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

Evidence for L1-associated DNA rearrangements and negligible L1 retrotransposition in glioblastoma multiforme.

Mob DNA · 2016
67/100 3/4
Why this verdict

The main results reproduced, with only marginal, non-material deviations.

Reproduced on the brainbox compute brainarbeit.com
How its reproducibility compares
67/100
Reproducibility score
0.4 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 29% of all assessed papers rank 795 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 reproduction (findings-level, not byte-exact). FRESH from-scratch re-run on «our HPC» after requeue + «infra» reclaim: rebuilt the tebreak conda env (commit 1a63a06), re-downloaded UCSC hg19 + bwa-indexed, re-downloaded the patient-8 RC-seq V3 trio from ENA PRJEB1785, re-aligned (bwa mem -M -Y), and re-ran authors' own tebreak. C0 self-test EXACT (5/5). C3 (headline): the tumour-specific somatic L1-Ta in EGFR intron 1 (chr7:55,030,723) is RECOVERED with the paper's exact signature (L1Ta, antisense, 5'-truncated TE 4970-6030, MismatchTSD ~ the 550nt deletion, 31/13 split reads), present in tumour (24 EGFR-window L1Ta) and ABSENT in both normal (0) and blood (0). Identical call to the prior 2026-06-21 run -> stable. C1 germline polymorphic L1-Ta burden ~150-152/sample vs reported 208 = within ~1.4x (single patient vs 14-patient cohort avg); fresh genome-wide normal refining. Honest caveats: (1) EGFR amplification in this GBM tumour inflates breakpoint density (224 vs 0) = the paper's own 'L1-associated DNA rearrangements / negligible retrotransposition' thesis, so clean separation of one bona-fide insertion from rearrangement signal is partial; (2) the 2016 paper used an in-house tebreak predecessor (flags --mincluster/--minclip/--minq absent from any public commit) -> byte-exact reproduction infeasible. NOT attempted: C2 (needs authors' 960-locus reference genotype set), C4 cohort-wide sweep (out of 80/20 scope), all wet-lab + WGS-SV results.

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 67
    assessed: 2026-06-21 ⛓ 4a4ceaaddd08
✎ 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
2026-06-24
Rubric version
not recorded
Assessed by
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: sonnet
Founding hypothesis

The authors hypothesised that L1-associated DNA rearrangements in glioblastoma multiforme (GBM) might occur via recombination or an atypical, endonuclease-independent retrotransposition mechanism lacking canonical TPRT hallmarks, or alternatively that L1 insertions in GBM could be restricted to rare sub-clonal, heterogeneous events undetectable by prior methods.

