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

Genomic regions and signaling pathways associated with indicator traits for feed efficiency in juvenile Atlantic salmon (Salmo salar).

Genet Sel Evol · 2020
L1 85/100 PQI 92
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: Q3 · Location of the main deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q6 · Severity of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Input / endpoint not comparable 1:1
+1 pts
From: Q1 · Data identity 🔴
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +9
✓ What held up
  • Nothing in this column.
What did not (or only partly)
  • 🔴Could not use the authors’ exact input data
  • 🟡Reported values were only indirectly comparable
  • 🟡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
85/100
Reproducibility score
0.6 SD above mean
vs. all fields · 1173 studies
🎯 Scores higher than 67% of all assessed papers rank 348 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 (honest mix; some claims reproduce 1:1, KEGG mismatches, the GWAS half is a data-restricted drop). The paper has two computational tracks. Track A 'Genomic regions/QTL' (Figs 1-4, Table 1; GCTA mlm-loco GWAS) is NOT reproducible: the Axiom SNP genotypes are owned by AquaGen AS and 'not publicly available', phenotypes only 'on request' -> data_restricted, not attempted. Track B 'signaling pathways' (RNAseq, Fig 5, Table 2): raw fastq ARE public (E-MTAB-8305, 184 liver samples = matches paper exactly) and the pipeline tool is public (bcbio-nextgen, third-party), BUT the EdgeR trait-association regression that produces every reported number needs per-fish phenotypes (ALC, ALN, growth FW/IW, family) that are restricted -- the E-MTAB-8305 SDRF carries only placeholder metadata (body weight '+/- 20'). So a from-raw-fastq rerun on «our HPC» would yield a count matrix that cannot be tied to any reported value; per the 80/20 rule I did NOT burn compute on an ungradeable result (no SLURM jobs). Instead I reproduced the paper's FINAL, clearly-specified steps directly from the authors' OWN deposited supplementary tables: applying q<0.05 + sign-of-slope to the deposited per-gene table (Table S1, 31235 genes) reproduces the TAG counts essentially exactly (ALC 799/741 exact, ALN 900->899 off-by-one at the rounding boundary, shared 317/281 exact), and all 7 transcription factors in Table 2 match the deposited slope+q values exactly to 2 dp -- no fabrication detected in the headline RNAseq numbers. The KEGG pathway counts are the one discrepancy: counting p<0.05 rows in the deposited Table S2 gives 50/68/15/14 vs the reported 59/88/24/35 (systematically fewer) -- flagged for human review (likely analysis-versioning/filtered table, not assertable as fabrication; the qualitative claims -- Proteasome most enriched, more positive than negative pathways -- hold). NOT attempted: the GWAS/QTL track (restricted genotypes), the full bcbio-nextgen alignment/counting rerun (ungradeable without restricted phenotypes), SalMotifDB TF DB query, and KEGG kegga rerun. This is a deposited-output / internal-consistency verification, NOT a from-raw-data pipeline reproduction, and is labelled as such throughout.

💻 Code ↗ 🗄 Data: E-MTAB-8305

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 85
    assessed: 2026-06-15 ⛓ 5aac8a46e0dc
✎ 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
2026-06-15
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

Can putative QTL and liver gene-expression signatures be identified for indicator phenotypes of growth, nitrogen/carbon metabolism, and feed efficiency in juvenile Atlantic salmon, in order to establish the genetic basis of feed-efficiency indicator traits?

