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

Profiling chromatin accessibility responses in human neutrophils with sensitive pathogen detection.

Life Sci Alliance · 2021
L1 54/100 3/4
⚑ 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 results reproduced, with only marginal, non-material deviations.

Reproduced on the brainbox compute brainarbeit.com
✓ What held up
  • Same input data as the authors
  • Reported values were directly comparable
What did not (or only partly)
  • 🔴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
54/100
Reproducibility score
1.1 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 14% of all assessed papers rank 997 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, honest. The brief's accession GSE153521 is the RNA-seq SubSeries (ATAC = GSE153520; SuperSeries GSE153522); the brief's code link (FelixKrueger/TrimGalore) is a text-mining artifact (one cited tool) — the authors' actual repo is nikhilram/neutrophil_ATACseq (ATAC-only: PEPATAC + DiffBind). RNA-seq reproduction (P16-style: described edgeR pipeline on the deposited raw-counts matrix, 25,702 genes x 8 libraries) was run directly with edgeR 4.4.2/R 4.4.2 on «our HPC» front1. Canonical config full_LRT_TMM (donor-blocked GLM, TMM, glmLRT, filterByExpr, FDR<0.05 & |logFC|>=1). RESULT: UP-regulation reproduces well — E.coli-4h up 2551 vs reported 2554 (essentially exact), E.coli-1h up 69 vs 66, total-4h DE within ~6% — but DOWN-regulation is systematically far below reported (1h 3 vs 55; 4h 2338 vs 2656; consistent-down 1 vs 10). A 24-config sweep (norm none/TMM/UQ/RLE x exact/QLF/LRT/Treat x pairwise/paired/full-donor designs) confirms no principled config reproduces both up and down counts simultaneously; only norm='none' lifts the down counts but then grossly overshoots 4h-down (~6500) and breaks the up side. Flagged for audit: the reported RNA-seq down counts are not cleanly derivable from the deposited data with standard edgeR (possible unstated normalization/filter, version effect, or reporting inconsistency). NOT attempted: (a) ATAC-seq DiffBind DAR table — deposited ATAC processed files are single-column merged coverage per condition, lacking the per-replicate counts/BAMs the authors' run_diffbind.R needs; reproducing it requires full PEPATAC re-alignment from SRA FASTQ (SRP265675, ~50 libraries) which is currently impossible because the shared «our HPC» account is OVER QUOTA on /home and /«infra» and the «infra» scheduler rejects job output on /usw (no SLURM job submittable); (b) Kraken pathogen detection — no script shipped, needs raw reads + DB. Datasets profiled in-pass: GSE153521 RNA-seq counts (grade A, delivers yes), GSE153520 ATAC coverage (grade C, delivers partial, n mismatch 13 deposited vs ~50 samples).

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 54
    assessed: 2026-06-18 ⛓ 40f03efef3b7
✎ 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
2026-06-18
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-19
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

Because the epigenome reacts before gene expression, the authors test whether profiling chromatin accessibility (ATAC-seq) responses in human neutrophils to different pathogen ligands and whole organisms can reveal challenge-specific and time-specific epigenomic signatures for early disease recognition, and how chromatin accessibility changes regulate downstream transcription.

