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

Dissecting lncRNA-mRNA competitive regulatory network in human islet tissue exosomes of a type 1 diabetes model reveals exosome miRNA markers.

Front Endocrinol (Lausanne) · 2022
L1 50/100 PQI 88
Why this verdict

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

Reproduced on the brainbox compute brainarbeit.com
✓ 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
50/100
Reproducibility score
1.4 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 8% of all assessed papers rank 1026 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

PARTIAL. This is a P16 third-party-reuse case: the linked 'code' (github.com/cytoscape/cytoscape) is the Cytoscape VISUALIZATION app, not an analysis pipeline; the paper re-analyzes public RNA-seq GSE139932 (10 human islet-exosome samples = 5 donors x {control, IL-1b+IFN-g cytokine}, GPL20301). Data was described well enough to start from the GEO-shipped processed matrices (raw + DESeq2 vsd counts, 17013 mRNA + 5711 lncRNA), faithful to the stated STAR+Gencode29+DESeq2 quant. REPRODUCED 1:1 (clean data point): the DESeq2 differential-expression step (paired ~donor+condition, Cyto vs Ctrl) -> under standard thresholds there is essentially NO differential expression in the exosomal fraction (0 genes at padj<0.05 for both mRNA and lncRNA; min padj 0.9998; nominal p<0.05 counts 143/32 are BELOW the chance expectation), which is consistent with the paper reporting only ~0.4% proportions and never a DE count -- no fabrication flag, but it shows the downstream ceRNA/miRNA story rests on n=5 co-expression (Spearman r>=0.9, trivially attainable with 5 samples) rather than on a robust DE gene set. The headline ceRNA NETWORK counts (ctrl 2864/249/1560; cyto 2529/244/1464) and hub degrees (NEAT1/KCNQ1OT1/XIST) are pipeline-derived and gradeable; the full reconstruction (ENCORI/starBase shared-miRNA>3 AND r>=0.9, per group) was coded, debugged (ENCORI mRNA bulk truncates -> solved by fetching mRNA targets per-miRNA only for the 636 miRNAs that hit expressed lncRNAs), validated, and LAUNCHED on «our HPC» («job», lncRNA layer = 16853 ENCORI pairs / 641 miRNAs confirmed), but the node/edge counts had NOT landed at the operator's finalize cutoff (job was ~55% through the per-miRNA fetch). No network number is therefore asserted -- honest in-progress partial, not a match claim; outputs will land on «infra» at reproductions/pmid-36440209/outputs/net_results.json for human folding. NOT ATTEMPTED (out of scope / hard ~20%): the 19-miRNA marker panel + coefficients and the validation AUCs (0.7569 GSE97123, 0.8125 GSE189107) -- a separate miRNA dataset + a fitted classifier not shipped; GO/KEGG term lists; Cytoscape figure layouts; re-alignment from FASTQ (GEO ships the matrices). KEY CAVEATS for the auditor: starBase v2.0 is retired so ENCORI v3 is a stand-in (network counts version-sensitive); paper states no DE thresholds; and the expression pre-filter feeding the network is unspecified, so an EXACT match to 2864/2529 is not expected even on completion -- order-of-magnitude + hub identity is the realistic 1:1 check. All grades provisional; a human reviewer decides.

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 50
    assessed: 2026-06-15 ⛓ ddbc84eebbb0
✎ 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

Can integrating lncRNA and mRNA expression data from human pancreatic islet-derived exosomes (with/without pro-inflammatory cytokine stimulation mimicking T1DM) to construct lncRNA-miRNA-mRNA ceRNA networks reveal regulatory mechanisms and identify exosomal miRNA/lncRNA biomarkers for type 1 diabetes?

