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

Indoleamine 2, 3-Dioxygenase 1 and CD8 Expression Profiling Revealed an Immunological Subtype of Colon Cancer With a Poor Prognosis.

Front Oncol · 2020
L1 48/100 PQI 83
⚑ Flagged for review — a reproduced result did not match the reported value

Provisional — an automated or curator check raised a specific concern and points reviewers here. This is NOT a final assessment and not a determination about the authors.

Why this verdict

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”.

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: Q3 · Location of the main deviation 🟡
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 +8
✓ What held up
  • Nothing in this column.
What did not (or only partly)
  • 🟡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
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
48/100
Reproducibility score
1.5 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 6% of all assessed papers rank 1092 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 to attempt: yes for the two named public datasets (GSE17538, TCGA-COAD via Xena); the cited 'code' (corrplot) is only a plotting package (P16). PARTIAL reproduction. The central thesis -- a CD8high/IDO1high subtype (group IV*) with poor prognosis -- reproduces in DIRECTION and statistical significance in both cohorts (TCGA 4-group OS P0.04; GSE17538 OS P0.015) under the paper's high IDO1 'secondary-peak' cutoff, and the IDO1~CD274(PD-L1) correlation reproduces cleanly (r=0.79 vs 0.73). HOWEVER the result is highly cutoff-fragile: group IV* is a tiny high-IDO1 tail (n=2-14) and the survival signal vanishes under a balanced/median split (P=0.29-0.58). GSE17538 DFS (0.015) did NOT reproduce as significant (best 0.08); the absolute IDO1 cutoff 9.576 is not recoverable from standard Xena log2-RSEM normalization (peaks at 7.08/14.34; paper used n=438 vs Xena's 286 tumors). NOT attempted (out of scope): web-database figures (GEPIA/TISIDB/TIMER/TIMER2.0), GSEA pathways, Fig7/8 immune-gene/E-cadherin/CMS/MSI/ROC sub-analyses, and the in-vivo murine model -- none are runnable local pipelines on public data. Three possible-discrepancy flags raised for human audit.

💻 Code ↗ 🗄 Data: GSE17538

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 48
    assessed: 2026-06-15 ⛓ be6185f54dd1
✎ 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
2026-06-15
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

Whether combining IDO1 expression with CD8A (CD8 T cell) infiltration levels yields independent prognostic and predictive significance for colon cancer beyond CD8 T cell density or CMS classification alone.

