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

Evolution and codon usage bias of mitochondrial and nuclear genomes in Aspergillus section Flavi.

G3 (Bethesda) · 2023
L1 85/100 3/4
Why this verdict

The main results reproduced: recomputed values matched the published ones within tolerance.

Reproduced on the brainbox compute brainarbeit.com
✓ What held up
  • Reported values are derivable from the shared data
  • Any deviation was negligible
  • The central claim held under reproduction
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
  • 🟡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

DESCRIBED WELL ENOUGH; largely 1:1 on the deterministic mitochondrial-side results. We independently recomputed mitogenome size (29100-39268 bp = paper 29.100-39.269 kbp, exact), GC min (24.94%=24.9%, exact), CDS 15-17, rRNA 2, tRNA 26, and the headline codon-usage RSCU values (AUA 1.62, CCC 0.043, CCU 2.72 vs paper 1.6304/0.0449/2.7140) directly from the authors' DEPOSITED GenBank mitogenomes (ON833077-ON833094) using an independent Biopython reimplementation -- a strong fabrication check: the reported numbers ARE derivable from the shipped data. Key methodological finding: paper RSCU = GENOME-level pooling (our 1.62 matches) not gene-level mean (1.56). Nuclear RSCU cross-checked vs the authors' figshare table (matches paper). NOT reproduced 1:1: exact ENc-GC3 R^2 (DAMBE-estimator-dependent; direction only), nuclear-side recompute / S-values / 2422 orthologs / phylogeny (JGI nuclear CDS login-gated + heavy compute). One data-quality flag: GenBank deposit labels A.transmontanensis twice (ON833080/81) and omits A.alliaceus (present in figshare) -- NCBI metadata labeling quirk, not a result error.

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: 2026-06-22 ⛓ 1325a765b69b
✎ 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
2026-06-22
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-22
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

The paper tests how the evolutionary history and codon usage bias of mitochondrial genomes in Aspergillus section Flavi compare to those of the nuclear genomes of the same species.

Core claims
  • 18 new mitochondrial genomes were assembled for Aspergillus section Flavi species, complementing 3 existing reference mitogenomes, for a total of 20 species analyzed. resource
  • All Flavi mitogenomes are circular DNA molecules of low GC content with highly conserved gene content and gene order (synteny). finding
  • Genomic content, including GC content and genome size, differs greatly between mitochondrial and nuclear genomes. finding
  • A phylogeny built from 14 concatenated mitochondrial genes predicts evolutionary relationships largely consistent with a phylogeny built from 2,422 nuclear genes. finding
  • Species cluster differently by codon usage bias depending on whether mitochondrial or nuclear relative synonymous codon usage (RSCU) values are used. finding
  • At the gene level, codon usage bias patterns are more similar among mitogenomes of different species than between the mitogenome and nuclear genome of the same species. finding
  • Mitogenomes are not under translational selection, whereas nuclear genomes are under moderate translational selection, despite most genes in both genomes deviating from neutral codon usage expectations. finding
  • The S-value (correlation between tRNA adaptation index and codon usage) can be used to compare the strength of translational selection between mitochondrial and nuclear genomes. method
Experimental setups
