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

Identity rather than 3D position informs splicing of rare introns in the human genome.

iScience · 2025
L1 98/100 PQI 95
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: Q7 · Core claim 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score 0
✓ What held up
  • Same input data as the authors
  • Reported values were directly comparable
  • No relevant deviation in data/preprocessing
  • No authors-side cause for any deviation
  • Reported values are derivable from the shared data
  • Any deviation was negligible
  • Overall, the reproduction was clean
What did not (or only partly)
  • 🟡The central claim did not (fully) hold under reproduction
How its reproducibility compares
98/100
Reproducibility score
1.4 SD above mean
vs. all fields · 1173 studies
🎯 Scores higher than 93% of all assessed papers rank 65 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 reproduce 1:1. The repo (saren-springer/introns-3Dgenome @756b1e7) ships its own GRCh38/Ensembl-v99 intron annotations, so the Figure-1 core analyses need no external download. All six intron-class counts (major 345317, major-like 29337, non-canonical 6070, minor 850, minor-like 458, hybrid 373) reproduce EXACTLY. Running the repo's verbatim inter_intron_distance.py (SLURM «job») and taking the per-class median reproduces the Fig 1C inter-intron distances: major 141 bp (exact), non-canonical 105 kb (exact), minor 1.84e6 (exact), minor-like 4.04e6 (exact); major-like 27.8 kb and hybrid 5.91e6 are within rounding of the reported ~27 kb / ~6.1e6 (within-tol). The Fig 1D 'R^2=0.98' is reproduced exactly as a log10-log10 fit of class size vs median distance (R^2=0.982; the non-log fit is only 0.19, identifying the transform). NOT attempted (the hard ~20%): (a) the 250 kb intron-density table — deterministic but no single pinnable scalar; (b) chromoMap Fig 1E — visualization only; (c) the bootstrapping/pollution 3D-compartment proportions (Fig 2, e.g. 'minor 71% in compartment A, p=4.7e-08') because the GSE63525-derived A/B/sub-compartment master overlap table is NOT shipped (only an example table is) and the step is randomized 1000x. No fabrication signal: every graded value is directly derivable from shipped data+code. Reported Fig 1C/1D values were read from the PMC article via automated extraction; a human should confirm exact panel numbers, but agreement across all six classes is far too tight to be coincidental.

💻 Code ↗ 🗄 Data: GSE63525

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 98
    assessed: 2026-06-16 ⛓ ba4fcf3df6c0
✎ 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
2026-06-16
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-16
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 authors hypothesize that 3D genome architecture spatially organizes minor introns and other rare intron classes into privileged nuclear regions (e.g., compartment A and speckle-associated domains) to ensure their proper splicing, and they test whether nuclear position versus intron identity governs splicing efficiency.