Core claims
  • The CD8A-high/IDO1-high colon cancer subtype (group IV*) has the worst survival despite high CD8 infiltrates. finding
  • IDO1/CD8A stratification is an independent prognostic factor of overall survival and a useful predictive biomarker in colon cancer. finding
  • The poor prognosis of the CD8A-high/IDO1-high group is associated with high immune response, immune checkpoint genes, and Th1/IFN-γ signatures, regardless of CMS classification. mechanism
  • IDO1 expression is strongly positively correlated with tumor-infiltrating lymphocytes, especially CD8 T cells, in colon adenocarcinoma. finding
  • IDO1 expression is strongly correlated with CD274 (PD-L1) expression in colon cancer. finding
  • A combined IDO1/CD8A classifier provides a tool for risk stratification to guide immunotherapy in colon cancer. resource
  • An in vivo murine liver-metastasis model validated that the CD8A-high/IDO1-high group correlates with late-stage metastasis and upregulation of immune checkpoints. method
Experimental setups
Assay System Perturbation Readout Platform
bulk RNA-seq expression and survival analysis TCGA-COAD cohort (human colon adenocarcinoma, 438 patients) none IDO1/CD8A/checkpoint gene expression and overall survival Illumina HiSeq and Illumina GA (RSEM values, UCSC Xena)
gene expression survival validation NCBI-GEO GSE17538 cohort (human colon cancer, 232 patients) none survival stratified by IDO1/CD8A expression
immune infiltration / TIL correlation analysis human colon adenocarcinoma (TCGA) none correlation of IDO1 with 28 TIL types and immune cell abundance TISIDB, TIMER/TIMER2.0, GEPIA databases
gene set enrichment analysis (GSEA) human colon cancer (group IV* vs others) none enriched Hallmark and KEGG immune pathways (NES, NOM P) GSEA, MSigDB v7.1
in vivo liver metastasis model with bioluminescence imaging C57BL/6J female mice with murine SL4 colon cancer cells (intrasplenic injection) tumor cell inoculation (3×10^6 SL4 cells) liver metastatic burden / photon flux at day 4 and day 12 IVIS Spectrum CT System (Xenogen), LivingImage software
Real-time quantitative PCR SL4 mouse liver metastatic foci tissue none mRNA of IDO1, CD8A, PD-1, PD-L2, TIM3, CD276, CD200, CD160 (normalized to GAPDH) ABI 7900 HT, SYBR Green Mix (TaKaRa)
Key results
  • CD8A-high/IDO1-high subtype had the worst overall survival despite high CD8 infiltrates
  • Higher IDO1 expression indicated improved OS using the first (highest peak) cut-off value
  • IDO1 associated with poor OS using the secondary cut-off value (trend, not significant) HR=1.621 (95% CI 0.801–3.281), P=0.101
  • IDO1 strongly correlated with CD274 (PD-L1) expression in TCGA COAD r=0.733, P<0.01
  • IDO1 correlated with effector memory CD8 T cells in COAD r=0.730
  • IDO1 correlated with activated CD8+ T cells in COAD r=0.710
  • IDO1 correlated with Th1 cells in COAD r=0.666
  • IDO1 correlated with Tfh (T follicular helper) cells in COAD r=0.664
Key statistics
  • correlation r=0.733, P<0.01 (IDO1 vs CD274 (PD-L1) in TCGA COAD)
  • correlation r=0.730 (IDO1 vs effector memory CD8 T cells (Tem CD8) in COAD)
  • correlation r=0.710 (IDO1 vs activated CD8+ T cells (Act CD8) in COAD)
  • correlation r=0.666 (IDO1 vs Th1 cells in COAD)
  • correlation r=0.664 (IDO1 vs Tfh cells in COAD)
  • other HR=1.621 (95% CI 0.801–3.281), P=0.101 (OS for IDO1 high vs low using secondary cut-off)
  • count 438 (TCGA colon cancer patients analyzed)
  • count 232 (GSE17538 validation cohort patients)

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 stratified two retrospective public cohorts (TCGA-COAD, n=438; GEO GSE17538, n=232) by IDO1 and CD8A expression, with thresholds selected by log-rank score maximization. Survival differences between the four resulting subgroups were assessed with Kaplan-Meier curves and log-rank tests, and the stratification was evaluated as an independent prognostic factor via univariate and multivariate Cox proportional hazards regression. Group-level transcriptomic differences were characterized by GSEA, Spearman correlations, t-tests/Wilcoxon/ANOVA depending on distributional properties, and immune deconvolution tools; findings were validated in a murine splenic-injection liver-metastasis model using RT-PCR.