Assay System Perturbation Readout Platform
mitogenome assembly (organelle genome assembly from WGS reads) 20 Aspergillus section Flavi species (whole genome sequencing reads / mycelia) none assembled circular mitochondrial genome sequence GetOrganelle v1.7.4.1
read mapping / variant calling for error correction Aspergillus section Flavi species (paired-end Illumina reads vs assembled mitogenome) none mapped reads, identified variants Bowtie2 v2.3.4.1, SAMtools v1.6, IGV v2.9.4
organellar genome annotation circularized mitochondrial genomes of Flavi species none annotated gene content, gene names, translations GeSeq v2.03, Geneious Prime v2021.1, OGDraw v1.1.1
multiple sequence alignment and maximum likelihood phylogenetics (mitochondrial) 14 core mitogenes across 20 Flavi species plus A. niger outgroup none mitochondrial phylogenetic tree with bootstrap support MAFFT v7, SequenceMatrix v1.9, ClipKIT v1.3.0, RAxML v8.2.11
ortholog identification and maximum likelihood phylogenetics (nuclear) 2,422 orthologous nuclear genes across Flavi species (excluding A. sojae) plus A. niger outgroup none nuclear phylogenetic tree with bootstrap support OrthoFinder v2.5.4, RAxML v8.2.11
relative synonymous codon usage (RSCU) hierarchical clustering mitochondrial and nuclear genomes/genes of Flavi species none genome-level and gene-level RSCU values, clustering of species DAMBE v7.3.5, RStudio v2021.09.1
correspondence analysis of codon usage gene-level RSCU values, mitochondrial vs nuclear genes across Flavi species none codons driving differences in codon usage signatures R packages ade4 v1.7-19 and factoextra v1.0.7
effective number of codons (ENc) vs GC3 analysis and S-value translational selection test mitochondrial and nuclear genomes/tRNAomes of Flavi species none ENc values, GC3 content, tRNA adaptation index (stAI), S-value with permutation-based significance DAMBE v7.3.5, tRNAscan-SE 2.0, stAIcalc v1.0, R package tAI.R v0.2
Key results
  • All 20 Flavi mitogenomes are small, circular DNA molecules with low GC content 24.9-26.9%
  • Each mitogenome contains the same 14 core genes with conserved order and shared synteny
  • A ribosomal protein S3 gene was found in all newly annotated Flavi mitogenomes
  • Mitochondrial phylogeny (14 genes) is largely consistent with nuclear phylogeny (2,422 genes)
  • Species group differently by codon usage bias depending on use of mitochondrial vs nuclear RSCU values
  • Gene-level codon usage bias is more similar between mitogenomes of different species than between mitochondrial and nuclear genome of the same species
  • Most nuclear and mitochondrial genes deviate from the neutral expectation for codon usage, but mitogenomes show no evidence of translational selection while nuclear genomes show moderate translational selection
Key statistics
  • other 24.9-26.9% (GC content range of Flavi mitogenomes)
  • count 14 core mitochondrial genes (conserved protein-coding genes per mitogenome)
  • count 2,422 orthologous nuclear genes (genes used to build concatenated nuclear phylogeny)
  • count 18 newly assembled mitogenomes (of 20 total analyzed) (mitogenome dataset composition)
  • other 1,000 bootstrap replicates (robustness evaluation for both mitochondrial and nuclear ML phylogenies)
  • other 100 permutations (permutation test for statistical significance of S-value)
  • other top 5% of permutation distribution (significance threshold for observed S-value)
  • other S-value (dos Reis et al. 2004 method) (metric quantifying strength of translational selection on codon usage)