Core claims
  • Splicing efficiency depends on intron identity rather than nuclear (SPAD) position; despite shared SPAD proximity, major-like and minor-like introns are less efficiently spliced than major and minor introns finding
  • The six intron classes (major, major-like, hybrid, minor-like, minor, non-canonical) exhibit distinct, non-random spatial positioning biases within the 3D genome finding
  • Minor intron-containing genes and their introns are enriched in compartment A and speckle-associated domains (SPADs) and depleted from lamina-associated domains (LADs) finding
  • Minor introns outside SPADs are more efficiently spliced in cancer cells, indicating minor intron splicing in cancer is buffered against spatial constraints finding
  • Rare introns are dispersed/scattered across the linear human genome, with inter-intron distance increasing as intron-class abundance decreases finding
  • Spatial enrichment of minor introns is tethered to intron identity, not gene identity or intron number, as shown by bootstrapping and pollution/permutation analyses finding
  • Mapped six intron classes across nuclear compartments, subcompartments, nuclear bodies, and 3D features by integrating Hi-C, TSA-seq, DamID-seq, and RNA-seq datasets method
  • MIGs with enhanced splicing in cancer cell lines are overrepresented for cell cycle regulators finding
Experimental setups
Assay System Perturbation Readout Platform
Hi-C (3D genome compartment/subcompartment/TAD/loop/boundary mapping) K562 human chronic myelogenous leukemia cells none intron-class distribution across compartments A/B, subcompartments A1/A2/B1/B2/B3, TADs, boundaries, loop anchors
TSA-seq K562, HFFc6 (fibroblast), H1 (embryonic stem cells), HCT116 (colorectal cancer) none speckle-associated domains (SPADs) / intron proximity to nuclear speckles
DamID-seq K562 (and additional human cell lines) none lamina-associated domains (LADs)
RNA-seq (transcriptome integration) K562, HFFc6, H1, HCT116 none intron splicing efficiency / intron retention and gene expression by intron class
Bioinformatic genome annotation / intron classification human reference genome none counts, genomic occupancy, inter-intron distance, and chromosomal density of six intron classes
Bootstrapping, permutation, and pollution analyses K562 intron sets (in silico) none whether compartment/subcompartment enrichment is tethered to intron identity vs gene identity/number
Functional/GO enrichment and essentialome intersection analysis intron-class-containing genes (in silico) none essential genes and enriched GO terms (e.g., cell cycle) per intron class
Key results
  • 71% of minor introns enriched in compartment A vs 62% of major introns (strongest enrichment of any class) 71% vs 62%; p=4.7e-08
  • Inter-intron distance increases as intron abundance decreases, a nearly linear relationship R^2=0.98
  • Pollution analysis shows progressive increase in compartment A enrichment with rising minor intron composition (10–100%) r=0.89, p=2.2e-308
  • Minor introns reciprocally reduced in compartment B relative to major introns p=3.2e-10
  • Hybrid (51%) and non-canonical (50%) introns underrepresented in compartment A, instead enriched in 'neither' category (hybrid 17%, non-canonical 22%) hybrid p=1.8e-05; non-canonical p=2.2e-16
  • Major-like introns enriched in compartment A (64%) and in A1 subcompartment (77% vs 73% major) 64%, p=4.2e-16; A1 77%, p=2.2e-16
  • Within SPADs, major and minor introns efficiently spliced while other rare classes show reduced splicing efficiency, consistent across 4 cell lines
  • Bootstrapped median compartment A enrichment for 1,000 sets of 850 major introns remained 62%, identical to full major set 62%
Key statistics
  • count Major introns 345,317 (93.99% of intronic space); major-like 29,337 (5.03%); non-canonical 6070 (0.74%); minor 850 (0.11%); minor-like 458 (0.06%); hybrid 373 (0.07%) (intron class abundance and genomic occupancy)
  • correlation R^2=0.98 (inter-intron distance vs intron number per class)
  • correlation r=0.89, p=2.2e-308 (pollution analysis of minor intron composition vs compartment A enrichment)
  • pvalue p=4.7e-08 (minor introns (71%) enriched in compartment A vs major)
  • pvalue p=3.2e-10 (minor intron reduction in compartment B vs major)
  • pvalue p=2.2e-16 (non-canonical introns underrepresented in compartment A / Fisher exact tests)
  • other Exons 3.83% of gene-body sequence; introns 96.17% (genomic occupancy of exons vs introns)
  • other Median inter-intron distance: major 141 bp, major-like 27 kb, non-canonical 105 kb, minor 1.8e06 bp, minor-like 4.0e06 bp, hybrid 6.1e06 bp (inter-intron distance per class)

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 mapped six human intron classes (major, major-like, hybrid, minor-like, minor, non-canonical) across 3D genomic features (compartments A/B, subcompartments A1/A2/B1/B2/B3, SPADs, LADs, TADs) in K562 and three additional human cell lines using publicly available Hi-C, TSA-seq, DamID-seq, and RNA-seq data. Enrichment of each rare intron class within compartments and subcompartments was quantified as proportions and tested against major introns as a background reference using Fisher exact tests, with exact p-values reported throughout. To separate identity-driven spatial bias from gene-composition or sampling artifacts, bootstrapping (1,000 replicates sized to match minor intron count) and progressive 'pollution' permutation analyses were employed. Splicing efficiency was integrated with TSA-seq proximity data across cell lines; the provided text excerpt ends before the full statistical treatment of those comparisons is described.