Core claims
  • ATAC-seq reveals unique neutrophil chromatin accessibility changes in response to different stimuli before transcriptional activation, with most differential regions being challenge-specific in position, function, and motif. finding
  • ATAC-seq of neutrophils enriches pathogen DNA, enhancing sensitive detection of microbial reads compared with traditional library preparation. finding
  • Neutrophil epigenomic changes are plastic over time, with only ~120 differential regions shared between E. coli challenges at 1 and 4 h, producing varied differential genes and associated processes. finding
  • Three classes of gene regulation are identified: chromatin access changes in the promoter; changes in the promoter and distal enhancers; and control of expression solely through distal enhancer changes. mechanism
  • Coupling neutrophil ATAC-seq host-response profiling with enriched microbial read detection in a single assay offers diagnostic potential for sepsis/bloodstream infections. method
  • Transcription factor footprinting plus positional and functional analysis reveal timely and challenge-specific mechanisms of transcriptional regulation in neutrophils. finding
Experimental setups
Assay System Perturbation Readout Platform
ATAC-seq Purified human neutrophils from healthy female volunteers 1 h challenge with TLR ligands (LTA/TLR2, LPS/TLR4, flagellin/TLR5, R848/TLR7-8, β-glucan peptide/dectin-1, HMGB1/DAMP) Genome-wide chromatin accessibility / differential accessible regions Tn5 transposase, Illumina sequencing
ATAC-seq Human neutrophils (and whole blood for S. aureus) Whole organism challenge with S. aureus and E. coli (1 h and 4 h) Differential chromatin accessibility and pathogen DNA reads Tn5 transposase, Illumina sequencing
RNA-seq Human neutrophils E. coli exposure for 1 h (EC1h) and 4 h (EC4h) Differential gene expression (logFC)
Genome-wide DNA sequencing (SPRI library prep comparison) Whole blood with negatively isolated neutrophils Live S. aureus spike at incremental CFU/ml for 1 h Relative abundance of pathogen reads vs ATAC-seq Solid-phase reversible immobilization (SPRI) library preparation
qRT-PCR Healthy donor human neutrophils Ligand or live organism challenge IL8 and TNFα expression (neutrophil activation confirmation)
SYTOX green assay Healthy volunteer human neutrophils Pathogen ligands (1 h) or live organism; PMA positive control Extracellular DNA as indication of NET formation
Key results
  • EC1h showed the most differential regions (5,010), with E. coli challenges producing strongly time-specific accessibility changes (EC4h: 1,688 DRs) 5,010 (EC1h) vs 1,688 (EC4h) DRs
  • Majority of differential regions are unique/challenge-specific; ~69.37% unique for ligand challenges and ~91% unique for whole organism challenges, with no DRs shared across all challenges ~69.37% (ligands), ~91% (whole organisms) unique
  • ATAC-seq retained higher relative abundance of S. aureus reads than SPRI at all concentrations; abundance at 10^3 CFU/ml by ATAC-seq comparable to 10^5 CFU/ml by SPRI; ~3x more pathogen reads 3-fold; 10^3 vs 10^5 CFU/ml equivalence
  • Only 118 (~120) differential regions shared between the two E. coli time points, indicating epigenomic plasticity 118 shared regions
  • RNA-seq showed marked temporal increase in differentially expressed genes: EC1h had 66 up/55 down, EC4h had 2,554 up/2,656 down regulated genes EC1h: 66 up/55 down; EC4h: 2,554 up/2,656 down
  • More than 40% of differential regions in each challenge located in distal intergenic or intronic regions, with similar genomic distribution across challenges >40% distal/intronic; >80% distal
  • On average ~95.8% (minimum ~89.2%) of differential regions were successfully associated with genes across challenges ~95.8% average, ~89.2% minimum
  • No NETs observed at 1 or 4 h of stimulation by SYTOX green assay, supporting nuclear integrity for ATAC-seq
Key statistics
  • correlation r2 = 0.70–0.95 (Genome-wide peak count correlation across ATAC-seq technical replicates per ligand)
  • correlation r2 ranging from 0.92 to 0.99 (RNA-seq correlation between replicates)
  • count 5,010 DRs (4,625 unique, 92%) (Differential regions for EC1h challenge vs unstimulated)
  • count 2,241 DRs (2,121 unique, 95%) (Differential regions for S. aureus challenge)
  • count 118 common DRs (Shared differential regions between EC1h and EC4h E. coli time points)
  • count 2,554 up- and 2,656 down-regulated genes (Differentially expressed genes at EC4h by RNA-seq)
  • fold_change 3 times more reads (ATAC-seq vs SPRI for S. aureus pathogen reads)
  • count 4,506 / 4,498 overlaps (EC1h differential regions overlapping each end of Hi-C interacting regions)

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.

Human neutrophils from healthy volunteers (n=4 for ligand/DAMP challenges, n=2 for whole-organism challenges) were profiled by ATAC-seq across nine conditions versus paired unstimulated controls; differential chromatin accessibility was identified using DiffBind (P<0.05, |logFC|≥1). RNA-seq from E. coli–challenged neutrophils at two time points was analyzed with edgeR (P<0.05). Motif enrichment in differential regions was assessed with HOMER (P<10⁻¹⁰), and functional annotation used ChIPseeker and clusterProfiler; results were reported primarily as region counts, logFC values, UpSet-plot overlaps, and pathway enrichment comparisons.

