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

Integrative transcriptome sequencing identifies trans-splicing events with important roles in human embryonic stem cell pluripotency.

Genome Res · 2013
L1 100/100 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: 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: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +5
✓ What held up
  • Same input data as the authors
  • No relevant deviation in data/preprocessing
  • No authors-side cause for any deviation
  • Any deviation was negligible
What did not (or only partly)
  • 🟡Reported values were only indirectly comparable
  • 🟡Reported values were not (fully) derivable from the shared data
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
100/100
Reproducibility score
1.5 SD above mean
vs. all fields · 1173 studies
🎯 Scores higher than 95% of all assessed papers rank 1 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.

💻 Code ↗ 🗄 Data: GSE20301

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.

✎ 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
2026-07-30
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-07-31
no human curator yet
Last updated
2026-07-31

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

Does genuine (non-artifactual) trans-splicing occur in human embryonic stem cells, and if so, do the resulting trans-spliced RNAs have biologically significant roles in pluripotency? The authors ask whether an integrative multi-platform transcriptome-sequencing approach can distinguish real trans-splicing from experimental artifacts and genetic rearrangements in hESCs.

Core claims
  • TSscan, a computational pipeline integrating long- and short-read transcriptome sequencing from multiple hESC lines, can detect trans-splicing while minimizing false positives from experimental artifacts and genetic rearrangements. method
  • Most chimeric RNA products detected in RNA-seq are platform-dependent experimental artifacts; ~99.9% of 454-nominated chimeric candidates were discarded by TSscan. finding
  • Four trans-spliced RNAs (tsCSNK1G3, tsARHGAP5, tsFAT1, tsRMST) were identified and experimentally confirmed in hESCs, while a fifth candidate (tsSOBP) was shown to be an MMLV-RTase-dependent artifact. finding
  • tsRMST is the first reported trans-spliced large intergenic noncoding RNA (lincRNA). resource
  • The four trans-spliced RNAs are highly expressed in human pluripotent stem cells (hESCs and iPSCs) and are differentially expressed during hESC in vitro differentiation. finding
  • tsRMST contributes to pluripotency maintenance of hESCs by suppressing lineage-specific gene expression. finding
  • tsRMST acts in trans, not in cis: it is nuclear-enriched, does not regulate neighboring genes within 1 MB, and interacts with the pluripotency transcription factor NANOG and the PRC2 complex factor SUZ12. mechanism
  • Comparing RT-PCR products generated with different reverse transcriptases (MMLV vs. AMV) is necessary to confirm trans-splicing, because template-switching artifacts are RTase-dependent. method
Experimental setups
Assay System Perturbation Readout Platform
Long-read whole-transcriptome RNA-seq (oligo-dT selected) H9 hESC line (generated in this study); H1 hESC 454 data downloaded from public database none Chimeric (noncolinear) RNA candidates and chimeric junction sites via alignment to human reference genome Roche 454
Short-read whole-transcriptome RNA-seq H9 hESC line (SOLiD, generated); H1 hESC (Illumina, public data) none Short-read support for long-read-nominated chimeric junction sites SOLiD and Illumina
RT-PCR with two different reverse transcriptases (MMLV-derived and AMV-derived RTase), plus RT-free control and elevated primer annealing temperature; amplicon Sanger sequencing hESC lines H1, H9, and NTU1 none (RTase type varied as artifact control) Presence/absence of trans-spliced amplicons; RTase dependence; chimeric junction site sequence identity
RNase protection assay (RPA), non-RTase-based validation Total RNA of hESC H9 none Probe protection vs. degradation for the five candidate trans-spliced RNAs