Replicationbiological Sample size438 TCGA-COAD patients and 232 GEO GSE17538 patients with complete OS data and clinicopathological parameters selected by exclusion criteria; animal group sizes not stated in the text; no formal power calculation described for either cohort or animal experiment GroupsFour IDO1/CD8A expression subgroups (I–IV*) in colon cancer; secondary comparisons across CMS subtypes and clinical stage subgroups Pairingunpaired Randomization/blindingnot stated DispersionCI Effect sizesyes Confidence intervalsyes Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Log-rank test Kaplan-Meier OS, DSS, and DFS comparisons between IDO1/CD8A stratification subgroups 438 (TCGA-COAD); 232 (GEO GSE17538) not stated
Univariate Cox proportional hazards regression Screening of clinical and molecular variables for association with OS 438 (TCGA-COAD); 232 (GEO GSE17538) not stated
Multivariate Cox proportional hazards regression Independent prognostic factor analysis for OS; IDO1 expression excluded from the model due to collinearity with risk groups 438 (TCGA-COAD); 232 (GEO GSE17538) not stated
ROC curve analysis with binary logistic regression (AUC) Evaluating combined predictive efficacy of IDO1 stratification with other variables to distinguish survival status; AUC > 0.7 and P < 0.05 as thresholds not stated
Spearman's correlation IDO1 vs CD274 (TCGA COAD, r=0.733); IDO1 vs 28 TIL subtypes (TISIDB); IDO1 vs CD8 T cell marker genes (pan-cancer via TIMER2.0) 438 (TCGA-COAD); pan-cancer sample sizes from external databases not stated
Unpaired t-test or Welch's t-test (selected based on homogeneity of variance) Comparisons between risk groups where normal distribution was assumed stated
Wilcoxon rank-sum test Comparisons between risk groups when data did not meet the normal distribution assumption stated
One-way ANOVA Comparisons involving more than two risk groups not stated
Fisher's exact test Categorical discrepancy between two groups in fourfold tables with total sample size < 40 stated
GSEA with permutation-based normalized enrichment score (NES) and nominal P-value Functional enrichment analysis of Hallmark and KEGG Canonical Pathway gene sets; 1,000 permutations; NOM P < 0.05 as significance threshold not stated
ComBat batch correction (sva R package) Normalization of TCGA RNA-seq expression values across IlluminaHiSeq and IlluminaGA platforms; effect verified by PCA 438 not stated
Approaches that could also have been used
  • Optimal IDO1 and CD8A cutoffs were selected by maximizing the log-rank test statistic in the same data used for survival analysis (cutp, survMisc), with a secondary peak used to define a third high-risk tier
    Could also: Pre-specified thresholds (e.g., median, published clinical cutoffs, or quartiles defined in a held-out training set) could also define the expression groups — Selecting a cutoff by maximizing the log-rank statistic in the same dataset inflates the Type I error rate because the split is optimized post hoc; pre-specified or cross-validated thresholds maintain the nominal error rate and transfer more directly to independent cohorts
  • GSEA significance was declared using the permutation-based nominal P-value (< 0.05) without reporting FDR q-values across the full set of tested gene sets
    Could also: Reporting GSEA FDR q-values (Benjamini-Hochberg, as provided natively by the GSEA tool) alongside or instead of nominal P-values could also summarize enrichment significance — Nominal P-values from GSEA do not account for the simultaneous testing of many gene sets; the FDR q-value is the metric the Broad Institute documentation recommends for interpreting results when multiple gene sets are evaluated
  • Multiple pairwise and multi-group comparisons across four subgroups and multiple immune gene sets were performed without explicit family-wise error rate or false discovery rate correction
    Could also: A Benjamini-Hochberg FDR adjustment or Bonferroni correction applied across the family of simultaneous comparisons could also control the error rate — When many tests are performed simultaneously, explicit multiplicity control makes the expected false-positive rate transparent and is standard practice in high-dimensional gene expression settings
  • The ROC curve was constructed by combining IDO1 stratification with other variables via binary logistic regression, summarized by AUC, with survival status treated as a fixed binary outcome
    Could also: Time-dependent ROC analysis (e.g., R packages timeROC or survivalROC) could also assess discrimination for a censored survival endpoint at a specified time horizon — Standard logistic regression ROC does not account for censoring or the time at which the event occurs; time-dependent ROC methods handle censored survival data and yield horizon-specific AUC estimates that are more appropriate for an overall-survival endpoint
  • The proportional hazards assumption underlying the Cox regression models was neither tested nor reported
    Could also: Testing the proportional hazards assumption via Schoenfeld residuals (cox.zph in R) and, when violated, using restricted mean survival time (RMST) or a stratified Cox model could also be applied — RMST does not require the proportional hazards assumption and provides an interpretable survival difference in absolute time units; reporting the assumption test allows readers to assess whether the HR is a valid summary across the follow-up period
  • Immune infiltration was estimated using two external web-based deconvolution tools (TIMER and TISIDB), and associations were summarized with Spearman correlations without cross-method comparison
    Could also: Applying additional bulk RNA-seq deconvolution methods (e.g., CIBERSORT, quanTIseq, or MCP-counter) to the same TCGA data could also estimate immune cell composition — Different deconvolution algorithms use different reference signatures and assumptions; consistency of infiltration estimates across multiple methods strengthens confidence in the reported immune cell abundance relationships
Software: R 4.0.0 · SPSS 25.0 · GraphPad Prism 8.0 · R/survMisc 0.5.5 · R/ComplexHeatmap 2.4.2 · R/corrplot 0.84 · R/sva (ComBat) · R/ggplot · GSEA (Broad Institute)

