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

Spatially clustered loci with multiple enhancers are frequent targets of HIV-1 integration.

Nat Commun · 2019
L1 77/100 3/4
Why this verdict

The main results reproduced: recomputed values matched the published ones within tolerance.

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: 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: Q6 · Severity of the deviation 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +4
✓ What held up
  • Reported values were directly comparable
  • The central claim held under reproduction
What did not (or only partly)
  • 🟡Could not use the authors’ exact input data
  • 🟡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
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
77/100
Reproducibility score
at the mean
vs. all fields · 1173 studies
🎯 Scores higher than 50% of all assessed papers rank 572 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: yes for the self-contained maja/ R analyses, which ship their own derived data objects (genidf.Robj, is.Robj, IS.txt, roc.Robj). Mostly 1:1. Reproduced from shipped data on «our HPC» (R 4.3.3 + GenomicRanges 1.54.1), repo @ 5c35ee8: C1 total integration sites 13,550 EXACT (in-vitro 4031 and patient 9519 subtotals both exact); C3 non-RIG protein-coding genes 13,140 EXACT; C4 super-enhancer-overlapping RIGs count 564 EXACT (reported %34.22 uses denominator 1648; shipped RIG total is 1605 -> 35.14%); C5 ROC Fig 1B directionality 9/9 marks match (enriched/ns/depleted). PARTIAL/DIFFERENT: C2 exact RIG total 1648 not re-derivable because the Brady-redefinition step needs two inputs absent from the repo (GRCh37AllGenes.Robj, IntegrationSitesInActivatedALLCells_bushman.txt); shipped genidf yields 1605 -- the 43-gene gap reconciles consistently with C4 (all added RIGs are non-SE). NOT ATTEMPTED: Hi-C .cool generation + A1/A2/B1/B2/AB sub-compartment clustering (Fig 4, clusters/ branch, large GSE122958 cool) -- deferred heavy last-20%; TableM1 %-IS-in-genes (blocked, un-shipped inputs); wet-lab + ngsplot metagene (non-pipeline); full independent ROC recompute (controls object structure mismatch). No fabrication indicators: the hardest-to-fake counts reproduce to the digit.

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 77
    assessed: 2026-06-14 ⛓ 14e7b69acd83
✎ 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
2026-06-14
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

What are the genomic and spatial nuclear features of genes recurrently targeted by HIV-1 integration in CD4+ T cells, and is the transcriptional activity of these genes the determinant of integration, or is it their spatial organization relative to super-enhancers?