RT-PCR and qRT-PCR expression profiling hESC lines (H1, H9, NTU1); human iPSC clones iCFB50 (foreskin fibroblast), iGRA2 (granulosa cells), iCD3 (dermal papilla cells) with parental somatic lines; ten human normal tissues; hESCs at differentiation day 14 and day 21 in vitro differentiation; somatic cell reprogramming Relative expression of trans-spliced isoforms vs. corresponding colinear isoforms
shRNA knockdown (shTS2 targeting the tsRMST chimeric junction site; shLuc control; shTS2-rescue by tsRMST re-expression) with microarray-based global gene expression profiling, qRT-PCR, and alkaline phosphatase staining hESCs (4 and 7 d post-viral transduction) shRNA knockdown (KD) and rescue/overexpression Pluripotency gene expression (NANOG, POU5F1, SOX2, TCF7L1), lineage-specific gene expression (T, MIXL1, GSC, GATA4, GATA6, SOX7, SOX17, PAX6, SOX1), alkaline phosphatase staining intensity microarray (platform not stated)
Fluorescence-activated cell sorting (FACS) and immunocytochemistry (ICC) hESCs transduced with shLuc, shTS2, or shTS2-rescue constructs (day 4 and day 7) shRNA knockdown and rescue Percentage/number of NANOG+, POU5F1+, T+, SOX17+, and PAX6+ cells
Nuclear/cytoplasmic fractionation qRT-PCR and RNA immunoprecipitation (RIP) hESCs none Subcellular localization of tsRMST; association of tsRMST with NANOG and SUZ12 proteins
Key results
  • TSscan reduced 8822 long-read-nominated chimeric RNA candidates to nine candidates supported by RNA-seq from both H1 and H9 hESCs ~99.9% of 454-nominated candidates discarded (8822 → 9)
  • Five of nine candidates (tsCSNK1G3, tsARHGAP5, tsFAT1, tsRMST, tsSOBP) were detected by MMLV-RTase RT-PCR in H1, H9, and NTU1; tsSOBP was MMLV-RTase-dependent and absent with AMV-RTase, and its RPA probes were degraded, identifying it as an artifact
  • tsRMST is classified as noncoding by the coding potential calculator, making it the first trans-spliced lincRNA identified with multiple experimental validations coding potential calculator score < 53
  • All four trans-spliced RNAs were expressed in every tested human iPSC clone (derived from skin fibroblast, dermal papilla, and granulosa cells)
  • Upon in vitro hESC differentiation, tsCSNK1G3, tsARHGAP5, and tsFAT1 increased while tsRMST significantly decreased; tsRMST was expressed far above RMST in pluripotent stem cells but was undetectable in ten normal tissues in which RMST was broadly expressed
  • tsRMST knockdown (shTS2) reduced alkaline phosphatase staining and significantly decreased pluripotency genes NANOG, POU5F1, SOX2, TCF7L1, while increasing mesodermal (T, MIXL1, GSC), endodermal (GATA4, GATA6, SOX7, SOX17), and neuroectodermal (PAX6, SOX1) genes
  • FACS and ICC showed decreased NANOG+ and POU5F1+ cells after shTS2, increased T+ and SOX17+ cells by day 4 and increased PAX6+ cells by day 7; re-expression of tsRMST (shTS2-rescue) significantly restored NANOG+/POU5F1+ cell numbers, pluripotency gene expression, hESC morphology, and alkaline phosphatase staining
  • tsRMST transcripts were highly enriched in hESC nuclei, and knockdown did not alter expression of neighboring genes/microRNAs within 1 MB (NEDD1, MIR1251, MIR135A2), indicating a trans- rather than cis-acting mechanism; RIP was used to test interaction with NANOG and SUZ12 no change for genes within 1-MB range
Key statistics
  • count 0.83 million long reads (averaging 353.7 bp) (Roche 454 whole-transcriptome sequencing of H9 hESCs generated in this study)
  • count 230.63 million short reads (50 bp) (SOLiD whole-transcriptome sequencing of H9 hESCs)
  • count 8822 (Chimeric RNA candidates extracted from long 454 reads of hESC H1/H9 (TSscan step 1))
  • count nine (trans-spliced RNA candidates remaining after all four TSscan filtering steps)
  • other ~99.9% (Proportion of 454-nominated chimeric candidates discarded by TSscan screening)
  • count four (Genuine trans-spliced transcripts confirmed by multiple experimental validation steps)
  • other score < 53 (Coding potential calculator score establishing tsRMST as noncoding)
  • pvalue P < 0.05 (*), P < 0.01 (**), P < 0.001 (***) (Significance thresholds for all comparisons, estimated by two-sample two-tailed t-test (Figs. 2, 3))