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 study is a comparative genomics analysis of mitochondrial versus nuclear genomes across 20 Aspergillus section Flavi species. The main analyses are maximum-likelihood phylogenetic reconstruction with bootstrap support, multivariate correspondence analysis of relative synonymous codon usage (RSCU) values, hierarchical clustering of RSCU patterns, and a permutation-based test (S-value, following dos Reis et al. 2004) to assess translational selection. Results are reported primarily as bootstrap support percentages, R² values (observed vs. expected effective number of codons), and S-values with permutation-derived significance thresholds, rather than through classical hypothesis-testing statistics like t-tests or ANOVA.

Replicationunclear Sample sizeSample size is described in terms of number of species (20 section Flavi species), genes (14 mitogenes, 2,422 nuclear orthologs), bootstrap replicates (1,000), and permutations (100), rather than a traditional biological-replicate count or power calculation GroupsMitochondrial vs. nuclear genomes/genes across 20 Aspergillus section Flavi species Pairingna Randomization/blindingnot stated Dispersionunclear Exact p-valuesno Effect sizesyes Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Maximum likelihood phylogenetic inference (RAxML v8.2.11, GTR+Γ model) with bootstrap support Mitochondrial phylogeny (14 concatenated mitogenes) and nuclear phylogeny (2,422 orthologous genes) 14 concatenated mitochondrial genes / 2,422 nuclear orthologs across up to 20 species; 1,000 bootstrap replicates stated
Correspondence analysis (CA) of gene-level RSCU values Identifying codons driving differences in codon usage between mitochondrial and nuclear genes and among species' mitogenomes not stated (gene-level RSCU values across annotated genes) stated
Hierarchical clustering of genome-level and gene-level RSCU values Comparing patterns of codon usage bias between mitochondrial and nuclear genomes not stated not stated
R² comparison of observed vs. expected effective number of codons (ENc) against the dos Reis et al. (2004) neutral expectation Evaluating the role of mutational bias (GC3) in codon usage not stated stated
S-value (dos Reis et al. 2004 framework) with permutation test (100 permutations; significance = top 5% of permuted distribution) Assessing strength of translational selection per mitochondrial and nuclear genome 100 permutations per genome stated
Approaches that could also have been used
  • Phylogenies were inferred using a single maximum-likelihood method (RAxML) with bootstrap support to assess node confidence.
    Could also: Bayesian phylogenetic inference (e.g., MrBayes or BEAST) or gene concordance/discordance factor analysis alongside ML — Bayesian posterior probabilities offer a complementary support metric, and concordance factors can additionally reveal how much individual gene trees agree with the concatenated species tree, which is often informative when combining many loci.
  • Correspondence analysis was chosen over principal component analysis because RSCU values for synonymous codons are interdependent (compositional).
    Could also: PCA applied after a compositional-data transformation (e.g., centered log-ratio or Hellinger transform) of RSCU values — This is a commonly used alternative for compositional codon-usage data that also addresses the interdependence among synonymous codon frequencies while allowing use of standard PCA machinery.
  • Significance of each S-value was determined using a permutation test based on 100 permutations, with significance defined as falling in the top 5% of the permuted distribution.
    Could also: A larger number of permutations (e.g., 1,000-10,000) with an exact empirical p-value reported — Increasing the permutation count narrows the resolution of the empirical null distribution and allows reporting of a precise p-value rather than a fixed percentile threshold, which some readers find easier to compare across genomes.
  • Translational selection (S-values) and ENc-GC3 fits were evaluated separately for each of the ~20 species' genomes.
    Could also: A false discovery rate procedure (e.g., Benjamini-Hochberg) applied across the set of per-genome permutation tests — When many similar tests are run across species in parallel, an FDR correction is a standard way to control the expected proportion of false positives among the results declared significant.
  • Hierarchical clustering was used to group genomes/genes by RSCU similarity without reported support values for the resulting clusters.
    Could also: Clustering with bootstrap or jackknife-based support values (e.g., via the pvclust R package) or an approximately unbiased test — Adding support values to clusters would convey the confidence in specific groupings, analogous to the bootstrap support already reported for the phylogenies.
  • The fit between observed and expected ENc values was summarized as a single R² statistic.
    Could also: Reporting a confidence interval around R² or a formal goodness-of-fit test statistic — An interval estimate would communicate the uncertainty around the point estimate of fit rather than relying on R² alone.
Software: RAxML 8.2.11 · MAFFT 7 · ClipKIT 1.3.0 · OrthoFinder 2.5.4 · DAMBE 7.3.5 · R/ade4 1.7-19 · R/factoextra 1.0.7 · tAI.R (R package) 0.2 · stAIcalc 1.0 · RStudio 2021.09.1

What was reproduced

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

Scope — pmid-36305682

Paper: Hugaboom, Hatmaker, LaBella, Rokas (2023). Evolution and codon usage bias of mitochondrial and nuclear genomes in Aspergillus section Flavi. G3 (Bethesda) 13(2):jkac285. PMID 36305682 · PMC9836360 · DOI 10.1093/g3journal/jkac285.

What the paper does (pipeline overview)

20 Aspergillus section Flavi species (+ A. niger outgroup). For each: mitogenome (3 reference + 18 newly assembled from WGS reads) and nuclear genome (JGI MycoCosm). They (a) assemble + annotate mitogenomes, (b) characterize genome size / GC / gene content, (c) compute codon usage bias (RSCU, ENc), (d) test translational selection (tAI / S-values), (e) correspondence analysis of codon usage, (f) build mito + nuclear phylogenies.

In scope (pipeline-derived, computationally reproducible)