Core claims
  • lncRNA-mRNA ceRNA networks differ substantially between control and cytokine-treated islet exosomes, sharing lncRNAs more than mRNAs, implying state-dependent regulatory functions finding
  • Differentially expressed lncRNAs in the ceRNA network are enriched in insulin secretion-related pathways including Hippo, TGF-beta, Wnt, FOXO, Neurotrophin and ErbB signaling finding
  • A 19-miRNA regression model derived from competitively regulated, differentially expressed miRNAs serves as a diagnostic marker panel for diabetes, validated in independent plasma and saliva exosome datasets resource
  • lncRNA PVT1, LINC00960 and hsa-miR-107 may be involved in the inflammatory response in T1DM and serve as candidate biomarkers/therapeutic targets mechanism
  • lncRNA-mRNA competitive regulatory pairs were defined by sharing >3 common miRNAs and a Spearman expression correlation ≥0.9 method
  • PI3K-Akt and FOXO signaling pathways are shared between control and cytokine network states and may be critical to diabetes occurrence finding
Experimental setups
Assay System Perturbation Readout Platform
RNA sequencing / transcriptome (lncRNA and mRNA expression) of islet exosomes Human pancreatic islet-derived exosomes (GSE139932, 10 samples) 50 U/ml IL-1β + 1000 U/ml IFN-γ for 24h (cytokines group) vs no stimulation (control) Differential lncRNA/mRNA expression; ceRNA network construction STAR aligner v2.5.3a, Gencode v29, DESeq2
Exosome isolation Human pancreatic islet tissue none/cytokine Isolated exosomes for RNA sequencing Ultracentrifugation combined with commercial kits
lncRNA-miRNA and miRNA-mRNA interaction annotation (CLIP-Seq based) Database-derived global interactions none Shared miRNA-mediated lncRNA-mRNA candidate associations StarBase V2.0
Plasma-derived exosome expression profiling (independent validation) Human plasma exosomes (GSE97123) none miRNA expression for ROC/AUC classification
Saliva exosome expression profiling (independent validation) Human saliva exosomes (GSE189107) none miRNA expression for ROC/AUC classification
KEGG functional enrichment analysis Competitively regulated mRNAs from ceRNA networks none Enriched signaling pathways
Logistic/regression modeling and ROC analysis 19 candidate miRNAs across GSE139932 training and validation datasets none Risk coefficients, AUC for control vs cytokine classification R / Cytoscape V3.8.0
Key results
  • Control group ceRNA network contained 2864 lncRNA-mRNA interactions, 249 lncRNAs and 1560 mRNAs 2864 interactions; 249 lncRNAs; 1560 mRNAs
  • Cytokines group ceRNA network contained 2529 lncRNA-mRNA interactions, 244 lncRNAs and 1464 mRNAs 2529 interactions; 244 lncRNAs; 1464 mRNAs
  • Networks shared lncRNAs more than mRNAs; shared lncRNA-mRNA pairs were very low lncRNA sharing 66.5%; mRNA sharing 36.0%; shared lncRNA-mRNA pairs 4.4% (230 pairs)
  • 19-miRNA model validated on plasma exosome dataset (GSE97123) AUC=0.7569 (10 miRNAs)
  • 19-miRNA model validated on saliva exosome dataset (GSE189107) AUC=0.8125 (13 miRNAs)
  • NEAT1, KCNQ1OT1, XIST, MALAT1 and PVT1 were high-degree hub lncRNAs in both networks NEAT1 degree 191 (control), KCNQ1OT1 degree 361 (cytokines)
  • Hsa-miR-17-5p had the largest positive risk coefficient; hsa-miR-802 most negative; hsa-miR-107 negative coefficient miR-17-5p=1947.24; miR-802=-1044.62; miR-107=-284.08
  • Differentially expressed mRNA proportion was higher in cytokines group while differential lncRNA proportion was nearly equal (~0.4%) lncRNA ~0.4% both states
Key statistics
  • correlation ≥0.9 Spearman cutoff for competitive regulatory pair (lncRNA-mRNA correlation threshold for ceRNA pair definition)
  • other AUC=0.7569 (ROC for miRNA model on plasma exosome dataset GSE97123)
  • other AUC=0.8125 (ROC for miRNA model on saliva exosome dataset GSE189107)
  • count 19 candidate miRNAs (miRNA markers screened for regression model)
  • count 230 shared lncRNA-mRNA pairs (interactions shared between two state networks)
  • count degree 361 (KCNQ1OT1 top degree lncRNA in cytokines group network)
  • other -284.08 (regression coefficient of hsa-miR-107 in 19-miRNA model)
  • count 66.5% lncRNA vs 36.0% mRNA sharing (degree of node sharing between control and cytokine networks)