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 reports comparisons of gene/transcript expression levels (e.g., trans-spliced vs. colinear isoforms, differentiation time points, knockdown vs. control conditions) using two-sample t-tests, with results summarized as significance thresholds (P < 0.05/0.01/0.001) rather than exact values. A microarray-based heat map was also used to visualize relative fold-change patterns across pluripotency- and lineage-associated genes in a knockdown experiment.

Replicationunclear Sample sizeOnly partially: 'Three independent transfections were performed to determine the mean' is stated for the FACS analyses in Fig. 3D and 3F; sample sizes for other qRT-PCR/RT-PCR comparisons are not stated Groupsundifferentiated vs. differentiated hESCs; trans-spliced vs. colinear transcript isoforms; shLuc control vs. shTS2 knockdown vs. shTS2-rescue hESCs; multiple tissue/cell types Pairingunclear Randomization/blindingnot stated Dispersionunclear Exact p-valuesno Multiplicity correctionno
Statistical tests used
Test Applied to n Assumptions
two-sample, two-tailed Student's t-test Figure 2 (expression comparisons across differentiation status, tissues, and isoform types) not stated
two-sample, two-tailed Student's t-test Figure 3 (qRT-PCR comparisons of pluripotency/lineage marker expression and FACS/ICC quantification in shTS2 knockdown, control, and rescue hESCs) three independent transfections stated for FACS quantifications (Fig. 3D,F); n not stated for other panels not stated
Approaches that could also have been used
  • Many individual two-sample t-tests were applied across a large family of genes/markers within the same figures (e.g., pluripotency and lineage markers in Fig. 3C, H) without a stated multiple-testing correction.
    Could also: A repeated-measures or one-way ANOVA across conditions per gene, followed by a post-hoc correction (e.g., Tukey HSD), or a false-discovery-rate procedure (e.g., Benjamini-Hochberg) applied across the full set of gene comparisons — These approaches would jointly control the family-wise error rate or false discovery rate across the many simultaneous comparisons, which can be a useful complement when testing many markers side by side.
  • Statistical significance is reported using threshold symbols (*, **, ***) rather than exact P-values.
    Could also: Reporting the exact P-value for each comparison alongside or instead of threshold symbols — Exact P-values let readers gauge the precise strength of evidence and enable meta-analytic or re-analysis use of the reported statistics.
  • Some comparisons (e.g., FACS quantification in Fig. 3D,F) were based on a small number of replicates (three independent transfections) analyzed with a parametric t-test.
    Could also: A non-parametric test such as the Mann-Whitney U test, or reporting exact/permutation-based P-values — With small sample sizes, non-parametric or exact/permutation methods do not rely on the normality assumption underlying the t-test and can be a useful cross-check.
  • Microarray-based global gene expression profiling (Fig. 3B) was summarized as a heat map of relative fold changes without a stated statistical test for differential expression.
    Could also: A moderated t-statistic approach designed for small-sample microarray/RNA-seq data (e.g., limma or a similar empirical Bayes method), potentially combined with FDR control — Such methods borrow information across genes to stabilize variance estimates in small-sample expression profiling and can also assign formal significance/FDR values to the fold-change patterns shown.
  • Sample size or replicate type (biological vs. technical) is stated explicitly for only some experiments (e.g., 'three independent transfections' for Fig. 3D,F) and not for others.
    Could also: Consistently reporting the number and type (biological/technical) of replicates for every quantitative comparison — Explicit, consistent replicate reporting alongside the existing t-tests would let readers evaluate the precision and generalizability of each comparison.

What was reproduced

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

C1_h9_read_count
Reported
~0.83 million reads
Reproduced
832,438 reads (435,350 from SRR315139 + 397,088 from SRR315140, independently downloaded from SRA/ENA and converted from FASTQ)
exact
C2_psl_provenance
Reported
n/a - this is a reproducibility/provenance check we devised, not a number quoted from the paper text
Reproduced
100% of the PSL's 10,135 unique query reads were traced to source files by exact read-ID matching: 2,981 from GSM517436 (H1), 3,729 from SRR315139 (H9 run1), 3,425 from SRR315140 (H9 run2), 0 from GSM517440 (H1 N2 control, correctly excluded as a differentiation-stage sample). Sum = 10,135, an exact match to the PSL's own unique-ID count.
exact
C3_tsscan_step1_candidate_count
Reported
8,822 candidates
Reproduced
nicht durchgefuehrt (kein Grund vermerkt)
m.public.grade.not-attempted
C4_step2to4_pipeline_and_h1h9_concordance
Reported
9 final candidates after all in-silico filters, each supported independently in both H1 and H9
Reproduced
nicht durchgefuehrt (kein Grund vermerkt)
m.public.grade.not-attempted
C5_experimentally_confirmed_trans_spliced_rnas
Reported
4 of 9 candidates experimentally confirmed
Reproduced
nicht durchgefuehrt — This claim rests on wet-lab experimental validation (RT-PCR, sequencing across candidate junctions, functional assays), not purely computational pipeline output. Out of scope for a computational reproduction; correctly excluded from grading against pipeline output.
m.public.grade.out-of-scope

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 100/100

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: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +5

What deviates: essentially nothing that was measured — the single directly comparable value reproduced exactly (832,438 H9 454 reads vs the paper's ~0.83 million), and the authors' deposited intermediate PSL passed a strict provenance audit with all 10,135 unique query IDs traced to raw archives (0 leaking in from the N2 differentiation control, exactly as it should be). What is missing: the two claims that actually carry the paper's computational argument, 8,822 step-1 candidates and 9 final candidates after the BFAST/filtering/H1-H9-concordance steps, were never executed, so the core claim is neither supported nor contradicted here. Whose side: the shortfall is ours (scope/compute time, Illumina files not fetched, VPN outage), while a real defect sits on the authors' side in their released code — TSscan1of4 segfaults on standard psLayout-header PSL, including their own deposited GSE30557_F5AGPVJ.psl, due to an unguarded first record hitting A[18]; that is a genuine reproducibility hazard even though it does not impugn the published numbers. Severity: low on observed error, moderate on coverage — an honest, well-documented partial reproduction with explicit drops rather than fabricated fills, which is why q5/q7/q8 sit at yellow rather than green or red.

🤝
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.