Replicationbiological Sample sizeFour female healthy volunteers for ligand and DAMP challenges (n=4); two donors for whole-organism (S. aureus, E. coli) challenges (n=2); technical replicates also performed for ATAC-seq (r²=0.70–0.95) GroupsNine challenge conditions (LTA, LPS, FLAG, R848, BGP, HMGB1, S. aureus, E. coli 1h, E. coli 4h) each vs. unstimulated neutrophil controls from the same donors Pairingunclear Randomization/blindingnot stated DispersionSEM Exact p-valuesno Effect sizesyes Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
DiffBind (differential accessibility analysis; internally uses DESeq2 or edgeR on count data) Identification of differentially accessible chromatin regions (DRs) for each of nine challenges vs. unstimulated control (ATAC-seq) n=4 donors for ligand/DAMP challenges; n=2 donors for whole-organism challenges not stated
edgeR (negative binomial GLM) Differential gene expression at E. coli 1 h and 4 h vs. unstimulated control (RNA-seq) not stated explicitly not stated
Pearson correlation (r²) Quality control of ATAC-seq technical replicates per donor and RNA-seq replicate concordance not stated not stated
HOMER motif enrichment (hypergeometric/binomial) Transcription factor motif enrichment in induced and repressed differential ATAC-seq regions per challenge null not stated
clusterProfiler compareCluster (hypergeometric/Fisher's exact) Reactome pathway enrichment of genes associated with differential chromatin regions across all challenges null not stated
Approaches that could also have been used
  • Nine challenges were each compared independently to unstimulated controls in separate DiffBind analyses
    Could also: A single multi-condition model (e.g., DESeq2 or edgeR with a multi-level factor and explicit contrasts, or limma-voom on count matrices) could analyze all conditions jointly with a shared dispersion estimate — Pooling information across groups in a unified model improves dispersion estimation, which can increase sensitivity particularly when per-group n is small (n=2–4); it also provides a natural framework for controlled pairwise contrasts with consistent error variance
  • Differential chromatin regions and differential genes were filtered at nominal P<0.05 thresholds; no explicit within-test FDR correction is described
    Could also: Applying Benjamini-Hochberg FDR correction within each DiffBind and edgeR analysis and reporting adjusted q-values (e.g., FDR<0.05 or FDR<0.10) is standard for genome-wide omics analyses — FDR-adjusted thresholds are the conventional standard for ATAC-seq and RNA-seq to control the expected proportion of false discoveries across thousands of simultaneous region- or gene-level tests
  • Dispersion for cytokine and SYTOX assays is reported as mean ± SE at n=2–4
    Could also: Mean ± SD, or individual data points overlaid on bar/line plots, could also convey spread; 95% CIs would additionally communicate estimation uncertainty — At small n, SD reflects actual biological variability rather than precision of the mean estimate; showing individual donor values is increasingly recommended by journals for small-n biological assays so readers can assess the underlying distribution
  • Overlap between challenge-specific differential regions was visualized with UpSet plots restricted to the top 100 regions
    Could also: Jaccard similarity indices or permutation-based overlap significance testing (e.g., regioneR) could also quantify the degree of sharing between any two conditions across the full region set — Quantitative overlap statistics complement visual UpSet plots by providing a normalized similarity measure and a null-distribution-based p-value for whether observed overlaps exceed chance expectation genome-wide
  • RNA-seq differential expression was performed with edgeR
    Could also: DESeq2 with its regularized log-fold-change shrinkage (lfcShrink) and Benjamini-Hochberg adjusted p-values is a widely used alternative for small-n RNA-seq experiments — DESeq2's shrinkage estimators stabilize logFC estimates for low-count genes, which can be frequent in neutrophils given their overall lower transcriptional activity; both tools are considered standard and results are often compared or cross-validated
  • No a priori power analysis or sample size justification is described, with n=2 for whole-organism conditions
    Could also: A formal power simulation based on pilot effect size estimates, or a post-hoc sensitivity analysis reporting the minimum detectable effect at the achieved n and significance threshold, could also be included — Reporting a power calculation or detectable-effect statement helps readers contextualize the biological interpretability of non-significant or variable results, especially for the n=2 whole-organism comparisons
Software: DiffBind (R/Bioconductor) · edgeR (R/Bioconductor) · ChIPseeker (R/Bioconductor) · clusterProfiler (R/Bioconductor) · HOMER · GREAT · T-gene · UpSetR (R)

What was reproduced

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

Scope — pmid-34145026

Paper: Ram-Mohan et al. 2021, Life Sci Alliance — "Integrative profiling of early host chromatin accessibility responses in human neutrophils with sensitive pathogen detection." PMID 34145026 · PMCID PMC8321655 · DOI 10.26508/lsa.202000976.

Accessions (brief's GSE153521 is the RNA-seq sub-series)

  • GSE153522 — SuperSeries (umbrella).
  • GSE153520 — ATAC-seq sub-series (peak-coverage BEDs deposited; raw FASTQ in SRP265675).
  • GSE153521 — RNA-seq sub-series (raw-counts matrix deposited; raw FASTQ in SRP269266). This is the accession named in the brief.

