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

Vascular adhesion protein-1 defines a unique subpopulation of human hematopoietic stem cells and regulates their proliferation.

Cell Mol Life Sci · 2021
L1 50/100 3/4
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: 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
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

Described well enough to be reproducible in principle: the paper's headline computational results come from scRNA-seq PRJNA729883 (SRA study SRP319821; 4 sorted-HSC 10x v3.1 samples, raw FASTQ only) via Cell Ranger 5.0.1 -> Seurat 4.0.1 -> UMAP/Monocle2 -> Nebulosa. BRIEF's PRJNA594799 is actually the BULK set; no GEO processed matrix or Seurat object is shipped and the paper's Code-availability section is blank, so reproduction must run the third-party pipeline on the FASTQ from scratch (valid per P16). OUTCOME = PARTIAL: I resolved the data discrepancy, pinned 4 claims (C1 AOC3 near-zero, C2 1267 HSC, C3 HSC1/HSC2 93/7%=89, C4 HSC2 cell-cycle enrichment), and built+submitted the full pipeline on «our HPC» (10x GRCh38-2020-A reference downloaded, STAR index built). NO numeric claim was graded: the ENA FASTQ staging step had a nested-xargs/wget quoting bug that wrote 32 zero-byte files, and the run was finalized on operator instruction before any count matrix existed; the alignment array was cancelled rather than run on empty input. I did NOT confirm or contradict any paper number, and explicitly did not attempt the bulk-RNA-seq DE markers or the Monocle2 pseudotime (hard/under-specified 20%). The STAR index, both sbatch scripts (STARsolo CR-emulation) and the Seurat+Nebulosa R script are staged on «infra» and re-runnable after a one-line download fix; deviation disclosed: Cell Ranger 5.0.1 -> STARsolo CR-emulation, same reference.

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 ⛓ 257826c4052c
✎ 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

The study tests whether Vascular Adhesion Protein-1 (VAP-1) marks a distinct subpopulation of human hematopoietic stem cells (HSC) and whether VAP-1-generated hydrogen peroxide regulates HSC proliferation and differentiation, both in vitro and in vivo.