Replicationunclear Sample sizeIntron counts per class stated explicitly (e.g., 850 minor, 345,317 major); four human cell lines used (K562, HFFc6, H1, HCT116); no formal power calculation or biological replicate count described in the provided text GroupsSix intron classes (major, major-like, hybrid, minor-like, minor, non-canonical) compared across nuclear compartments, subcompartments, and nuclear body-associated domains; four cell lines compared for splicing efficiency Pairingna Randomization/blindingna DispersionIQR Exact p-valuesyes Effect sizesno Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Fisher exact test (two-sided, pairwise) Pairwise comparison of compartment (A, B, neither) and subcompartment (A1, A2, B1, B2, B3) distributions between major introns and each rare intron class; also major vs minor introns stratified by gene essentiality Per-class intron counts: major 345,317; major-like 29,337; non-canonical 6,070; minor 850; minor-like 458; hybrid 373 not stated
Bootstrap resampling (1,000 replicates) Random draws of 850 major introns to generate a null distribution of compartment A enrichment proportions; used to assess whether minor intron enrichment is an artifact of sampling a small subset of the major class 1,000 replicates, each of n=850 (matching minor intron count) not stated
Permutation / 'pollution' analysis Progressive substitution of 10–100% minor introns into the 1,000 bootstrapped major intron lists to test identity-dependence of compartment A enrichment; also applied to subcompartment distributions for other rare classes 1,000 lists × 11 substitution levels (0–100% in 10% steps) not stated
Pearson correlation / linear regression Scatterplot of inter-intron distance vs intron class abundance (6 class-level data points; R²=0.98, Figure 1D); correlation between minor intron pollution proportion and compartment A enrichment across bootstrap replicates (r=0.89, p=2.2e-308) n=6 intron classes for R²; n=11 pollution levels × 1,000 replicates for r not stated
Gene Ontology (GO) functional enrichment analysis Rare intron-containing essential genes (hybrid, minor-like, minor, non-canonical) intersected with the essentialome; overlapping GO terms across classes yielded 1,314 terms, including 'cell cycle'; 146 cell-cycle genes per class intersected null not stated
Approaches that could also have been used
  • Many Fisher exact tests were conducted across intron classes and compartment/subcompartment categories without a stated multiplicity correction
    Could also: Apply a Benjamini-Hochberg false discovery rate correction (or Bonferroni) across the full family of pairwise tests — With five rare intron classes tested against multiple compartment and subcompartment categories—plus essentiality-stratified analyses—controlling the FDR is standard in genomics and would characterize which enrichments are robust to the multiple-comparison structure; reporting adjusted q-values alongside raw p-values is widely accepted practice
  • Compartment enrichment was reported as proportions; no effect-size measure accompanied the Fisher exact p-values
    Could also: Report odds ratios with 95% confidence intervals derived from the same 2×2 Fisher tables — Odds ratios quantify enrichment magnitude and precision alongside significance, which is especially useful when cell counts differ by orders of magnitude (850 minor vs 345,317 major introns), making the practical size of any enrichment easier to interpret and compare across classes
  • Distributions across more than two compartment categories (A, B, neither; or A1, A2, B1, B2, B3) were decomposed into separate Fisher exact tests per category
    Could also: Precede pairwise tests with a single chi-square or multinomial goodness-of-fit test across all categories simultaneously — An omnibus test first establishes that the overall distribution differs from the major intron reference, after which pairwise post-hoc decompositions are better motivated; this is a common approach when comparing count distributions across more than two categories
  • The inter-intron distance versus abundance relationship was summarized with R² from a linear regression on six class-level summary statistics (medians)
    Could also: Report Spearman's rank correlation with a 95% bootstrap confidence interval, or explicitly note the degrees of freedom (n=6 classes, df=4) — With only six data points the uncertainty around R² is substantial and unstated; Spearman correlation is also robust to the highly skewed scale of inter-intron distances across classes, and either approach with an accompanying CI conveys the imprecision of the fit
  • Bootstrapping demonstrated that minor intron compartment A enrichment is not attributable to random sampling of 850 major introns, with results shown as IQR boxplots
    Could also: Complement with a label-permutation test that shuffles intron-class assignments genome-wide and reports an empirical null distribution and formal p-value — Label permutation directly tests the null that intron class has no effect on compartment assignment under the actual genomic coordinate structure, providing a p-value grounded in the observed data geometry rather than parametric assumptions, and is widely used in genomic enrichment analyses
  • Splicing efficiency within and outside SPADs was compared across four cell lines using HFFc6 as a non-cancerous reference, apparently with per-comparison tests
    Could also: A linear mixed-effects model with intron class, SPAD proximity, and cell line as fixed effects—and an intron-class × cell-line interaction term—could jointly model all comparisons — A unified model would formally test whether the cancer-specific SPAD-independence of minor intron splicing is statistically significant as an interaction effect, control for intron-level covariates, and reduce the number of separate tests relative to pairwise per-cell-line comparisons
Software: not stated in provided text

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
0
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.

GSE63525 GEO in Methods (http://purl.org/orb/Methods)
also used by 1 paper:
GSM2687371 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSM2687379 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSM958729 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSM958733 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSM958749 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet

Downstream reach in the literature

100 downstream papers · 1 datasets

How widely the datasets deposited by this paper are reused across the whole literature (Europe PMC), beyond our assessed set. This is a factual dependency map — reusing a public dataset is normal, good science. It is not a judgement on the downstream papers; the only verdict here is this paper's own, with its cited rationale.