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.

This bioinformatics study constructed lncRNA-miRNA-mRNA ceRNA networks from RNA-sequencing data of human pancreatic islet-derived exosomes (GSE139932, n=10) comparing control and cytokine-stimulated (IL-1β + IFN-γ) conditions. Differential expression was identified with DESeq2, and Spearman correlation (r ≥ 0.9) defined competitive regulatory lncRNA-mRNA pairs; KEGG pathway enrichment characterized functional roles. An unspecified regression model was built from 19 candidate miRNAs and evaluated in two independent GEO datasets (GSE97123, GSE189107) via ROC analysis, yielding AUC values of 0.7569 and 0.8125.

Replicationunclear Sample size10 samples total stated; per-group split not explicitly reported; no power calculation mentioned Groupscontrol exosomes vs cytokine-stimulated (IL-1β + IFN-γ) exosomes from human pancreatic islets; independent validation in plasma (GSE97123) and saliva (GSE189107) exosome GEO datasets Pairingunclear Randomization/blindingnot stated Dispersionnone Exact p-valuesno Effect sizesyes Confidence intervalsno
Statistical tests used
Test Applied to n Assumptions
DESeq2 (Wald test, implied by package default) Differential expression of lncRNAs and mRNAs between control and cytokine-stimulated exosome groups (GSE139932) 10 not stated
Spearman correlation (fixed cutoff r ≥ 0.9) ceRNA network construction — identifying competitive lncRNA-mRNA regulatory pairs within each group 10 not stated
Regression analysis (type unspecified) Derivation of risk coefficients for 19 candidate miRNAs using training dataset GSE139932 10 not stated
ROC / AUC analysis Validation of miRNA biomarker model in plasma exosome dataset GSE97123 and saliva exosome dataset GSE189107 not stated
KEGG enrichment analysis (statistical method not specified) Functional annotation of mRNAs in ceRNA networks and shared lncRNA-mRNA interactions not stated
Approaches that could also have been used
  • Regression coefficients for 19 miRNA features were derived from a training dataset of n=10 samples without reported cross-validation or regularization
    Could also: LASSO or ridge regression with leave-one-out or k-fold cross-validation could also be used to build a penalized model and estimate out-of-sample performance in the same dataset — With 19 predictors and n=10 observations, a penalized or cross-validated approach also quantifies model stability and expected generalization performance before applying coefficients to external datasets
  • AUC values for the two independent validation datasets were reported as single point estimates (0.7569 and 0.8125) with no measure of uncertainty
    Could also: Bootstrap-derived 95% confidence intervals for AUC, or DeLong's method, could also be reported alongside point estimates — CIs around AUC convey the precision of the discrimination estimate and allow readers to assess whether the interval excludes 0.5 (chance performance), which is particularly informative when validation n is small or unstated
  • Competitive lncRNA-mRNA pairs were defined by a fixed Spearman correlation cutoff of r ≥ 0.9, applied to n=10 samples
    Could also: A permutation-based significance test or hypergeometric test for shared miRNA counts could also provide a statistically grounded threshold for defining competitive regulatory pairs — A significance-based threshold also accounts for the number of possible pairs and chance co-occurrence, and communicates the expected false-positive rate among selected pairs alongside the magnitude cutoff
  • Pathway enrichment was performed using KEGG alone for all network components
    Could also: Gene Ontology (GO) enrichment or Reactome pathway analysis could also be run alongside KEGG — Different pathway databases annotate complementary aspects of biology; using multiple databases also captures processes not well-covered by KEGG, such as detailed molecular-function or cellular-component terms from GO
  • DESeq2 was the sole differential expression tool applied to the n=10 RNA-seq samples
    Could also: edgeR (quasi-likelihood F-test or exact test) could also have been applied to the same count data as a parallel analysis — edgeR is another standard Bioconductor method for small-n RNA-seq using empirical Bayes dispersion estimation; comparing results from both tools is a common robustness check that highlights findings consistent across methods
  • The 19 candidate miRNAs entering the regression model were selected by combining membership in the competitive network with a differential miRNA list from the source paper, without a stated FDR threshold
    Could also: An explicit FDR threshold (e.g., BH-adjusted p < 0.05 or 0.10) for both differential expression and network-membership criteria could also be stated to make the selection criterion fully reproducible — Stating an explicit FDR cutoff also allows readers to understand the expected false-positive rate among the 19 candidates and to reproduce the feature-selection step independently
Software: STAR aligner 2.5.3a · Gencode annotation v29 · DESeq2 (R/Bioconductor) · Cytoscape 3.8.0 · R (AUC/ROC calculation, custom code) · StarBase V2.0