Result convergence & founder nodes

Findings this paper shares with others that ran a comparable experiment. A node’s strength is how many independent papers report it (replication breadth) — not how often it is cited, so a heavily-replicated but under-cited founder still stands out.

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
23
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.

GSE17538 GEO in Methods (http://purl.org/orb/Methods)
also used by 2 papers:

Downstream reach in the literature

100 downstream papers · 1 datasets

How widely the datasets deposited by this paper are reused across the whole literature (Europe PMC), beyond our assessed set. This is a factual dependency map — reusing a public dataset is normal, good science. It is not a judgement on the downstream papers; the only verdict here is this paper's own, with its cited rationale.

This paper is currently under reproducibility review (see the verdict above). The map below shows where the data in question has propagated — so reuse can be traced, not so the downstream work is presumed affected.

What was reproduced

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

Figures / tables: Fig 4GFig 4HFig S2Fig 4CFig 1DFig 5B
R1a
Reported
GSE17538 OS group IV* vs III* P=0.010 (Fig4G)
Reproduced
antimode P=0.0147 (IV* n=8); balanced split P=0.289
within tolerance
R1b
Reported
GSE17538 DFS group IV* vs III* P=0.015 (Fig4H)
Reproduced
best faithful P=0.0815 (none significant)
did not match
R2
Reported
TCGA-COAD IDO1~CD274 r=0.733, P<0.01 (SuppFigS2)
Reproduced
Pearson r=0.792 (P=1.1e-62); Spearman 0.784; n=286
within tolerance
R3
Reported
TCGA-COAD four-group OS P=0.032 (Fig4C)
Reproduced
antimode P=0.0415 (IV* n=3); balanced P=0.5815
partial
R4
Reported
TCGA-COAD IDO1 secondary-peak cutoff ~9.576 (Fig1D)
Reproduced
density peaks 7.083/14.343; secondary=14.343
did not match
R5
Reported
TCGA-COAD Cox group IV* HR=3.016 (1.081-8.416) P=0.035, n=295 (Fig5B)
Reproduced
HR=6.079 (1.415-26.112) P=0.0152, n=185
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 48/100

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: Q3 · Location of the main deviation 🟡
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 +8

The core conclusion — a CD8high/IDO1high colon-cancer subtype (group IV*) with poor prognosis — reproduces in direction and significance in both cohorts (TCGA 4-group OS P0.04, GSE17538 OS P0.015) and the IDO1–PD-L1 correlation reproduces cleanly (r=0.79 vs 0.73), so this is not a fabrication case. However the result is strongly cutoff-fragile: group IV* is a tiny high-IDO1 tail (n=2-14) and the survival signal disappears under a balanced split (P=0.29-0.58). The deviations sit mainly on the input/preprocessing and self-chosen-cutoff side — the absolute cutoff 9.576 is not recoverable from public Xena normalization and the cohort (n=286 vs 438) is redefined — with one genuine non-reproduction (GSE17538 DFS, P=0.015→0.08) and an inflated/unstable Cox HR (6.08 vs 3.016). Overall a solid partial reproduction with explainable, moderate deviations.

🤝
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.

154.7 k
tokens (I/O) · 6.4 M incl. cache
16 min
runtime · 0.03 CPU-h
1.6 GB
peak RAM
4 (1 failed)
HPC jobs
hummel
machine