# Result Pipeline / tool Input data Tractability
S1 Mitogenome size range 29.1–39.3 kbp (Table 2) parse GenBank deposited ON8330xx mitogenomes high — direct
S2 Mitogenome GC% 24.9–26.9% (Table 2) seq GC calc deposited mitogenomes high — direct
S3 Mitogenome CDS count 15–17; rRNA 2; tRNA 26 (Table 2 / text) parse GenBank features deposited mitogenomes high — direct
S4 RSCU values (Fig 5; figshare RSCU table) e.g. AUA 1.63 mito / 0.45 nuc; CCC 0.045/1.05; CCU 2.71/1.07 RSCU from CDS (DAMBE → we use codon counts) mito CDS (+ nuc CDS) high for mito; compare to authors' own figshare RSCU xlsx
S5 ENc + ENc–GC3 regression R²: mito 0.211, nuc 0.598 ENc (Wright) vs GC3 mito/nuc CDS medium
S6 Correspondence analysis variance: dim1 15.6%, dim2 7% (codon level) ade4 CA on RSCU RSCU matrix medium
S7 Translational selection S-values (mito median 0.162 / mean 0.137, none sig; nuc median 0.432 / mean 0.427, all sig) stAIcalc + tAI.R tRNA copy numbers + CDS lower — needs tRNA GCN + permutation
S8 Mito phylogeny (14-gene concat, RAxML GTR+Γ, 1000 bs); branch dist A.flavus–A.nomiae 0.244 mito / 0.086 nuc MAFFT→catfasta2phyml/SequenceMatrix→ClipKIT→RAxML mito gene alignments heavy («our HPC» SLURM)

Primary target (the clean 1:1)

S1–S3 (genome size/GC/gene content) and S4 (RSCU) computed directly from the authors' own deposited annotated mitogenomes (GenBank ON833077–ON833094) and cross-checked against their own figshare RSCU table. This is a faithful 1:1 reproduction that does not require re-running the (heavy, semi-manual) assembly, and is fully human-auditable.

Out of scope (not pipeline-reproducible / wet-lab / manual / external)

  • WGS sequencing itself (wet lab) — only reuse the SRA reads.
  • Mitogenome assembly GetOrganelle/Bowtie2 + manual Geneious inspection + GeSeq annotation with manual curation — semi-manual; not a deterministic pipeline. We instead reproduce downstream metrics from the deposited result. (Optionally attempt one assembly on «our HPC» as a spot-check if time allows.)
  • Nuclear ortholog calling / RAxML species tree (2,422 orthologs) — heavy; attempt only as stretch.
  • Any narrative/qualitative claims ("largely consistent topology", "high support").

Code artifact (P16 note)

The linked repo github.com/nylander/catfasta2phyml is a third-party Perl tool for concatenating FASTA alignments into PHYLIP/phyml input — used by the authors for the mito phylogeny. Per brief P16, applying it to the paper's data is a valid reproduction, but it is peripheral to the central codon-usage results. There is no custom authors' analysis-code repository; codon usage was computed in DAMBE/R (GUI tools). We therefore reproduce the codon-usage numbers with an independent, scripted reimplementation (Biopython codon counts → RSCU/ENc), which is the stronger fabrication check.

Data inventory

  • GenBank mitogenomes (this study): ON833077, ON833078, ON833079, ON833081–ON833094
Figures / tables: TableFig 2Fig 5Fig 7Fig 8
S1
Reported
29.100-39.269 kbp
Reproduced
29100-39268 bp
exact
S2a
Reported
GC% min 24.9%
Reproduced
24.94%
exact
S2b
Reported
GC% max 26.9%
Reproduced
26.19%
partial
S3a
Reported
CDS 15-17
Reproduced
15-17 (incl. ref A.sojae=17)
within tolerance
S3b
Reported
rRNA 2
Reproduced
2
exact
S3c
Reported
tRNA 26 (conserved)
Reproduced
26 (figshare S4 26/26); GenBank parse 25-26
within tolerance
S4a
Reported
RSCU AUA mito 1.6304
Reproduced
1.62 (genome-level)
within tolerance
S4c
Reported
RSCU CCC mito 0.0449
Reproduced
0.043 (genome-level)
within tolerance
S4e
Reported
RSCU CCU mito 2.7140
Reproduced
2.72 (genome-level)
within tolerance
S4b
Reported
RSCU AUA nuc 0.4525
Reproduced
0.4544 (authors figshare table)
within tolerance
S4d
Reported
RSCU CCC nuc 1.0457
Reproduced
1.0968 (authors figshare table)
within tolerance
S4f
Reported
RSCU CCU nuc 1.072
Reproduced
1.0526 (authors figshare table)
within tolerance
S5b
Reported
ENc-GC3 R^2 mito 0.211
Reproduced
method-dependent; direction reproduced (poor fit)
partial
N1
Reported
20 (+1 outgroup)
Reproduced
20 Flavi + A.niger
exact
N2
Reported
18 newly assembled
Reproduced
18 ON deposits
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)
🤝
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.

247.3 k
tokens (I/O) · 12.8 M incl. cache
80 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.