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.

Authors · 6
1Fang Tian 2Xue Gong 3Wu Jianjun 4Long Wei 5Zhang Xiaomeng 6Yang Fan
Citations
16
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.

GSE139932 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSE189107 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSE55098 GEO in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
GSE55099 GEO in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
GSE94649 GEO in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
GSE97123 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-36440209

Paper: Fang T. et al. (2022) Dissecting lncRNA-mRNA competitive regulatory network in human islet tissue exosomes of a type 1 diabetes model reveals exosome miRNA markers. Front Endocrinol 13:1015800. DOI 10.3389/fendo.2022.1015800.

Nature of this RU (P16 third-party reuse)

The "code" link (github.com/cytoscape/cytoscape) is a generic network-visualization desktop app, not an analysis pipeline — a text-mining link. The paper itself is a re-analysis of public RNA-seq data: it reuses GEO GSE139932 (Evans-Molina lab, Indiana Univ.; 10 human pancreatic-islet exosome RNA-seq samples = 5 donors × {control, IL-1β+IFN-γ cytokine}, Illumina HiSeq 4000, GPL20301). Reproduction = re-running the described pipeline on the paper's own data. Per BRIEF rule P16 this is equally valid.

Pipeline as described in Methods

  1. Alignment/quant: STAR v2.5.3a, Gencode v29. → GEO already ships the result as processed matrices: GSE139932_Rawcounts_mRNAs_and_lncRNAs.xlsx and ..._vsdcounts_...xlsx (vsd = DESeq2 variance-stabilized). We start from these (faithful: same counts the authors used; we skip re-aligning raw FASTQ = the 20%).
  2. Differential expression: DESeq2 (named). Thresholds not stated in Methods → reproducibility gap; we report counts at a standard threshold and document.
  3. ceRNA network: for lncRNA–mRNA pairs, shared miRNAs from starBase v2.0 (CLIP-seq supported miRNA–lncRNA + miRNA–mRNA interactions); keep pairs with >3 shared miRNAs AND Spearman |r| ≥ 0.9 (computed within each group, n=5). Networks built per group; degrees computed; visualized in Cytoscape v3.8.0.