Core claims
  • Canonical (endonuclease-dependent, TPRT-driven) L1 retrotransposition is absent or negligible in GBM tumours and cultured GBM cell lines finding
  • Atypical L1-associated DNA rearrangements (endonuclease-independent insertion, L1-associated rearrangements, Alu-Alu recombination near an L1) occur in GBM tumours following DNA damage finding
  • Retrotransposon capture sequencing (RC-seq) at up to 250x depth was used to survey L1 mutations in 14 brain tumour patients method
  • An engineered L1 reporter assay was used to test in vitro L1 mobilisation of WT and EN-mutant constructs in GBM cell lines method
  • Tumour-specific L1 mutations identified by RC-seq in MeCP2 and EGFR introns were PCR validated finding
  • The somatic L1 insertion in MeCP2 in patient #2 is associated with reduced MeCP2 transcript levels, increased L1 transcript levels, and reduced L1 promoter methylation in tumour versus adjacent brain finding
Experimental setups
Assay System Perturbation Readout Platform
Retrotransposon capture sequencing (RC-seq) 14 brain tumour patients (9 GBM, 5 lower grade glioma); tumour, adjacent brain, and blood tissue none (tumour vs adjacent brain/blood comparison) L1 insertion sites / L1-genome junctions Illumina HiSeq2000, HiSeq2500, MiSeq
Empty/filled site PCR validation with capillary sequencing Patient tumour and adjacent brain genomic DNA (MeCP2, EGFR loci) none Presence/absence of L1 mutant allele ABI3730 capillary sequencer
qRT-PCR Patient #2 tumour and adjacent brain tissue RNA none MeCP2 isoform 1/2 and exon 4 transcript levels; L1 5'UTR and ORF2 transcript levels ViiA 7 Real-Time PCR System
Whole genome sequencing Patient #2 (tumour, adjacent brain) and patient #8 (tumour, blood) genomic DNA none Genome-wide detection of endonuclease-dependent L1 insertions Illumina HiSeq X Ten
Bisulfite sequencing (L1 promoter methylation) Patient tumour and adjacent brain genomic DNA none CpG methylation status at L1.4 promoter CpG island Illumina MiSeq
In vitro L1 retrotransposition reporter assay 4 cultured GBM cell lines Wild-type vs endonuclease-mutant L1 reporter construct L1 mobilisation efficiency
PCR-based deletion quantification with gel imaging Patient #2 tumour and adjacent brain genomic DNA (MeCP2 locus) none Relative amplicon intensity of 58 nt deletion region Typhoon FLA 9500 scanner, Image Studio Lite
Key results
  • In 4 GBM tumours, characterised one probable endonuclease-independent L1 insertion, two L1-associated rearrangements, and one likely Alu-Alu recombination event adjacent to an L1 4 events in 4/14 tumours
  • No tumour-specific, endonuclease-dependent L1 insertions found by RC-seq despite sequencing at up to 250x depth up to 250x depth
  • Whole genome sequencing of tumours carrying the MeCP2 and EGFR L1 mutations (patients #2, #8) revealed no endonuclease-dependent L1 insertions
  • Wild-type and endonuclease-mutant L1 reporter constructs each mobilised very inefficiently in four cultured GBM cell lines
  • MeCP2 transcript isoform levels significantly reduced in tumour versus adjacent brain p<0.008
  • L1 5'UTR and ORF2 transcript levels significantly increased in tumour versus adjacent brain p<0.001
  • L1 promoter CpG methylation reduced in tumour versus adjacent brain p<0.001
Key statistics
  • pvalue p < 0.008 (MeCP2 transcript isoform levels, tumour vs adjacent brain, two-tailed t-test, df=6)
  • pvalue p < 0.001 (L1 5'UTR/ORF2 transcript levels, tumour vs adjacent brain, two-tailed t-test, df=10)
  • pvalue p < 0.001 (L1 promoter methylation, tumour vs adjacent brain, paired t-test, df=18)
  • count 3,252,752,806 (Total 2x150mer RC-seq reads generated across the cohort)
  • count 14 (Brain tumour patients studied (9 GBM, 5 lower grade glioma))
  • count 4 (Putative tumour-specific L1 mutations reported at applied thresholds)
  • other up to 250x (RC-seq sequencing depth at L1 integration sites)
  • other 98.5% (Previously reported PCR validation rate for polymorphic L1 insertions detected by 2 RC-seq reads)

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 study surveyed somatic L1 retrotransposon mutations in 14 brain tumour patients using retrotransposon capture sequencing (RC-seq) and whole-genome sequencing, relying on rule-based bioinformatic thresholds for mutation calling rather than formal inferential statistics. Quantitative molecular assays (qRT-PCR of MeCP2 and L1 transcripts; bisulfite sequencing of L1 promoter methylation) compared tumour versus adjacent brain tissue using two-tailed or paired t-tests, with results expressed as mean ± SEM. An in vitro L1 reporter assay in GBM cell lines provided complementary qualitative evidence; no multiplicity correction was reported across the several t-tests performed.