What was reproduced

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

Scope — pmid-41561379

Paper: Springer, Fleck et al., "Identity rather than 3D position informs splicing of rare introns in the human genome", iScience (2026). DOI 10.1016/j.isci.2025.114532. Code: https://github.com/saren-springer/introns-3Dgenome (HEAD 756b1e7). Data field (brief): GSE63525 (Rao et al. 2014 in-situ Hi-C, GM12878) — used only for the 3D-genome compartment/subcompartment annotations in the bootstrapping/pollution arm. The core intron analyses use intron annotations that ship inside the repo (GRCh38 Ensembl v99 intron BED/CSV files).

Pipeline-derived results (candidate in-scope)

# Result (paper) Pipeline Input (shipped?) Determinism Decision
1 Intron counts per class (Fig 1, Results): major 345317, major-like 29337, non-canonical 6070, minor 850, minor-like 458, hybrid 373 line-count / class tally of repo BED+CSV yes (repo) deterministic IN SCOPE (done)
2 Median inter-intron distance per class (Fig 1C) repo inter_intron_distance.py (pandas) yes (repo BEDs) deterministic IN SCOPE
3 R² distance vs abundance (Fig 1D, "R²=0.98") derived from #2 yes deterministic IN SCOPE
4 Intron density per 250 kb window (Fig 1, intron_density.R) repo R script yes (all_human_introns.csv) deterministic in scope (optional, lower-value: no single pinnable number)
5 chromoMap visualization (Fig 1E) repo chromoMap.R yes deterministic but a figure, no scalar out (visual only)
6 Bootstrapping/pollution compartment proportions, e.g. minor 71% in compartment A, p=4.7e-08 (Fig 2B–D, Supp 4/5) repo bash shuf + bedtools intersect with GSE63525-derived A/B/sub-compartment master table partial: repo ships only an example_overlap_table_forPlotting.txt; the real 3D master table (from GSE63525 Hi-C → compartments) is NOT shipped and must be rebuilt random (1000× shuf) + heavy upstream HARD 20% — not fully attempted (see AUDIT)

Out of scope

  • Wet-lab / cell-culture / RT-PCR splicing assays (cancer-cell minor-intron splicing) — experimental, not a pipeline.
  • Deriving A/B compartments & subcompartments de novo from GSE63525 Hi-C — large upstream pipeline; the repo presupposes the annotated master table. Marks #6 as the optional tail.

Note (P16)

Repo is the authors' own code, mostly thin wrappers over shipped data + standard libs (pandas/dplyr/chromoMap/bedtools). Fully eligible: resolvable code + shipped public data + pinnable reported values (#1–#3).

Figures / tables: Fig 1Fig 1CFig 1D
C1_count_major
Reported
345317
Reproduced
345317
exact
C2_count_majorlike
Reported
29337
Reproduced
29337
exact
C3_count_noncanonical
Reported
6070
Reproduced
6070
exact
C4_count_minor
Reported
850
Reproduced
850
exact
C5_count_minorlike
Reported
458
Reproduced
458
exact
C6_count_hybrid
Reported
373
Reproduced
373
exact
C7_dist_major
Reported
141 bp
Reproduced
141.0 bp
exact
C8_dist_majorlike
Reported
~27 kb
Reproduced
27822 bp
within tolerance
C9_dist_noncanonical
Reported
~105 kb
Reproduced
105226 bp
exact
C10_dist_minor
Reported
~1.8e6 bp
Reproduced
1835364 bp
exact
C11_dist_minorlike
Reported
~4.0e6 bp
Reproduced
4036939.5 bp
exact
C12_dist_hybrid
Reported
~6.1e6 bp
Reproduced
5912001 bp
within tolerance
C13_r2_dist_abundance
Reported
0.98
Reproduced
0.982 (log-log)
exact

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 98/100

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: Q7 · Core claim 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score 0

All 13 attempted Fig 1 claims reproduce essentially perfectly from repo-shipped GRCh38/Ens99 annotations: six intron-class counts and four inter-intron distances are exact, C8/C12 are within rounding, and the Fig 1D log-log R²=0.982 matches the reported 0.98 (the agent correctly identified the log transform, non-log R² being only 0.19). No fabrication signal — every graded value is derivable from shipped data+code. The only limitation is scope and lies on the data-availability side (not the authors' fault): the paper's actual central thesis ('identity rather than 3D position', Fig 2) rests on a GSE63525-derived compartment master table that is not deposited, so it could not be reproduced. Hence a strong q8=green for what was tested, with q7=yellow because the title-level conclusion remains unconfirmed.

🤝
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.

76.6 k
tokens (I/O) · 4.7 M incl. cache
8 min
runtime · 0 CPU-h
0.3 GB
peak RAM
1
HPC jobs
hummel
machine