IN SCOPE (pipeline-derived, attempt)

  • C-NET-CTRL: Control ceRNA network size — 2864 lncRNA–mRNA interactions, 249 lncRNAs, 1560 mRNAs (Results/Fig).
  • C-NET-CYTO: Cytokine ceRNA network size — 2529 interactions, 244 lncRNAs, 1464 mRNAs.
  • C-OVERLAP: Shared between groups — 230 interactions, 801 shared mRNAs (36.0%), ~66.5% shared lncRNAs.
  • C-HUB-CTRL: Top hub lncRNAs (degree) control — NEAT1 191, KCNQ1OT1 179, XIST 150 (Table 1).
  • C-HUB-CYTO: Top hub lncRNAs cytokine — KCNQ1OT1 361, NEAT1 261 (Table 2).
  • C-DE (support): DESeq2 DE on the matrices (no exact paper count to grade; sanity).

OUT OF SCOPE (not attempted; the hard ~20% / non-pipeline)

  • miRNA marker panel (Table 3: 19 candidate miRNAs w/ coefficients; hsa-miR-17-5p 1947.24 etc.) and validation AUCs (GSE97123 plasma 0.7569; GSE189107 saliva 0.8125): the miRNA layer comes from a different miRNA dataset + a fitted regression/classifier model on other GEO sets — separate data, model not shipped.
  • GO/KEGG enrichment specific term lists (qualitative, many tools/versions).
  • Cytoscape visual figures (layout is cosmetic; node/edge counts cover the substance).
  • Re-alignment from FASTQ (STAR) — GEO ships the count matrices; re-aligning adds no comparison value and is the deliberately-skipped 20%.

Key reproducibility risks (to report honestly)

  • DE thresholds unstated → DE counts not gradeable against paper.
  • starBase v2.0 is retired; current = ENCORI/starBase v3. miRNA-interaction tables differ by version → exact network counts (2864 etc.) are version-sensitive.
  • n=5 per group makes Spearman |r|≥0.9 easy to hit → network size very sensitive to the expression pre-filter (which genes enter), which the paper does not specify.
  • Therefore: exact match on 2864/2529 is unlikely; we target order-of-magnitude agreement + hub-identity (NEAT1/KCNQ1OT1/XIST as top hubs) as the honest 1:1 check.

Data/code pointers

  • Data: GEO GSE139932 (SRA SRP228580), processed xlsx on GEO FTP.
  • starBase/ENCORI: https://rnasysu.com/encori/ (miRNA-lncRNA, miRNA-mRNA, human/hg).
  • Cytoscape (viz only): github.com/cytoscape/cytoscape.
  • «infra» work dir: «path»
Figures / tables: TableFig 4
C-DE-support
Reported
no DE count (only ~0.4% proportions, Fig 4); pipeline STAR+Gencode29+DESeq2
Reproduced
DESeq2 paired (~donor+condition) Cyto vs Ctrl on GSE139932 matrices: 0 DE at padj<0.05 (mRNA 17013 & lncRNA 5711); min padj 0.9998; nominal p<0.05 = 143 mRNA / 32 lncRNA (below chance)
partial
C-NET-CTRL
Reported
2864 edges / 249 lncRNA / 1560 mRNA
Reproduced
reconstruction built+validated+launched («our HPC» «job»); counts not landed at finalize cutoff
partial
C-NET-CYTO
Reported
2529 edges / 244 lncRNA / 1464 mRNA
Reproduced
same «job», running at cutoff
partial
C-OVERLAP
Reported
230 shared edges; 801 shared mRNA
Reproduced
computed by same job; not landed at cutoff
partial
C-HUB
Reported
NEAT1 191/KCNQ1OT1 179/XIST 150 (ctrl); KCNQ1OT1 361/NEAT1 261 (cyto)
Reproduced
degree tables emitted by job; not landed at cutoff
partial
C-MIR-PANEL
Reported
19 miRNA markers + coefficients; AUC 0.7569/0.8125
Reproduced
OUT OF SCOPE — miRNA layer from different GEO datasets + unshipped fitted model (hard ~20%)
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 50/100

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.

198.8 k
tokens (I/O) · 14 M incl. cache
36 min
runtime · 0.92 CPU-h
1.9 GB
peak RAM
4
HPC jobs
hummel
machine