Hepatocyte dedifferentiation in 2D culture reveals extensive transcriptomic and proteomic rewiring.
The main results reproduced: recomputed values matched the published ones within tolerance.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- Nothing in this column.
- 🟡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
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 reproduce the transcriptomic branch; essentially 1:1 there, partial overall. The repo is the AUTHORS' own code (R targets+renv; github.com/Mortendall/PrimHepMultiOmics @ f24acb6). The BULK RNA-seq differential-expression pipeline — the paper's core 'transcriptomic rewiring' result — reproduced cleanly by running the authors' verbatim edgeR pipeline (filterByExpr->calcNormFactors->estimateDisp->glmQLFit->glmQLFTest, DEG@FDR<0.05) on the DEPOSITED GEO count matrix GSE173406_count_matrix.xlsx. metadata.xlsx is not deposited but was fully reconstructed from the GEO series matrix (WT mice {544,548,554,556,558,564}), yielding the SAME 15-sample/5-per-group WT design after the repo's outlier removal (558L,544CS,544PH) [C3 exact]. DEG counts reproduced the headline Fig-1C upset claims: >10,000 DE for PH-vs-L (10,715) and PH-vs-CS (10,480) [C1a/C1b EXACT], ~1000 for L-vs-CS (1,288) [C2 within-tol], with the central overlap 9,746 vs reported ~8,700 [C1c within-tol; the paper's upset is computed on de-duplicated SYMBOLs which lowers the count]. GO Cellular-Component enrichment reproduced the headline biology: genes DOWN in cultured hepatocytes enriched for mitochondrial/OXPHOS terms (respiratory chain complex, NADH dehydrogenase, mito matrix) [C4], genes UP enriched for ribosome terms [C5 exact]. NOT reproduced 1:1: (i) proteomics per-group protein counts [C6 mismatch] and the limma DAP pipeline [C8] — the repo reads a combined file morten_proteome_combined_hybrid.txt that is NOT deposited; PRIDE PXD068742 ships only 3 separate per-batch DIA-NN matrices needing bespoke merging (hard 20%); (ii) the single-nucleus RNA-seq branch [C7] — the README states the Seurat .rds objects cannot be uploaded due to size and invites contacting the authors (data effectively on-request/restricted). No fabrication concern: all reproduced bulk RNA-seq values are independently derivable from public GEO data via the authors' shipped code. Provisional grades; human reviewer decides via AUDIT.md.
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.
-
v1 current initial assessment Score 77assessed: 2026-06-14 ⛓ 18b77b9e1423
✎ 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
- 2026-06-14
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-15no human curator yet
- Last updated
- 2026-09-19
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: sonnetThe study tests whether isolating primary mouse hepatocytes and culturing them in 2D for 24 hours causes dedifferentiation-associated rewiring of the transcriptome and proteome, and whether such changes reflect genuine cellular adaptation rather than shifts in cell-type composition.
- ★ 2D culturing of primary hepatocytes for 24 hours causes extensive transcriptomic (>10,000 genes) and proteomic (>3000 proteins) rewiring compared with freshly isolated cells finding
- ★ Culturing decreases transcriptional heterogeneity and causes loss of zonal markers finding
- ★ Culturing alters proteins associated with the extracellular matrix, mitochondria, and ribosomes, and increases acute-phase response proteins finding
- ★ Mitochondrial inner membrane genes show increased protein abundance but decreased gene expression in cultured hepatocytes compared with liver, revealing mRNA-protein discordance mechanism