Code

  • Brief lists github.com/FelixKrueger/TrimGalore — that is just one cited tool (RNA-seq adapter trimming), a text-mining artifact, not the authors' repo.
  • Authors' own repo: github.com/nikhilram/neutrophil_ATACseq — ships run_pepatac.sh, run_diffbind.R, and 4 Perl post-processing scripts. It is an ATAC-only repo: no RNA-seq edgeR script, no Kraken script.

Pipelines named in Methods

  • ATAC-seq: PEPATAC (Trimmomatic → Bowtie2 --very-sensitive -X 2000 hg19 → Picard dedup → SAMtools MAPQ<10, drop chrM/chrY → MACS2 -q 0.01 --shift --nomodel) → DiffBind (dba.report th=0.05 bUsePval=TRUE fold=1, 0.66 consensus overlap) → ChIPseeker / HOMER (P<1e-10) / TOBIAS footprinting.
  • RNA-seq: FastQC → Trim Galore → HISAT2 --rna-strandness RF hg19 → Rsubread featureCounts (strand-specific) → edgeR (FDR<0.05, |logFC|≥1) → clusterProfiler.
  • Pathogen detection: Kraken on human-depleted reads, abundance as counts-per-million.

In scope (pipeline-derived, reproducible from deposited processed data)

  1. RNA-seq differential expression (edgeR) — PRIMARY. The deposited GSE153521_Raw_counts_for_each_replicate.txt.gz is the exact featureCounts matrix (25,702 genes × 8 libraries: 2 reps × {EC-1h, EC-4h, noEC-1h, noEC-4h}). Re-running edgeR (FDR<0.05, |logFC|≥1) reproduces the DEG counts (paper Fig 5 / text): EC-1h 66 up / 55 down; EC-4h 2554 up / 2656 down; consistent across both time-points 93 up / 10 down. Deterministic, low-compute.

Out of scope or harder

  1. ATAC-seq DiffBind DAR counts (LTA 1331, LPS 1729, FLAG 2963, R848 3105, β-glucan 2030, HMGB1 2930, S.aureus 2241, EC-1h 5010, EC-4h 1688 — Table/Fig 2). The authors' run_diffbind.R needs per-replicate BAMs + per-replicate peaks; GEO deposits only single-column merged-coverage BEDs per condition (per-replicate resolution lost). Exact DAR reproduction therefore requires re-running full PEPATAC from raw FASTQ (≈50 ATAC libraries, hg19 alignment) — heavy; attempted only if the compute path opens.
  2. Kraken pathogen detection (S. aureus 3× higher abundance vs SPRI; 100-fold sensitivity gain) — no script shipped; needs raw reads + a Kraken DB. Harder.
  3. Motif/footprint/Hi-C association numbers — HOMER/TOBIAS/GREAT; downstream, not attempted in the first pass.

Infrastructure note

Shared «our HPC» account («user») is over quota on /home AND «infra», front1 /tmp is full, and the «infra» submit filter forbids /usw for job stdout → no SLURM job can be submitted at present. Workaround for the tiny RNA edgeR step: run R 4.4.2 (from the existing wp9_rnavar env) directly on front1, edgeR compiled into a /usw personal library. Heavy ATAC/Kraken work stays blocked until the account quota frees.

Figures / tables: Fig5Fig2Table
rna_ec1h_up
Reported
66 up DEGs (E.coli 1h)
Reproduced
69
within tolerance
rna_ec1h_dn
Reported
55 down DEGs (E.coli 1h)
Reproduced
3
did not match
rna_ec4h_up
Reported
2554 up DEGs (E.coli 4h)
Reproduced
2551
exact
rna_ec4h_dn
Reported
2656 down DEGs (E.coli 4h)
Reproduced
2338
partial
rna_ec4h_total
Reported
5210 total DEGs (E.coli 4h)
Reproduced
4889
within tolerance
rna_cons_up
Reported
93 consistently up (1h&4h)
Reproduced
62
partial
rna_cons_dn
Reported
10 consistently down (1h&4h)
Reproduced
1
did not match
atac_dar_table
Reported
DiffBind DAR counts per condition (LTA 1331 ... EC-1h 5010)
Reproduced
NOT_ATTEMPTED
partial
pathogen_kraken
Reported
S.aureus 3x abundance via ATAC vs SPRI (Kraken)
Reproduced
NOT_ATTEMPTED
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 54/100

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.

252.5 k
tokens (I/O) · 13.7 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.