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

EDGE COVID-19: a web platform to generate submission-ready genomes from SARS-CoV-2 sequencing efforts.

Bioinformatics · 2022
L1 49/100 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 result did not reproduce in this reproduction attempt. Where our recomputation produced values that differ from the published ones, those discrepancies are listed below. This is a single automated attempt — not peer review and not a finding of error or misconduct — and differences can also arise from data access, undocumented parameters or the computing environment. The verdict can be contested via “report an error”.

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: 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 +6
✓ What held up
  • Reported values were directly comparable
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
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
49/100
Reproducibility score
1.4 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 7% of all assessed papers rank 1081 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

EC-19 is an applications-note web pipeline (FaQCs -> bwa/minimap2 vs NC_045512.2 -> variant call -> consensus -> Pangolin --usher). Its quantitative Results come NOT from the paper's own deposit PRJNA714680 (44 Illumina amplicon runs, 'dozens of genomes', no numeric claim) but from 9 OTHER reprocessed public bioprojects whose EC-19 output the authors deposited as LANL-data.txt. Tier-1 audit: re-derived the paper's aggregate numbers directly from that deposited table. STRONG reproductions: per-sample SNV rate 13.45 vs 13; linear coverage 86.5% vs 88%; min coverage 1.41% vs 1.4%. MISMATCHES flagged for human audit: deposited table has 386 samples vs 239 reported (superset); indel counts 2337 (6.05/sample) vs 139 (0.6/sample) ~10x higher because the deposit holds unfiltered ONT homopolymer insertions while the paper reports a filtered set whose filtering step is undocumented. Total SNVs (5192) scale to ~3115 once restricted to 239. NOT attempted: Table 1 EC-19-vs-Galaxy concordance (needs external Galaxy workflow). PENDING («our HPC» down centrally during this pass): Tier-2 = run EC-19 on individual accessions to validate single rows of LANL-data.txt, and on the 55 CDC benchmark SRRs to reproduce 100% lineage accuracy. Honest verdict: core per-sample claims reproduce; the deposited supplement does not 1:1 match the printed N and indel totals -> partial.

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 49
    assessed: 2026-06-19 ⛓ 1f018ead43d9
✎ 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
2026-06-25
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: sonnet
Founding hypothesis

Can an integrated, web-based bioinformatic platform (EDGE COVID-19) generate consistent, high-quality, submission-ready SARS-CoV-2 consensus genomes from both Illumina and Oxford Nanopore sequencing data, with variant and lineage calls that are accurate and concordant with established workflows like the COVID-19 Galaxy Project?

Core claims
  • EDGE COVID-19 (EC-19) is a web-based platform that automates QC, reference-based variant/consensus calling, lineage determination, and submission of SARS-CoV-2 genomes and metadata to GenBank, GISAID and INSDC for both Illumina and ONT data. resource
  • EC-19 correctly assigned lineages for all 55 CDC benchmark datasets tested. finding
  • EC-19 shows strong concordance with the Galaxy workflow on Illumina data, with 95% agreement for both SNVs and indels. finding
  • EC-19 and Galaxy show greater discordance on ONT data, especially for indels (55% agreement) versus SNVs (95% agreement). finding
  • Most discrepancies between EC-19 and Galaxy occur in low-coverage samples/regions or near amplicon primer binding sites. mechanism
  • EC-19 uses FaQCs for QC, Minimap2 (ONT) or BWA mem (Illumina) for mapping, and Pangolin with usher mode for lineage assignment. method
  • EC-19 was used to successfully process and submit real-world New Mexico surveillance genomes to GISAID and GenBank. finding
Experimental setups
Assay System Perturbation Readout Platform
Amplicon/shotgun sequencing (Illumina) with EC-19 workflow (QC, mapping, variant/consensus calling) SARS-CoV-2 clinical samples, New Mexico, USA none consensus genome, SNVs/indels, submission success to GISAID/GenBank Illumina
Amplicon sequencing (ONT) with EC-19 workflow SARS-CoV-2 clinical samples, New Mexico, USA none consensus genome, SNVs/indels, submission success to GISAID/GenBank Oxford Nanopore Technologies (ONT)
Lineage assignment benchmarking (Pangolin --usher mode) 55 SRR benchmark datasets with known lineages (CDC) none concordance of predicted vs known lineage Pangolin
Cross-workflow SNV/indel comparison (EC-19 vs Galaxy) 88 Illumina SRA runs, ARTIC protocol, 6 Bioprojects/98 Biosamples none shared vs platform-specific SNV/indel calls, allele frequency, depth of coverage BWA mem / Galaxy pipeline
Cross-workflow SNV/indel comparison (EC-19 vs Galaxy) 151 ONT SRA runs, ARTIC protocol, 6 Bioprojects/98 Biosamples none shared vs platform-specific SNV/indel calls, allele frequency, depth of coverage Minimap2 / Galaxy pipeline
Key results
  • EC-19 correctly predicted all lineages in the CDC benchmark set 55/55 (100%)
  • Illumina SNV and indel calls agreed between EC-19 and Galaxy 95% agreement
  • ONT SNV calls agreed between EC-19 and Galaxy 95% agreement
  • ONT indel calls agreed between EC-19 and Galaxy 55% agreement
  • ONT deletion calls agreed between EC-19 and Galaxy 82% (69/84) agreement
  • Majority of Illumina workflow differences occurred in low-coverage regions (DP<100x) 59/68 differences
  • Test dataset quality across 239 samples avg 1834x fold coverage, 88% linear coverage
  • Total variants reported by EC-19 across all projects 3115 SNVs, 139 indels (avg 13 SNVs, 0.6 indels/sample)
Key statistics
  • count 239 SRA datasets (151 ONT, 88 Illumina) (cross-workflow comparison dataset size)
  • count 55/55 lineages correctly predicted (CDC benchmark lineage validation)
  • other 95% agreement (Illumina SNV/indel concordance between EC-19 and Galaxy)
  • other 95% agreement (SNV), 55% agreement (indel) (ONT SNV vs indel concordance between EC-19 and Galaxy)
  • other 82% agreement (69 shared, 4 Galaxy-specific, 10 EC-19-specific) (ONT deletion concordance)
  • mean average fold coverage 1834x; average linear coverage 88% (overall test dataset quality)
  • count 3115 SNVs and 139 indels reported; avg 13 SNVs and 0.6 indels per sample (total EC-19 variant calls across all projects)
  • count 36 low-quality samples with <75% linear genome coverage (subset of low-quality samples in comparison dataset)