- ★ The isolation process itself (Cell Suspension vs Liver) also alters gene expression and protein abundance, though to a lesser extent than culturing finding
- Proteins involved in peptidase and glutathione transferase activity remain largely stable in abundance during culturing finding
- ★ The authors developed the Shiny app 'Hepamorphosis' to allow exploration of RNA/protein correlations, zonation profiles, and cell-type-specific transcription resource
- ★ snRNAseq confirms that transcriptional changes upon isolation and culturing reflect cellular adaptation rather than solely a shift in cell-type composition mechanism
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| bulk RNA-seq (reanalysis of GSE173406) | female C57BL/6JBomTac mouse liver/hepatocytes (Liver, Cell Suspension, Primary Hepatocytes cultured 24h on collagen) | hepatocyte isolation and 24h 2D culturing | differential gene expression, GO enrichment | — |
| quantitative proteomics | male C57BL/6NTac mouse liver/hepatocytes (Liver, Cell Suspension, Primary Hepatocytes cultured 24h) | hepatocyte isolation and 24h 2D culturing | differential protein abundance, GO enrichment | — |
| single-nucleus RNA sequencing (snRNAseq), analyzed as pseudo-bulk and by clustering | same male C57BL/6NTac mouse liver/hepatocyte samples as proteomics cohort (Liver, Cell Suspension, Primary Hepatocytes) | hepatocyte isolation and 24h 2D culturing | pseudo-bulk differential gene expression, UMAP clustering/cell-type heterogeneity, GO enrichment | — |
- – Over 10,000 genes differentially expressed between Primary Hepatocytes and both Liver and Cell Suspension (bulk RNAseq), ~8700 overlapping >10,000 genes; ~8700 overlapping
- – ~1000 genes differed between Cell Suspension and Liver, indicating the isolation process alone alters gene expression ~1000 genes
- – ~3000 proteins differentially abundant between Primary Hepatocytes and both Liver and Cell Suspension, with 2335 overlapping ~3000 proteins; 2335 overlapping
- ▼ ~400 proteins differed between Liver and Cell Suspension, indicating protein loss during isolation ~400 proteins
- – Mitochondrial inner membrane genes showed reduced protein abundance but increased RNA levels in Liver versus Primary Hepatocytes (inverse relationship)
- – snRNAseq pseudo-bulk analysis confirmed >10,000 DEGs between Primary Hepatocytes and Liver and between Cell Suspension and Primary Hepatocytes, and >5000 genes differed between Liver and Cell Suspension >10,000 genes; >5000 genes
- – Genes downregulated in Primary Hepatocytes were enriched for mitochondrial-associated GO terms, while upregulated genes were enriched for ribosomal components
- ▲ 51 GO cellular component terms were significantly enriched in cultured Primary Hepatocytes versus Liver, largely ribosomal and mitochondrial 51 GO terms
- count >10,000 genes differentially expressed (Primary Hepatocytes vs Liver and vs Cell Suspension, bulk RNAseq)
- count ~8700 overlapping genes (overlap between PH-vs-Liver and PH-vs-CS DEG sets)
- count ~1000 genes (Cell Suspension vs Liver DEGs)
- count ~3000 differentially abundant proteins (Primary Hepatocytes vs Liver and vs Cell Suspension, proteomics)
- count 2335 overlapping proteins (overlap between PH-vs-Liver and PH-vs-CS differentially abundant proteins)
- count ~400 proteins (Liver vs Cell Suspension differentially abundant proteins)
- count 5755 (Liver), 5676 (Cell Suspension), 5680 (Primary Hepatocytes) proteins detected (total proteins identified per sample type)
- count >10,000 genes (PH vs Liver; CS vs PH); >5000 genes (Liver vs CS) (snRNAseq pseudo-bulk differential expression)
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.