Core claims
  • VAP-1 is expressed on a subset of human HSC and on bone marrow vasculature, forming a hematogenic niche. finding
  • VAP-1+ HSC are a transcriptionally unique small subset of differentiated, proliferating HSC, whereas VAP-1- HSC are the most primitive HSC. finding
  • VAP-1-generated hydrogen peroxide acts via the p53 signaling pathway to regulate HSC proliferation. mechanism
  • VAP-1 functions as a check point-like inhibitor of HSC differentiation; its inhibition enhances HSC expansion and differentiation into colony-forming units. finding
  • VAP-1 in bone marrow vasculature supports HSC expansion, confirmed using VAP-1 knockout mice, enzymatically inactive VAP-1 knock-in mice, and an enzyme inhibitor. finding
  • VAP-1 expression enables characterization and prospective isolation of a new subset of human HSC. resource
  • VAP-1 inhibitor (LJP-1586) can be used to expand HSC for potential clinical use. method
Experimental setups
Assay System Perturbation Readout Platform
Bulk RNA-seq FACS-sorted VAP-1+ and VAP-1- HSC from human bone marrow (n=4 donors) none (VAP-1+ vs VAP-1- sorting) genome-wide gene expression / differentially expressed genes SMART-Seq v4 Ultra Low Input RNA Kit (Takara), Nextera XT (Illumina), HiSeq 3000
Single-cell RNA-seq Sorted VAP-1+ and VAP-1- CD34+ Lin- human bone marrow cells (1 donor) none (VAP-1+ vs VAP-1- sorting) single-cell transcriptomes, clustering, trajectory 10x Genomics Chromium Single Cell 3' v3.1; Illumina NovaSeq 6000; Cell Ranger 5.0.1
Flow cytometry / FACS sorting Human BM and cord blood cells; mouse BM cells none VAP-1 and HSC marker expression; cell sorting FACSAria IIu, LSR Fortessa, Sony SH800
Quantitative RT-PCR Sorted human BM VAP-1+ and VAP-1- HSC (Lin-CD34+CD38-CD45RA-CD90+) none expression of TPX2, CDCA8, PCNA, MLLT3, TYMS normalized to B2M TaqMan Fast Advanced Master Mix; 7900HT Fast Real-Time PCR System
Immunocytochemistry / immunofluorescence FACS-sorted CD34+ cord blood cells; murine tissue sections; mouse femur whole-mount none VAP-1, CD31, CD150, Lineage localization Olympus BX60, Zeiss LSM780 confocal, 3i Marianas spinning disk confocal
Colony-forming unit (CFU) assay Human CB/BM CD34+ cells; mouse BM and peripheral blood cells VAP-1 inhibitor LJP-1586 treatment; VAP-1-KO vs WT mice number/type of colonies (CFUs) MethoCult H4435/H4436/M3434 (STEMCELL Technologies)
Long-term culture-initiating cell (LTC-IC) assay BM cells from WT and VAP-1-KO mice on irradiated stromal feeder layers VAP-1 knockout LTC-IC frequency (positive/negative scoring) MyeloCult M5300 / MethoCult GF M3434
ROS production measurement Human CD34+ BM cells in liquid culture (9 days) VAP-1 inhibitor LJP-1586 reactive oxygen species production
Key results
  • VAP-1+ HSC represent a transcriptionally distinct, small subset of differentiated and proliferating HSC, while VAP-1- HSC are the most primitive HSC.
  • Bulk RNAseq identified 687 VAP-1+ enriched and 378 VAP-1- enriched genes (fold change >1, p<0.05). 687 vs 378 genes
  • scRNAseq identified 371 VAP-1+ and 50 VAP-1- marker genes. 371 vs 50 genes
  • HSC expansion and differentiation into colony-forming units are enhanced by inhibition of VAP-1.
  • VAP-1-generated hydrogen peroxide regulates HSC proliferation via the p53 signaling pathway.
  • Contribution of VAP-1 to HSC proliferation confirmed with VAP-1-deficient mice, mutated (enzymatically inactive) VAP-1 mice, and enzyme inhibitor treatment.
Key statistics
  • count 687 VAP-1+ and 378 VAP-1- genes (fold change >1, p-value <0.05) (DEGs from bulk RNAseq used for Metascape analysis)
  • count 371 VAP-1+ and 50 VAP-1- genes (marker genes from scRNAseq used for analysis)
  • count 14,943 expressed genes (down from 60,619 total) (genes remaining after low-expression filtering in bulk RNAseq)
  • count 117–150 VAP-1+ and VAP-1- HSC per individual (cells used for bulk RNA sequencing)
  • count 11 donors (human bone marrow donors providing fresh BM cells)
  • count 260–320 M reads/lane (sequencing depth on HiSeq3000 run)
  • other ~10,000 cells per sample targeted; 400 cells/μl loaded (10x Genomics scRNAseq loading)

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 characterizes VAP-1+ versus VAP-1− human HSC subpopulations using a multi-modal design: bulk RNA-seq (n=4 BM donors, TMM-normalized, GSEA pathway analysis), single-cell RNA-seq (one donor, two technical replicates per group, Seurat/Wilcoxon-based marker identification), qPCR validation, and functional assays (CFU, LTC-IC, liquid culture) in both human cells and VAP-1 KO/KI/WT mouse models. Differential gene expression in bulk RNA-seq is reported at fold change >1 and p<0.05 thresholds; scRNA-seq cluster markers are identified via Wilcoxon rank-sum test through Seurat's FindMarkers. The paper text is truncated before the functional assay statistics sections, so those tests cannot be fully characterized.