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.

This is a bioinformatics platform description and validation paper rather than a hypothesis-testing study. The authors validate the EC-19 workflow using benchmark and real-world sequencing datasets and compare its SNV, indel, and lineage calls against those from the COVID-19 Galaxy Project workflow. Results are reported as raw counts, percentages of agreement (concordance), and average values (e.g., mean depth of coverage, mean allele frequency) across samples, without inferential statistical hypothesis tests, p-values, or formal measures of uncertainty.

Replicationunclear Sample sizeSample counts for each comparison are explicitly stated (e.g., 239 SRA datasets, 55 CDC benchmark datasets, dozens of New Mexico surveillance genomes), but no a priori power or sample-size justification is described. GroupsSNV, indel, and lineage calls generated by the EC-19 workflow versus the same calls generated by the COVID-19 Galaxy Project workflow on identical sequencing datasets Pairingpaired Randomization/blindingnot stated Dispersionnone Exact p-valuesno Effect sizesno Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Descriptive concordance/percent agreement calculation (no inferential statistical test) Comparison of SNV and indel calls between EC-19 and Galaxy workflows (Table 1, Section 4.4-4.5) 239 SRA datasets (151 ONT, 88 Illumina) from six Bioprojects and 98 Biosamples na
Simple accuracy tally (correct/total) Lineage assignment validation against CDC benchmark datasets (Section 4.2) 55 SRR datasets with known lineages na
Approaches that could also have been used
  • Agreement between EC-19 and Galaxy workflows on SNV/indel calls is summarized as raw counts and percent agreement (e.g., 95% for Illumina SNVs).
    Could also: A formal paired-agreement statistic such as Cohen's kappa (for categorical call/no-call agreement) or McNemar's test (for paired discordant pairs) could also be used — These measures quantify agreement while accounting for chance concordance and can accompany the descriptive percentages already reported, giving a standardized effect size for inter-workflow agreement.
  • Sequencing depth of coverage and allele frequency are summarized using single average values (e.g., average fold coverage of 1834x, average AF of 0.46 for certain SNVs).
    Could also: Reporting the SD, IQR, or a 95% confidence interval alongside these means would also be a standard approach — A dispersion measure conveys how variable coverage and AF are across the sample set, which is relevant given the wide range noted (e.g., linear coverage from 1.4% to high values).
  • Discrepancies between workflows are attributed narratively to factors such as low depth of coverage or proximity to primer-binding sites.
    Could also: A regression model (e.g., logistic regression predicting discordant call probability from depth of coverage, allele frequency, and primer distance) could also be used — This would let the relationship between these factors and call discordance be quantified with effect estimates and confidence intervals rather than described qualitatively.
  • Lineage assignment accuracy is reported as 55/55 correct calls with no accompanying interval.
    Could also: A binomial confidence interval (e.g., Wilson or Clopper-Pearson) around the accuracy estimate could also be reported — With a relatively small benchmark set, an interval estimate communicates the plausible range of the true accuracy more fully than the point estimate alone.
  • The comparison dataset (239 SRA runs) was drawn from an existing collaborative resource (ACTIV TRACE) rather than a pre-specified statistical sample size.
    Could also: For this type of workflow-validation/benchmarking study, using an opportunistic or convenience benchmark set is standard practice and formal power analysis is not typically expected — Noted here for completeness: unlike hypothesis-driven experiments, tool-validation studies commonly rely on available reference datasets rather than a priori sample-size calculations.
Software: FaQCs (quality control) · Minimap2 (read mapping, ONT default) · BWA-MEM (read mapping, Illumina default) · Pangolin, --usher mode (lineage assignment) auto-updated to latest release per run · COVID-19 Galaxy Project workflow (comparator)

What was reproduced

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

Scope — pmid-35561186 (EDGE COVID-19 / EC-19)

Paper: Lo CC et al., EDGE COVID-19: a web platform to generate submission-ready genomes from SARS-CoV-2 sequencing efforts. Bioinformatics 2022. PMID 35561186 · PMC9113274 · DOI 10.1093/bioinformatics/btac176.