The paper characterizes transcriptomic and proteomic changes in primary mouse hepatocytes across three states (whole liver, freshly isolated cell suspension, 24-hour 2D-cultured hepatocytes) using bulk RNAseq re-analysis, label-free quantitative proteomics, and pseudo-bulk single-nucleus RNAseq from overlapping cohorts. Differential expression and abundance were assessed for all pairwise group combinations using DESeq2 (RNA) and an unstated method (proteomics, described only in Supplemental Methods), with FDR correction applied. Results were reported primarily as counts of differentially expressed or abundant features, log2 fold changes, and Gene Ontology enrichment analyses rather than inferential statistics with effect sizes and uncertainty measures.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| DESeq2 Wald test (negative binomial GLM) | Pairwise differential gene expression among Liver, Cell Suspension, and Primary Hepatocytes — bulk RNAseq re-analysis of GSE173406 | not stated in main text (control animals only from public dataset) | not stated |
| DESeq2 Wald test on pseudo-bulk aggregated counts (negative binomial GLM, design ~group) | Pairwise differential gene expression from snRNAseq aggregated per sample: Liver vs Cell Suspension, Liver vs Primary Hepatocytes, Cell Suspension vs Primary Hepatocytes | not stated in main text; ≥6 mice implied by mouse-ID references | not stated |
| Differential protein abundance test (method not named in main text; see Supplemental Methods) | Pairwise protein abundance comparisons: Liver vs Cell Suspension, Liver vs Primary Hepatocytes, Cell Suspension vs Primary Hepatocytes (proteomics cohort) | not stated in main text | not stated |
| Overrepresentation analysis / hypergeometric test via ClusterProfiler | Gene Ontology enrichment of differentially expressed genes and differentially abundant proteins for each pairwise comparison; detected/expressed feature list used as background | na | not stated |
-
Pseudo-bulk differential expression aggregated snRNAseq counts per sample and analyzed them with DESeq2↳ Could also: Mixed-model approaches such as dream (variancePartition) could also analyze snRNAseq data while retaining cell-level observations and explicitly modeling donor as a random effect — Mixed-model methods can increase statistical power when per-sample cell counts are unequal and preserve cell-level variability information that pseudo-bulk aggregation collapses into a single value per donor
-
GO enrichment used a threshold-based overrepresentation approach (ClusterProfiler ORA) with detected genes as the background universe↳ Could also: Gene Set Enrichment Analysis (GSEA) on the complete ranked list of log2 fold changes could also be applied — GSEA avoids a hard significance cutoff for defining the hit list and can detect coordinated, directional shifts across a pathway even when no single gene clears a significance threshold
-
Three pairwise group comparisons were each corrected for multiple testing independently within their own FDR family↳ Could also: A single multi-group model with planned contrasts and one joint FDR adjustment across all contrasts could also be applied — A joint correction controls the overall false-discovery rate uniformly across the full set of tests; separate per-comparison FDR families treat each contrast as an independent experiment, which is one recognized analytic choice but differs in the error-rate guarantee
-
Transcriptomic (bulk RNAseq) and proteomic cohorts used mice that differed in sex, strain, and diet; the mRNA–protein discordance was subsequently re-examined in a matched cohort (snRNAseq + proteomics from the same animals)↳ Could also: Collecting RNA and protein from the same animals in a single matched cohort could also have been used as the primary design — A matched design eliminates cohort as a confound when directly comparing mRNA and protein fold changes — the authors themselves identify cross-cohort differences as a potential bias motivating the second snRNAseq experiment
-
Sample-level structure was visualized with multidimensional scaling (MDS) plots↳ Could also: Principal component analysis (PCA) or UMAP could also visualize between-sample variation — PCA provides explicit loadings identifying which genes drive sample separation; UMAP can reveal non-linear structure; both are widely used in multi-omics contexts and facilitate cross-study comparison
-
The specific statistical test used for proteomics differential abundance is described only in Supplemental Methods and is not named in the main text↳ Could also: Commonly used LFQ proteomics approaches include limma with empirical Bayes moderation or t-tests with permutation-based FDR (e.g., Perseus); either would be a standard, well-documented choice — Stating the test in the main methods section allows readers to evaluate distributional assumptions and reproducibility without needing to consult the supplement
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.
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.
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-41111196
Paper: Dall M, Stocks B, Cervone DT, Deshmukh AS, Treebak JT. Hepatocyte dedifferentiation in 2D culture reveals extensive transcriptomic and proteomic rewiring. Hepatol Commun 2025. PMID 41111196 / PMC12506984 / DOI 10.1097/hc9.0000000000000795.
Repo: https://github.com/Mortendall/PrimHepMultiOmics (authors' own code; R
targets + renv workflow). Pipeline funcs in R/functions.R (bulk RNA),
R/functionsProteomics.R (proteomics), R/function_singleCell.R (snRNA).
Three groups throughout: L = whole-liver biopsy (after saline perfusion), CS = washed hepatocyte cell suspension (pre-seeding), PH = primary hepatocytes after 24 h 2D culture.