Core claims
  • HIV-1 recurrently integrates into genes that are proximal to super-enhancer (SE) genomic elements in both patients and in vitro T cell cultures. finding
  • Recurrent integration genes (RIGs) are proximal to SEs irrespective of their transcriptional levels, and disruption of SE activity (JQ1) does not alter HIV-1 integration patterns. finding
  • HIV-1 insertion hotspots cluster together in 3D nuclear space and preferentially contact super-enhancers, occupying the same 3D sub-compartment. finding
  • Spatial clustering of targeted genes together with their transcriptional activity are the major determinants of HIV-1 integration, with SEs contributing indirectly via genome reorganization during T cell activation. mechanism
  • RIGs were defined as genes with ≥1 HIV-1 integration in at least 2 of 8 datasets, yielding 1648 RIGs, used to compare chromatin/expression features against non-RIGs. method
  • RIGs show higher H3K27ac, H3K4me1, H3K4me3, BRD4, MED1, H3K36me3, and H4K20me1 and lower H3K27me3/H3K9me2 relative to non-RIGs. finding
  • A high-resolution Jurkat Hi-C dataset (~1.5 billion contacts) was generated and used to segment the genome into spatial sub-compartments (A1, A2, B1, B2, AB). resource
  • HTLV-1 integration sites are not enriched in SE marks whereas MLV shows strong enrichment, distinguishing HIV-1's SE-proximal bias. finding
Experimental setups
Assay System Perturbation Readout Platform
HIV-1 integration site mapping (compiled datasets + inverse PCR + linear amplification-mediated PCR) activated primary CD4+ T cells (in vitro infection) and HIV-1 patient samples HIV-1 infection genomic location/frequency of proviral integration sites
ChIP-Seq primary CD4+ T cells (activated) none genomic levels of H3K27ac, H3K4me1, H3K4me3, BRD4, MED1, H3K36me3, H4K20me1, H3K9me2, H3K27me3; SE identification
RNA-Seq CD3/CD28-activated primary CD4+ T cells none (baseline expression) transcript abundance (regularized log read counts) of protein-coding genes
RNA-Seq (transcriptional profiling) activated CD4+ T cells JQ1 (BET/BRD4 bromodomain inhibitor) differential expression of SE-proximal vs non-SE genes, RIGs vs non-targeted
HIV-1 integration site mapping by inverse PCR CD4+ T cells JQ1 treatment vs control HIV-1 insertion profiles (38,964 sites mapped)
Hi-C uninfected Jurkat lymphoid T cells none genome-wide chromosomal contact frequencies; TADs, loop domains, A/B compartments, 3D sub-compartments; inter-chromosomal contact density
RNA and protein quantification (MYC control for JQ1) CD4+ T cells JQ1 treatment MYC RNA and protein levels
Key results
  • The more datasets a RIG is found in, the closer it lies to super-enhancers on average.
  • 21.4% of protein-coding genes targeted by HIV-1 are in the top 10% most expressed genes vs 6.07% of non-targeted genes. 21.4% vs 6.07%
  • 19.05% of silent RIGs have a proximal SE, versus only 1.5% of silent genes never targeted by HIV-1. 19.05% vs 1.5%
  • 2584 SEs identified, intersecting 564 RIGs (34.22%). 564/1648 = 34.22%
  • Loci most targeted by HIV-1 engage in stronger inter-chromosomal Hi-C contacts with each other than non-targeted loci, and SEs cluster with HIV-1 hotspots in 3D.
  • JQ1 treatment does not alter HIV-1 insertion biases at chromosome scale nor spatial localization of provirus/RIGs.
  • Insertion rate per chromosome is similar between primary T and Jurkat cells, with ~3-fold increase on chromosomes 17 and 19. ~3-fold (chr17, chr19)
  • HTLV-1 insertion sites not enriched in SE marks; MLV strongly enriched in all SE marks.
Key statistics
  • count 4031 HIV-1 integration sites from in vitro activated primary CD4+ T cells (in vitro infection integration sites assembled)
  • count 9519 insertion sites from 6 HIV-1 patient studies (patient-derived integration sites)
  • count 10,735 integrations in gene bodies (77% patient avg, 84% in vitro avg) targeting 5601 genes (integrations within gene bodies)
  • count 1648 RIGs (recurrent integration genes defined (≥1 integration in ≥2 of 8 datasets))
  • count 13,140 non-RIG protein-coding genes (comparison set without HIV-1 insertions)
  • pvalue <2.2 x 10^-16 (Wilcoxon rank-sum test of mRNA abundance difference for genes without HIV integrations and genes on one list (SE vs no SE))
  • pvalue 3.7 × 10^-12 (Wilcoxon rank-sum test of mRNA abundance difference for RIGs (SE vs no SE))
  • count ~1.5 billion informative Hi-C contacts (Jurkat Hi-C dataset depth)

Statistical methods review

Model: opus

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 is a genomics/computational study characterizing HIV-1 integration site preferences in CD4+ T cells, integrating published and newly generated integration-site datasets with ChIP-Seq, RNA-Seq, and Hi-C data. Group comparisons of genomic/expression features between recurrently targeted genes (RIGs) and non-targeted genes were assessed mainly with the non-parametric Wilcoxon rank-sum test, enrichment of chromatin marks at integration sites was quantified using ROC area analysis against distance-matched control sites, and distributions were displayed primarily as box and violin plots. Exact p-values were reported for the key expression comparisons.

