresholds (FDR<0.05 & |log2FC|>1.5). This is a method-divergent reproduction on the authors' own data (the paper's count used Cuffdiff on BAM alignments, which the FPKM table cannot reconstruct).

Figures / tables: Figure 2ETablesFig 2Table
C3-data-structure
Reported
FPKM table, 20 samples (10 UC + 10 normal organoids), GRCh38
Reproduced
35786 genes x 20 samples; 10 CTd=UC + 10 NL=normal, confirmed vs series-matrix status
exact
C2-direction
Reported
84% of DE transcripts upregulated in UC (219/260)
Reproduced
82.2-85% upregulated across filters (0.822 at |log2FC|>1.5; 0.85 at p<0.05 & |log2FC|>1.5)
within tolerance
C1-de-count
Reported
260 DE transcripts at FDR<0.05 & |log2FC|>1.5
Reproduced
9 (FDR<0.05 & |log2FC|>1.5); 40 (p<0.05 & |log2FC|>1.5); 90 (|log2FC|>1.5 only); 756 (FDR<0.05 only)
did not match

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 65/100

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)
Scoring basis — itemised

Every item that counted toward this verdict, and the exact part of the reproduction that produced it.

Supporting (toward a concern)
Content-critical question only partially held
+2 pts
From: Q5 · Derivability / plausibility 🟡
Content-critical question only partially held
+2 pts
From: Q7 · Core claim 🟡
Content-critical question only partially held
+2 pts
From: Q8 · Severity of the miss (overall human judgment) 🟡
Minor / cosmetic deviation
+1 pts
From: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q6 · Severity of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +7

Re-derived DE from the authors' own shipped FPKM table using their stated thresholds. The data structure is exact (35786 genes x 20 samples; 10 UC + 10 normal verified) and the directional asymmetry reproduces robustly (84% up reported vs 82-85%). The headline 260-DE-transcript count does NOT reproduce with a Welch t-test (9 strict / 756 FDR-only) — because the paper used Cuffdiff on BAM alignments, a less-conservative model than a t-test on FPKM. This is a method divergence on our side, not a fabrication signal: 756 genes pass FDR before the FC filter, so a few-hundred count under Cuffdiff is plausible. Reproducing 260 exactly needs TopHat2+Cuffdiff from raw FASTQ (the deliberately-skipped hard 20%). Note the listed code link (TrimGalore) is a harvest false-positive, not the analysis pipeline.

🤝
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.

91.4 k
tokens (I/O) · 6.3 M incl. cache
14 min
runtime · 0 CPU-h
0.3 GB
peak RAM
1
HPC jobs
hummel
machine