In scope (pipeline-derived, attempted)
A. Bulk RNA-seq differential expression — PRIMARY, fully reproducible
- Data: GEO GSE173406, supplementary
GSE173406_count_matrix.xlsx(raw gene counts, Ensembl IDs) — the exact file the repo'scount_matrix_assembly()reads. Public, directly downloadable. - Pipeline (
functions.R): filter toGenotype==WT, drop 3 outliers (558L, 544CS, 544PH) → edgeRDGEList→filterByExpr(design)→calcNormFactors→estimateDisp→glmQLFit→glmQLFTestfor 3 contrasts (L−CS, L−PH, CS−PH) →topTags. DEG = FDR < 0.05 (UpsetplotGeneration). - Metadata reconstruction:
metadata.xlsxis NOT deposited, but the sample→group→genotype map is fully recoverable from the GEO series matrix (!Sample_description= sample IDs like544L;!Sample_characteristicsgenotype = WT/HNKO). WT mice = {544,548,554,556,558,564}; 6/group → minus 3 outliers = 15 samples, 5/group. Group from the ID suffix (L/CS/PH). - Claims to compare (Fig 1C upset, Results):
- C1: "Over 10,000 genes DE between PH and both L and CS, ~8700 overlapping"
- C2: "Around 1000 genes differed between CS and L (L−CS)"
-
- sample design (15 samples, 5/group) and gene count after filterByExpr.
B. Bulk RNA-seq GO Cellular Component enrichment — SECONDARY
GOCCSplit(): clusterProfilercompareCluster(enrichGO, ont="CC", OrgDb=org.Mm.eg.db)on up/down DE genes (FDR<0.05) per contrast, universe = all tested genes. Headline claim: down-in-PH genes enriched for mitochondrial inner membrane (GO:0005743); up enriched for ribosome (GO:0005840).
C. Proteomics protein-count sanity — BONUS (cheap)
- Data: PRIDE PXD068742 ships 3 DIA-NN
*.pg_matrix.tsv(processed protein groups) + raw.d.rar. Claim: "5755 proteins in Liver, 5676 in CS, 5680 in PH". Attempt a per-group detected-protein count from the pg_matrix files.
Out of scope (hard 20% / not feasible — not attempted)
- Single-nucleus RNA-seq (GSE280301) — the README states the Seurat
.rdsobjects "cannot be uploaded" due to size and invites readers to contact the authors. Thetargets/Figures code reads a local220503_liver_full-seurat_updated.rdsthat is not deposited. → effectively data_restricted / on-request for the snRNA branch (Figs 3–6, cluster proportions, pseudobulk DE). Recorded, not attempted. - Full proteomics DAP pipeline — the repo's
ProteomicsDataLoaderreads a single combined filemorten_proteome_combined_hybrid.txt(cols Liver1..8, CS1..8, PH1..8) that is not deposited; PRIDE ships 3 separate DIA-NN matrices (hybrid/ph/cs) that would need bespoke merging into the authors' exact format to rerun limma. This reconstruction is the hard 20%; the limma DAP counts (Fig 2C: ~3000 DAP, 2335 overlap, ~400 L−CS) are not reproduced 1:1. Per-group protein counts (C) attempted as a partial check only. - Wet-lab (mouse isolations, MS acquisition), the Hepamorphosis Shiny app, and all figure cosmetics — out of scope by definition.
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.
The paper's core transcriptomic-rewiring result reproduced essentially 1:1 by running the authors' own edgeR pipeline on the deposited GEO count matrix: >10,000 DE genes for PH-vs-L (10,715) and PH-vs-CS (10,480), ~1,000 for L-vs-CS (1,288), with mitochondrial loss and ribosomal gain confirmed (C1–C5). The main deviations are on the data-availability side: proteomics per-group counts (C6, 4877/5249/4025 vs reported 5755/5676/5680) and the snRNA branch could not be reproduced because the merged proteome matrix and Seurat .rds objects are not deposited, and the C1c overlap gap is a SYMBOL-vs-ENSEMBL preprocessing artifact. Severity is moderate and explainable with no fabrication concern — every reproduced value is independently derivable from public GEO data — so this is a solid partial reproduction, fully confirming the transcriptomic conclusion but leaving the proteomic/snRNA claims untestable from shared artifacts.
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 [email protected].
Reproduction footprint
claude-opus-4-8Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.