Core claims
  • A QTL for pre-smolt growth was identified on chromosome 9. finding
  • A QTL for carbon metabolism in the liver (ALC) was identified on chromosome 12, closely related to tank-level feed conversion ratio. finding
  • Isotope-derived feed-efficiency indicator traits (from 15N and 13C of muscle and growth) showed no convincing QTL, suggesting they are polygenic. finding
  • Fish with high carbon and nitrogen metabolism in the liver convert feed protein more efficiently, primarily due to higher expression of proteasome, lipid, and carbon metabolic pathways in liver. mechanism
  • Seven transcription factors associated with carbon and nitrogen metabolism were located within the identified QTL regions. finding
  • Stable-isotope-based indicator traits (IFCR/IFER) enable individual-level assessment of feed efficiency usable in GWAS and marker-assisted selection. method
  • Including nitrogen and carbon metabolism traits substantially improved prediction of tank-FCR (73% of variance explained vs 53-63% by growth and sampling day alone). finding
Experimental setups
Assay System Perturbation Readout Platform
Stable isotope analysis (element analysis isotope ratio mass spectrometry, atom % 13C and 15N) Atlantic salmon (Salmo salar) juveniles, muscle/liver/adipose tissue; 2249-2280 fish per trait diet labeled with 15N (2%) and 13C (1%) during 12-day feed conversion test atom % 13C/15N in muscle (AMC, AMN), liver (ALC, ALN), adipose (AAC); IFCR/IFER indicators isotope ratio mass spectrometry (Institute for Energy Technology, Kjeller, Norway)
SNP genotyping array / GWAS Atlantic salmon, fin-clip DNA from 2300 fish, 23 full-sib families none SNP associations with growth, metabolism, and feed-efficiency indicator traits (54,200 SNPs after filtering) AquaGen custom Axiom SNP array (Thermo Fisher/Affymetrix, 56,177 SNPs)
RNAseq (transcriptomic analysis) Atlantic salmon liver, 184 fish from all families none gene expression associated with carbon and nitrogen metabolism in liver
Growth phenotyping (weight) Atlantic salmon, 2281 fish in family tanks 12-day feed conversion test, fishmeal-based diet weight gain (WG), relative weight gain (RG), initial/final weight
DNA extraction and quantification Atlantic salmon fin clips (20 mg) none DNA concentration Sbeadex livestock kit (LGC Genomics); Nanodrop 8000 (Thermo Fisher Scientific)
Key results
  • QTL for pre-smolt growth detected on chromosome 9
  • QTL for liver carbon metabolism (ALC) detected on chromosome 12, related to tank-FCR
  • No convincing QTL for isotope-derived muscle feed-efficiency indicator traits, indicating polygenic architecture
  • Higher liver expression of proteasome, lipid, and carbon metabolic pathways in more protein-efficient fish
  • Seven transcription factors associated with carbon/nitrogen metabolism located in QTL regions
  • Growth, isotope-based indicators, and sampling day jointly explained 73% of variance in tank-FCR 73%
  • Genetic correlation of tank-FCR with muscle nitrogen/carbon metabolism indicators ~1.0 and with liver carbon metabolism (ALC) ~0.9 (prior study) rg~1.0; rg~0.9
Key statistics
  • count 2281 (total fish in 12-day feed conversion test)
  • count 54,200 SNPs (SNPs included after filtering)
  • count 56,177 SNPs (SNPs on AquaGen custom Axiom array)
  • pvalue 9.23×10^-7 (−log10 p = 6.03) (Bonferroni 5% genome-wide significance threshold (0.05/54,200))
  • correlation rg ~1.0 (genetic correlation of tank-FCR with muscle 15N/13C metabolism indicators)
  • correlation rg ~0.9 (genetic correlation of tank-FCR with liver carbon metabolism (ALC))
  • other 73% vs 53-63% (variance in tank-FCR explained with vs without isotope metabolism traits)
  • count 184 fish (number of fish used in RNAseq liver analysis)

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 GWAS study in Atlantic salmon (n = 2249–2280 per trait, from 2281 fish across 23 full-sib families) tested associations between 54,200 SNPs and 11 indicator traits for feed efficiency, growth, and nitrogen/carbon metabolism using a linear mixed model with a leave-one-chromosome-out genomic relationship matrix (GCTA --mlm-loco), with significance assessed via a built-in likelihood-ratio test. Genome-wide significance was set using a Bonferroni threshold (α = 0.05 / 54,200 SNPs = 9.23 × 10⁻⁷; −log10(p) = 6.03), with chromosome-wide thresholds also calculated. A separate RNA-seq analysis on 184 fish was conducted to identify genes associated with carbon and nitrogen metabolism in liver, though the RNA-seq pipeline and differential expression method are not described in the provided text excerpt. Results were visualized using Manhattan plots and QQ-plots; no allelic effect sizes or confidence intervals were reported.