Replicationmixed Sample size11 human BM donors total for flow cytometry/phenotyping; n=4 BM donors (Lonza) for bulk RNA-seq; one BM donor with two technical replicates per group for scRNA-seq; mouse experiments used sex- and age-matched littermates from heterozygous VAP-1-KO colonies; sample sizes for functional assays not extractable from the provided text GroupsVAP-1+ HSC vs VAP-1− HSC (human); VAP-1 KO vs KI vs WT mice; LJP-1586-treated vs untreated cells Pairingunpaired Randomization/blindingnot stated Dispersionunclear Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Wilcoxon rank-sum test (Seurat FindMarkers, default parameters) scRNA-seq differential expression for cluster marker gene identification Cells from one biological donor; two technical replicates per group (VAP-1+ and VAP-1−) not stated
GSEA pre-ranked Pathway and gene set enrichment analysis on bulk and scRNA-seq gene lists 687 VAP-1+ and 378 VAP-1− genes from bulk RNAseq; 371 VAP-1+ and 50 VAP-1− genes from scRNAseq not stated
Bulk RNA-seq differential expression (specific test not named; TMM normalization applied; likely edgeR or limma-voom per cited Kumar et al. pipeline) VAP-1+ vs VAP-1− HSC bulk transcriptome comparison n=4 human BM donors (Lonza); 117–150 cells per individual per group not stated
PCA and UMAP dimensional reduction (non-statistical inference tests) scRNA-seq dimensionality reduction and visualization Cells from one donor; four samples (two technical replicates × two populations) na
Monocle 2 trajectory inference (pseudotime ordering) Developmental trajectory construction from scRNA-seq data Cells from one donor not stated
Metascape GO and pathway over-representation analysis Gene Ontology biological process and pathway enrichment on DE gene lists Gene lists derived from bulk and scRNA-seq analyses not stated
Approaches that could also have been used
  • Bulk RNA-seq used TMM normalization (edgeR framework per the cited Kumar et al. pipeline) for differential expression between VAP-1+ and VAP-1− HSC
    Could also: DESeq2 with median-of-ratios normalization and negative binomial Wald test — DESeq2 is another widely adopted framework that models count overdispersion explicitly; comparing results across both frameworks is a common robustness check in low-n transcriptomic studies (n=4 here)
  • scRNA-seq differential expression used the Wilcoxon rank-sum test via Seurat's FindMarkers with default parameters
    Could also: MAST (Model-based Analysis of Single-cell Transcriptomics) or a mixed-effects model incorporating donor as a random effect — MAST accounts for the bimodal dropout structure of scRNA-seq data; a mixed-effects formulation would additionally accommodate the fact that all cells originate from a single donor, making the effective independent unit the cell rather than the individual
  • The scRNA-seq experiment used one biological donor with two technical replicates per group (VAP-1+ and VAP-1−)
    Could also: Multiple biological donors with pseudo-bulk aggregation per donor before differential testing — Pseudo-bulk approaches (e.g., summing counts per donor then applying bulk-RNA-seq methods) treat the donor as the unit of replication, which better separates inter-individual from within-individual cell-to-cell variance and is increasingly recommended for single-cell differential expression
  • Trajectory analysis used Monocle 2 to order cells along a pseudotime axis
    Could also: Monocle 3 or PAGA (partition-based graph abstraction, implemented in Scanpy) — Monocle 3 and PAGA use different graph-based algorithms that can model branching topologies more flexibly; comparing trajectories across methods is a standard robustness check given sensitivity of pseudotime to algorithm choice
  • Bulk RNA-seq DE gene lists were generated at a p<0.05 and FC>1 threshold with no explicitly described multiple-testing correction
    Could also: Benjamini-Hochberg false discovery rate (FDR) correction at a defined q-value threshold (e.g., q<0.05 or q<0.10) — With ~15,000 expressed genes tested, the expected number of false positives under a nominal p<0.05 threshold alone is substantial; FDR control explicitly quantifies and limits the proportion of false discoveries in the reported gene list
  • GSEA pre-ranked was used for pathway enrichment on the full ranked gene list from bulk and scRNA-seq comparisons
    Could also: Over-representation analysis (ORA) using a hypergeometric or Fisher's exact test on a discretized DE gene set — ORA is computationally simpler and more interpretable when a well-defined threshold gene list is already available; GSEA pre-ranked is more sensitive to coordinated moderate shifts across a pathway but requires a meaningful ranking metric, making the two approaches complementary
Software: R 3.6 and 4.0.5 · Seurat 3.1 and 4.0.1 · Cell Ranger 5.0.1 · Monocle 2 2.10.1 · Nebulosa (R package) · Metascape · GSEA (Broad Institute) · STITCH database · STAR (RNA aligner) · Trimmomatic · Rsubread · FastQC · RNA-SeQC · FlowJo (Tree Star) · ImageJ · SlideBook 6 (Intelligent Imaging Innovations) · Zen 2010 (Carl Zeiss)

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
1
Impact: low
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.

CG000315 ENA in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSM3305359 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
PRJNA594799 BioProject 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-34719737

Paper: Iftakhar-E-Khuda et al. (2021) Vascular adhesion protein-1 defines a unique subpopulation of human hematopoietic stem cells and regulates their proliferation. Cell Mol Life Sci. PMID 34719737 · PMCID PMC8629906 · DOI 10.1007/s00018-021-03977-6.