Type: Applications Note describing a bioinformatics web platform (EC-19). The pipeline: FaQCs read QC → align to modified RefSeq NC_045512.2 (33-nt poly-A trimmed) with minimap2 (ONT) / BWA-mem (Illumina) → variant calling (indel AF>0.5 Illumina, >0.6 ONT, >0.8 in homopolymers; ambiguous base if ≤5 supporting reads) → consensus genome → Pangolin (--usher) lineage → GenBank/GISAID/SRA submission packaging. JBrowse for visual inspection.

Code

Data

  • PRJNA714680 = the authors' OWN new data (New Mexico surveillance). ENA read_run: 44 runs, all Illumina MiSeq AMPLICON. Paper calls them "dozens of genomes" — no specific reproducible numeric claim is attached to this accession.
  • The paper's quantitative Results (239 samples, SNV/indel/coverage totals, Table 1 concordance) come from 9 OTHER public bioprojects that EC-19 reprocessed (PRJEB41737, PRJEB37886, PRJNA650037, PRJNA707211, PRJNA627229, PRJNA613958, PRJNA614995, PRJNA628662, PRJNA616147), whose EC-19 output the authors deposited in LANL-data.txt. PRJNA714680 is NOT in that table.

In scope (pipeline-derived) — what we reproduce

  • Tier 1 (done, no-compute): Re-derive the paper's aggregate Results numbers (N, avg/total SNVs & indels, avg/linear coverage, low-quality count, min coverage) directly from the authors' deposited LANL-data.txt and compare to the printed values. This is a fabrication-detection audit: does the deposited data support the paper's numbers? → claims C1–C9.
  • Tier 2 (pending «our HPC»): Run EC-19 (or its documented equivalent: minimap2/bwa→consensus→Pangolin) on a few accessions that have ground-truth rows in LANL-data.txt, plus the 55 CDC benchmark SRRs, to confirm per-sample variant calls / coverage / lineage reproduce the deposited values → C10, and individual-row validation of C2/C3/C6/C7.

Out of scope

  • C11/C12 (Table 1 EC-19-vs-Galaxy concordance): requires also running the external Galaxy SARS-CoV-2 workflow and intersecting variant sets — not a single-pipeline reproduction; not attempted.
  • Web-platform UI, JBrowse, automated GenBank/GISAID/SRA submission (manual / interactive / external services) — wet-lab/manual, out of scope.
Figures / tables: Table
C1
Reported
239 SRA datasets (151 ONT + 88 Illumina)
Reproduced
386 in deposited table (226 ONT + 160 Illumina)
did not match
C2
Reported
13 SNVs/sample
Reproduced
13.45 (all)
within tolerance
C3
Reported
3115 total SNVs
Reproduced
5192 on 386; ~3115 consistent at 239 samples
partial
C4
Reported
0.6 indels/sample
Reproduced
6.05 (all)
did not match
C5
Reported
139 total indels
Reproduced
2337 (all)
did not match
C6
Reported
1834x avg fold
Reproduced
2405 all / 1712 ONT; median 734
partial
C7
Reported
88% avg linear coverage
Reproduced
86.5% (all)
within tolerance
C8
Reported
36 samples <75% coverage
Reproduced
65 all / 36 ONT-only
partial
C9
Reported
min 1.4% coverage
Reproduced
1.41% (Illumina min)
within tolerance
C10
Reported
55/55 CDC lineages correct
Reproduced
pending («our HPC»)
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 49/100

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: 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 +6

Tier-1 audit re-derived the paper's Results 4.3 aggregates from the authors' own deposited per-variant table. Reproduces cleanly: per-sample SNV rate (13.45 vs 13), linear coverage (86.5% vs 88%), min coverage (1.41% vs 1.4%) — good evidence these numbers are genuine. Does not match 1:1: the deposit is a 386-sample superset of the reported 239, and its indel totals are ~10x the printed figures (2337 vs 139) because the deposit is unfiltered and the filtering/subsetting steps are undocumented. The discrepancies sit on the authors' side (incomplete documentation, not derivable as printed) but are explainable and do not overturn the tool's central claim; the flagship 55/55 CDC lineage benchmark remains pending. Overall partial / yellow.

🤝
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.

115.7 k
tokens (I/O) · 4.5 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.