Replicationmixed Sample size4031 in vitro and 9519 patient integration sites compiled; 1648 RIGs vs 13,140 non-RIGs; RNA-Seq averaged over three replicates; Hi-C ~1.5 billion contacts; formal power/sample-size justification not stated GroupsRIGs vs non-RIGs; SE-proximal vs SE-distal; active vs silent; HIV vs No-HIV loci Pairingunpaired Randomization/blindingna DispersionIQR Exact p-valuesyes Effect sizesno Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Wilcoxon rank-sum (Mann-Whitney) test differences in median mRNA abundance between genes with vs without a proximal super-enhancer, across gene groups (Fig. 2c) expression averaged over three RNA-Seq replicates; gene-level n not stated numerically not stated
ROC area (area under ROC curve) analysis co-occurrence/enrichment of integration sites with each ChIP-Seq epigenetic mark vs distance-matched control sites (Fig. 1b) na
Inter-chromosomal Hi-C contact density comparison across gene classes (box-plot distributions) contact strength among active/silent, HIV/No-HIV, SE/No-SE gene aggregates (Fig. 3d, e) not stated
Approaches that could also have been used
  • Median mRNA abundance between gene groups was compared with the non-parametric Wilcoxon rank-sum test.
    Could also: A permutation/bootstrap test, or a generalized linear model on counts (e.g., negative-binomial regression as in DESeq2) with expression group and SE status as covariates. — A modeling approach would also estimate effect sizes and let one jointly account for expression level and SE proximity, complementing the rank-based two-group test.
  • Enrichment of chromatin marks at integration sites was quantified using the ROC area method against distance-matched control sites.
    Could also: Logistic regression or a precision-recall (PR) analysis could also summarize the same association. — PR curves are often informative under class imbalance, and a regression framework would allow simultaneous adjustment for several genomic covariates while still yielding an enrichment estimate.
  • Multiple group comparisons (e.g., several gene groups in Fig. 2c) were each reported with a p-value.
    Could also: A family-wise or false-discovery-rate adjustment (e.g., Benjamini-Hochberg, Bonferroni, or a Kruskal-Wallis test with post-hoc comparisons) could also be applied across the related comparisons. — An explicit multiplicity adjustment controls the error rate across the family of tests and is commonly reported when several related comparisons are made.
  • Spread of distributions was conveyed primarily through box plots and violin plots.
    Could also: Reporting accompanying effect-size measures (e.g., Cliff's delta, rank-biserial correlation, or median differences with 95% confidence intervals) could also be included. — Effect sizes and CIs convey the magnitude and precision of differences, which complements significance values, especially given very small p-values on large gene sets.
  • Integration-site enrichment used control sites matched on distance to the nearest gene.
    Could also: Multiple matched control sets or a covariate-balanced random-control resampling scheme could also be used. — Repeated random control draws would provide an empirical null distribution and a sense of variability around the enrichment estimates.
Software: DESeq2 (implied by regularized log-transformed read counts for RNA-Seq)

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
122
Impact: high
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.

What was reproduced

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

C1
Reported
13550 integration sites (4031 in-vitro + 9519 patient)
Reproduced
13550; subtotals 4031 & 9519 exact
exact
C3
Reported
13140 non-RIG genes
Reproduced
13140 (protein-coding, 0 lists)
exact
C4
Reported
564 RIGs intersect super-enhancers (34.22%)
Reproduced
564 (count exact); 35.14% of shipped 1605 RIGs
within tolerance
C2
Reported
1648 RIGs (gene in >=2 of 8 datasets)
Reproduced
1605 all / 1509 protein-coding (shipped genidf predates Brady-redefinition; required GRCh37AllGenes.Robj + bushman.txt not shipped)
partial
C5
Reported
Fig 1B ROC: H3K27ac/H3K4me1/BRD4/MED1/H3K36me3/H4K20me1 enriched(>0.5), H3K4me3 ns(~0.5), H3K27me3/H3K9me2 depleted(<0.5)
Reproduced
9/9 directions match from shipped roc.Robj @bin1000 (e.g. H3K27Ac 0.637, H3K4me3 0.492, H3K27me3 0.375); independent recompute blocked (controls shipped as list RS, script expects matrix IS_controls_075 not shipped)
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 77/100

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: 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: Q6 · Severity of the deviation 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +4

Counts reproduce to the digit (13,550 with exact 4031/9519 split; 13,140 non-RIGs; 564 SE-overlapping RIGs) and Fig 1B ROC directionality matches 9/9, so the core claim holds and there is no fabrication signal. The only real deviation is the exact RIG total 1648 vs shipped 1605 — not re-derivable because two redefinition inputs (GRCh37AllGenes.Robj, bushman activated-cell IS) were never shipped, also explaining the C4 % drift (34.22% uses 1648 as denominator vs 35.14% on 1605). The gap is on the authors'/deposit side (incomplete repo, value not derivable from shared data) but is small (+43 genes, all non-SE) and reconciles consistently rather than contradicting, so overall a yellow-quality reproduction.

🤝
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.

125.2 k
tokens (I/O) · 4.8 M incl. cache
12 min
runtime · 0.01 CPU-h
1.9 GB
peak RAM
3
HPC jobs
hummel
machine