Data accessions (clarified — the BRIEF pointer is the BULK set)

  • PRJNA594799 = BULK RNA-seq (this is the accession in BRIEF.md).
  • PRJNA729883 = single-cell RNA-seq (SRA study SRP319821) — the source of the headline computational/Nebulosa results. 4 sorted-HSC 10x samples (v3.1 dual-index), raw FASTQ only:
    • S2 210011_2_VAP_1_neg_replicate2 (VAP-1⁻ rep2)
    • S3 210011_3_VAP_1_neg_replicate3 (VAP-1⁻ rep3)
    • S4 210011_4_VAP_1_pos_replicate1 (VAP-1⁺ rep1)
    • S5 210011_5_VAP_1_pos_replicate2 (VAP-1⁺ rep2)
  • No GEO Series / no processed count matrix / no Seurat object is shipped (GDS search empty; SRA records carry no GSM cross-ref). "Code availability" section is blank — no author code repo. → reproduction = apply the described third-party pipeline to the deposited FASTQ (valid per BRIEF rule P16).

Reported pipeline (Methods)

10x Chromium 3' v3.1 → Cell Ranger 5.0.1 (GRCh38) → Seurat 4.0.1 (R 4.0.5; FindVariableFeatures/ScaleData/RunPCA/FindMarkers Wilcoxon) → UMAP → Monocle 2 v2.10.1 (trajectory) → Nebulosa (kernel density estimation of gene expression). Nebulosa is the third-party GitHub tool in the BRIEF (powellgenomicslab/Nebulosa) and IS genuinely used.

IN SCOPE (pipeline-derived, reproducible)

  • C1 AOC3/VAP-1 mRNA "practically negative" in the HSC scRNA-seq (the central Nebulosa-shown point). Robust, low-QC-sensitivity. → reproduce.
  • C2 ~1267 high-quality HSC after QC (aggregate). → reproduce (approx; exact count is QC-threshold-sensitive = hard 20%).
  • C3 Two clusters HSC1 (93%) / HSC2 (7%, ≈89 cells). → reproduce (proportion).
  • C4 HSC2 enriched for proliferation / cell-cycle (S.Score, G2M.Score). → reproduce via Seurat CellCycleScoring.

DEVIATION (transparent)

  • Cell Ranger 5.0.1 → STARsolo in CellRanger-emulation mode. Cell Ranger old-version binaries are license-gated (no public direct URL). STARsolo with CR-emulation flags (CB16/UMI12, 3M-february-2018 whitelist, --soloUMIdedup 1MM_CR, --soloCBmatchWLtype 1MM_multi_Nbase_pseudocounts, --soloUMIfiltering MultiGeneUMI_CR, --clipAdapterType CellRanger4, --soloCellFilter EmptyDrops_CR) is the accepted free equivalent. Genome built from the public 10x GRCh38-2020-A fasta+GTF (same reference build the paper used). Counts will not be byte-identical to Cell Ranger; this is disclosed and graded accordingly.

OUT OF SCOPE (not attempted / hard 20%)

  • Bulk RNA-seq DE marker derivation (PRJNA594799) — specific reported value not pinnable from the open text; would be a separate alignment+DESeq2 effort.
  • Monocle 2 pseudotime trajectory — depends on exact cell set & is highly parameter-/seed- sensitive; not a crisp reported number.
  • Exact 1267 / 89 cell counts to the unit — QC thresholds (mito%, nFeature) underspecified; we report our values and grade as partial/within-order rather than claim exact.
  • All wet-lab / FACS / functional-proliferation assays — non-pipeline.
C1
Reported
AOC3/VAP-1 mRNA 'practically negative' in HSC scRNA-seq (Nebulosa density)
Reproduced
not obtained
partial
C2
Reported
1267 high-quality HSC after QC
Reproduced
not obtained
partial
C3
Reported
HSC1 93% / HSC2 7% (HSC2 = 89 cells)
Reproduced
not obtained
partial
C4
Reported
HSC2 / VAP-1+ enriched S.Score & G2M.Score
Reproduced
not obtained
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)
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

This is an honest partial / setup-only reproduction: the agent correctly resolved a data-accession mismatch (bulk PRJNA594799 vs the real scRNA-seq PRJNA729883), rebuilt the GRCh38-2020-A reference + STAR index, and staged the full STARsolo→Seurat→Nebulosa pipeline, but an our-side download bug (32 zero-byte FASTQ files) meant no count matrix and zero graded numeric claims. The deviations that exist are on our/data-availability side — no processed matrix shipped, blank code-availability section, Cell-Ranger→STARsolo substitution — not evidence against the authors. The internally-consistent arithmetic (89 = 7% of 1267) is neither confirmed nor refuted, so there is no fabrication signal; criticality is yellow purely because the reproduction is incomplete, not because any claim broke.

🤝
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.

155 k
tokens (I/O) · 8.6 M incl. cache
28 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.