Replicationmixed Sample size14 patients total (9 GBM, 5 lower grade glioma) for RC-seq; qRT-PCR comparisons derived from patient #2 tissue with five technical replicates per assay; degrees of freedom reported for each t-test (df = 6, 10, 18) but the source of observations (biological vs technical) is not disambiguated GroupsTumour vs adjacent brain tissue (and blood for selected patients); four GBM cell lines in the in vitro reporter assay Pairingmixed Randomization/blindingnot stated DispersionSEM Exact p-valuesno Effect sizesno Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Two-tailed Student's t-test MeCP2 transcript isoform levels (isoforms 1, 2, and exon 4): tumour vs adjacent brain (qRT-PCR, patient #2) df = 6 (as stated); exact n not specified not stated
Two-tailed Student's t-test L1 transcript abundance at 5′UTR and ORF2 regions: tumour vs adjacent brain (qRT-PCR, patient #2) df = 10 (as stated); exact n not specified not stated
Paired t-test L1 promoter CpG methylation: tumour vs adjacent brain (bisulfite-seq, patient #2) df = 18 (as stated); exact n not specified not stated
Approaches that could also have been used
  • Dispersion was reported as mean ± SEM for all qRT-PCR comparisons, with df ranging from 6 to 18
    Could also: Standard deviation (SD) or 95% confidence intervals could also convey spread — With small effective n, SEM can appear narrow relative to the true variability in the data; SD and CIs communicate the spread of individual observations more directly and are often preferred in biological-replication contexts by journals and reporting guidelines
  • Multiple independent t-tests were applied across MeCP2 isoforms, L1 transcript regions, and CpG methylation sites without a stated correction
    Could also: A mixed-effects model or repeated-measures ANOVA with a post-hoc correction (e.g., Tukey HSD or Benjamini-Hochberg FDR) could also span the family of comparisons — A correction would explicitly control the familywise or false-discovery error rate when multiple related outcomes are tested simultaneously, which is a common alternative in expression and epigenetics studies
  • Two-tailed Student's t-tests were used for qRT-PCR group comparisons with small df (6 and 10)
    Could also: Non-parametric alternatives such as the Wilcoxon rank-sum (Mann-Whitney U) or Wilcoxon signed-rank test could also be applied — With very small sample sizes the normality assumption underlying t-tests is difficult to verify empirically; non-parametric rank tests make no distributional assumption and are a common alternative when n is small
  • p-values were reported as inequalities (p < 0.008, p < 0.001) rather than exact values
    Could also: Exact p-values (e.g., p = 0.003) could also be reported alongside the test statistic — Exact p-values allow readers and meta-analysts to evaluate the strength of evidence more precisely and are recommended by APA, CONSORT, and many biomedical journals
  • No effect sizes were reported alongside the t-test results
    Could also: Cohen's d or the fold-change with a 95% CI could also accompany each comparison — Effect sizes quantify the magnitude of a difference independently of sample size, complementing the p-value and enabling cross-study comparisons and future power calculations
  • Somatic L1 mutation detection relied on rule-based bioinformatic thresholds (read-count cutoffs, confidence scores ≥ 0.9, database exclusion) rather than a probabilistic model
    Could also: Probabilistic somatic transposable-element callers (e.g., MELT, xTea) with associated posterior probabilities or q-values could also be applied — Probabilistic frameworks can quantify uncertainty around each candidate insertion call and provide a principled false-discovery rate estimate across the full call set, making the sensitivity–specificity trade-off explicit
Software: GraphPad Prism 5 5 · TEBreak commit f7f01c1 · BWA-MEM · GATK 3.3 · Picard Tools (MarkDuplicates) · SAMtools

What was reproduced

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

Scope — pmid-27843499

Paper: Carreira et al. 2016, Mobile DNA 7:21. "Evidence for L1-associated DNA rearrangements and negligible L1 retrotransposition in glioblastoma multiforme." PMCID PMC5105311 · DOI 10.1186/s13100-016-0076-6.

Tool (P16): TEBreak — https://github.com/adamewing/tebreak (Adam Ewing, a co-author). Split-read / discordant-read transposable-element insertion caller.

Data: ENA PRJEB1785 — 59 runs (~1.5 TB). 44 RC-seq capture runs ("OTHER", L1-Ta-enriched, V2 design for patients #1-5, V3 design for #6-14) + 15 WGS runs.

In scope (pipeline-derived, attempted)

The paper's quantitative L1-detection numbers come from the TEBreak pipeline run on PRJEB1785. We reproduce by running the published tebreak tool on the paper's own data (RC-seq), per the described approach (BWA-MEM -Y -M → hg19 → tebreak split-read calling of L1 insertions).