Replicationbiological Sample size2281 fish from 23 full-sib families; 50 fish per tank, 2 tanks per family (46 tanks total); 184-fish subset for RNA-seq; families randomly allocated to tanks; no formal power calculation described Groups23 full-sib families with divergently selected parents (high vs. low EBV for growth); GWAS treats SNP genotype as a continuous additive covariate (coded 0/1/2) Pairingunpaired Randomization/blindingstated Dispersionnone Exact p-valuesno Effect sizesno Confidence intervalsno Multiplicity correctionBonferroni (0.05 / 54,200 = 9.23 × 10⁻⁷; −log10(p) = 6.03); chromosome-wide Bonferroni thresholds also computed per chromosome
Statistical tests used
Test Applied to n Assumptions
Linear mixed model with likelihood-ratio test (GCTA --mlm-loco GWAS) SNP association testing for 11 traits: WG, RG, AMC, AMN, ALC, ALN, AAC, IFCR_AMC, IFCR_AMN, IFER_AMC, IFER_AMN 2249–2280 fish per trait (from 2281 total genotyped individuals) not stated
RNA-seq differential expression analysis (method and software not specified in provided text) Identifying genes associated with carbon and nitrogen metabolism in liver tissue 184 fish not stated
Approaches that could also have been used
  • Genome-wide significance was assessed with a Bonferroni threshold, and the authors themselves note it is conservative for LD-correlated SNPs
    Could also: A permutation-based genome-wide threshold (e.g., 1,000 phenotype permutations) or a Benjamini-Hochberg FDR correction could also be used — Permutation-based thresholds directly account for the LD correlation structure among SNPs, potentially providing better-calibrated genome-wide error rates and recovering associations that a strict Bonferroni correction would miss
  • Eleven traits were analyzed in separate single-trait GWASs without any multiplicity correction across traits
    Could also: A multi-trait GWAS (e.g., using GEMMA's multivariate LMM, MultiPhen, or mvBIMBAM) or a Bonferroni/FDR adjustment across traits could also be applied — Because several traits showed near-unity genetic correlations, a multi-trait framework jointly models correlated phenotypes, which can increase power for pleiotropic loci while naturally accounting for the cross-trait testing burden
  • The GWAS model included a polygenic random effect via a GRM (--mlm-loco) but did not include tank as an explicit random effect, even though 50 fish per tank shared a common 12-day tank environment
    Could also: Tank could be included as an additional random effect in the linear mixed model (Yi = a + bx + tankj + gi⁻ + εi) — Fish sharing the same tank may exhibit correlated residuals due to shared feeding conditions and social effects; explicitly modeling tank variance partitions this source of environmental covariance from the polygenic component and may improve precision of SNP effect estimates
  • Allelic substitution effect estimates and the proportion of phenotypic or genetic variance explained by identified QTL are not reported
    Could also: The b coefficients from the fitted model, along with the QTL variance explained (e.g., as % of total genetic variance or heritability), could also be reported for each lead SNP — Effect size metrics allow assessment of biological magnitude independently of sample size, facilitate cross-study comparison, and are directly relevant to the marker-assisted selection application the paper proposes
  • The RNA-seq analysis method, normalization strategy, and differential expression tool are not described in the provided text excerpt
    Could also: Standard RNA-seq DE workflows such as DESeq2 (negative binomial Wald test with BH FDR), edgeR (quasi-likelihood F-test), or limma-voom (moderated t-test) are widely used for this data type — Specifying the pipeline, normalization method, and FDR threshold enables reproducibility assessment and direct comparison with other transcriptomic studies of feed efficiency in salmonids
  • Family structure was handled by including a genome-wide GRM as a random effect; the experimental design used only 23 families with divergent EBVs
    Could also: A pedigree-based mixed model (using a numerator relationship matrix A rather than the genomic G) or a within-family regression approach could also control for family structure in this full-sib design — With only 23 families and strong divergent selection of parents, a within-family design or haplotype-based association test may better separate QTL effects from between-family confounding that could be incompletely captured by the GRM in a small, structured population
Software: GCTA · Axiom Power Tools (Affymetrix) — genotype calling only

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
18
Impact: medium
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.

E-MTAB-8305 ArrayExpress in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet

What was reproduced

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

Scope — pmid-33158415

Dvergedal et al. 2020, Genet Sel Evol 52:66. "Genomic regions and signaling pathways associated with indicator traits for feed efficiency in juvenile Atlantic salmon (Salmo salar)." DOI 10.1186/s12711-020-00587-x.

The paper has two distinct computational tracks.

Track A — "Genomic regions" / QTL (Figs 1–4, Table 1) → OUT OF SCOPE (data_restricted)

  • Pipeline: GCTA --mlm-loco linear mixed-model GWAS over 54,200 Axiom SNPs vs 12 phenotypes; Bonferroni −log10(p)=6.03 threshold; QTL on Ssa9 (growth) and Ssa12 (carbon metabolism).
  • Inputs are NOT public. Genotypes are "owned by AquaGen AS and used under license … not publicly available" (Availability statement). Phenotypes only "on request." No accession. → cannot be reproduced from public data.
  • Not attempted. drop_reason for this track: data_restricted.