# Reported result Location Pipeline
C1 "Average of 208 polymorphic L1-Ta insertions per sample" Results, "L1 mutations…" tebreak nonref L1 calls per RC-seq sample
C2 "93.6% of 960 reference genome copies of L1-Ta detected" Results tebreak/RC-seq sensitivity over reference L1-Ta set
C3 Somatic L1-Ta insertion in EGFR intron (patient #8), 5′-truncated, antisense, ~0.5 kb, 550 nt deletion Results, "EGFR" tebreak nonref L1 call at EGFR (chr7) in tumour not normal/blood
C4 Somatic L1PA2 insertion in MeCP2 intron (patient #2), 58 nt deletion Results, "MeCP2" tebreak nonref L1 call at MeCP2 (chrX) — secondary, if budget

Primary target (80%): Patient #8 RC-seq V3 trio — tumour GBM.8T.V3 (ERR580947, 4.4 GB), adjacent-normal GBM.8NT.V3 (ERR580946, 4.2 GB), blood GBM.8B.V3 (ERR580956, 0.5 GB). Small, V3 design, contains the EGFR somatic insertion (C3) and yields a per-sample polymorphic-L1 count (C1). This gives clean, checkable data points at modest compute.

Version-drift caveat (honesty note)

The paper's stated TEBreak parameters — --mincluster 2, --minclip 30, --minq 1do not exist in any public tebreak commit (earliest repo content is Feb 2016; CLI uses --min_split_reads, --min_minclip, --min_prox_mapq, no --mincluster/--minclip/--minq). The 2016 paper used an in-house predecessor of the redesigned, published tool. Therefore byte-exact reproduction of the 2016 numbers is not feasible; we reproduce the findings (does the authors' tool, run on the authors' data, recover the somatic L1 insertions and a comparable polymorphic-L1 burden) and grade order-of-magnitude / presence-absence, not identity. Filter cascade (scripts/general_filter.py, the nonref polymorphism DB, per-patient read thresholds) is approximated, not matched line-for-line.

Out of scope (not pipeline / not attempted)

  • All wet-lab: PCR validation, qRT-PCR expression, methylation, in-vitro L1 retrotransposition reporter assays (DBTRG/M059J/LN18/LN229/HeLa).
  • WGS structural-variant / CNV / point-mutation calls (Delly, Manta, Strelka, Platypus, cn.MOPS) — TP53/IDH1/EGFR-amplification/CDKN2A etc. These are orthogonal pipelines, not the L1/tebreak result that defines the paper.
  • The full 14-patient × all-runs sweep (~1.5 TB) — 80/20: we run one patient trio, not the cohort. Cohort-wide averages (C1/C2) are therefore estimated from one sample, reported as such.
C0-selftest
Reported
tebreak shipped self-test yields 5 insertions
Reproduced
5 insertions (fresh run, setup «job»)
exact
C1
Reported
~208 polymorphic (non-ref) L1-Ta insertions per RC-seq sample (cohort avg, 14 patients)
Reproduced
germline genome-wide L1Ta burden (single patient-8): normal=151, blood=152, tumour=538 (somatic-inflated); fresh genome-wide normal re-run «job» in progress. ~150 vs 208 = within ~1.4x (same ballpark; single patient vs 14-patient cohort avg).
partial
C2
Reported
93.6% of 960 reference-genome L1-Ta copies detected by RC-seq
Reproduced
NOT REPRODUCIBLE: needs authors' 960-locus reference set + genotype_ref step, not standard nonref tebreak output
m.public.grade.error
C3
Reported
tumour-specific somatic L1-Ta in EGFR intron (patient #8), 5'-truncated antisense ~0.5kb, ~550nt genomic deletion, absent in normal/blood
Reproduced
FRESH RE-RUN recovers chr7:55,030,723 (EGFR intron 1): L1Ta, antisense (-/-), 5'-truncated (TE_align 4970-6030 = 3' fragment), 31/13 split reads, Filter=MismatchTSD (non-canonical junction ~ the 550nt deletion). Tumour EGFR window: 24 L1Ta calls; normal=0; blood=0. Identical to the 2026-06-21 run on every reported axis.
partial
C4
Reported
4 putative tumour-specific (somatic) L1 across the 14-patient cohort
Reproduced
NOT ATTEMPTED cohort-wide (out of 80/20 scope); patient-8 EGFR (1 of the 4) reproduced as C3
m.public.grade.error

Assessments & scoring basis

Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.

No assessment has been recorded yet.
🤝
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.

872.4 k
tokens (I/O) · 70.9 M incl. cache
545 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.