Track B — "Signaling pathways" / RNAseq (Fig 5, Table 2) → PARTIALLY IN SCOPE

  • Pipeline as described: bcbio-nextgen (trim, align to ICSASG_v2, count) on E-MTAB-8305 → EdgeR per-gene regression of expression on the trait (ALC or ALN), with growth (FW−IW)/FW as a covariate and family intercepts → TAG at FDR q<0.05 → KEGG over-representation via limma kegga → TF annotation via SalMotifDB.
  • Raw fastq are public (E-MTAB-8305, 184 liver samples — matches the paper's "184 fish" exactly). The pipeline code is a public third-party tool (bcbio).
  • BUT the regression's covariates — per-fish ALC, ALN, growth (FW,IW), family — are NOT public (same "on request" phenotype restriction). The E-MTAB-8305 SDRF carries only placeholder metadata (body weight = "+/- 20", genotype = "wild type", sex = "not available"); no isotope/trait/growth/family fields. → A from-raw-fastq reproduction of the reported TAG/KEGG numbers is not possible from public data (the heavy bcbio run would yield a count matrix that cannot be tied to any reported value).

What IS reproducible in Track B (no compute)

The authors deposited their per-gene regression output (Additional file 1 / Table S1: geneID, baseMean, ALC.slope, ALC.padj, ALN.slope, ALN.padj for 31,235 genes) and the KEGG pathway p-values (Additional file 2 / Table S2, 155 pathways). The paper's final, clearly-specified steps can therefore be reproduced directly from the deposited data:

  1. Apply q<0.05 + sign-of-slope to Table S1 → the reported TAG counts.
  2. Apply p<0.05 to Table S2 → the reported KEGG pathway counts.
  3. Cross-check the 7 transcription factors in Table 2 against Table S1.

This is an internal-consistency / deposited-output verification (and a fabrication check), not a from-raw-data pipeline reproduction. Graded and clearly labelled as such.

80/20 decision

The clearly-specified low-hanging outputs (TAG counts, Table-2 TFs, KEGG counts) are checkable from deposited data with zero compute and are done here. The hard last 20% — re-running bcbio-nextgen on 184 fastq to regenerate the count matrix — was deliberately NOT attempted: it cannot reach any reported number without the restricted phenotypes, so it would consume «our HPC» compute for an ungradeable result. No SLURM jobs were submitted.

Figures / tables: Fig 5aFig 5bTableFig1Fig2
C1
Reported
799 TAG positively associated with ALC (13C liver), q<0.05
Reproduced
799 (apply q<0.05 + slope>0 to deposited Table S1, 31235 genes)
exact
C2
Reported
741 TAG negatively associated with ALC, q<0.05
Reproduced
741
exact
C3
Reported
900 TAG positively associated with ALN (15N liver), q<0.05
Reproduced
899 (off-by-one at the q=0.05 rounding boundary; deposited padj 4dp, many at exactly 0.0500)
within tolerance
C4
Reported
978 TAG negatively associated with ALN, q<0.05
Reproduced
978
exact
C5
Reported
317 genes shared among positive TAG of ALC and ALN
Reproduced
317
exact
C6
Reported
281 genes shared among negative TAG of ALC and ALN
Reproduced
281
exact
C11
Reported
7 transcription factors in the QTL regions (Table 2: b1 + q for ALC & ALN)
Reproduced
all 7 TFs found in deposited Table S1; every slope and q matches Table 2 exactly to 2 dp
exact
C7-C10
Reported
KEGG pathways p<0.05: positive 59 (ALC) / 88 (ALN); negative 24 (ALC) / 35 (ALN)
Reproduced
from deposited Table S2: 50 / 68 / 15 / 14 -- systematically fewer than reported
did not match
C12-C13
Reported
QTL for pre-smolt growth on Ssa9 and for liver carbon metabolism on Ssa12 (GCTA mlm-loco, -log10p>6.03)
Reproduced
not attempted -- AquaGen-proprietary SNP genotypes are not public and phenotypes are on-request
partial
C14
Reported
RNA from liver of 184 fish
Reproduced
E-MTAB-8305 SDRF contains exactly 184 individuals
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 85/100

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: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q6 · Severity of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Input / endpoint not comparable 1:1
+1 pts
From: Q1 · Data identity 🔴
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +9

The RNAseq track reproduces well from the authors' own deposited tables: TAG counts (799/741/978, shared 317/281) and all 7 Table-2 transcription factors match Table S1 essentially 1:1, with only a single rounding off-by-one (ALN 900→899) — no fabrication evident in the headline numbers. The one real deviation is the KEGG pathway counts (59/88/24/35 reported vs 50/68/15/14 derivable from deposited Table S2), a moderate, systematic gap most likely from table versioning/filtering, with the qualitative enrichment conclusions still holding. The GWAS/QTL half is data-restricted (proprietary genotypes, on-request phenotypes) and therefore untested rather than refuted, so this is a partial but honest reproduction whose deviations sit on the data-availability and deposited-table-version side, not on a demonstrable authors' computational defect.

🤝
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.

100.1 k
tokens (I/O) · 5 M incl. cache
12 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.