Stem rust resistance in wheat is suppressed by a subunit of the mediator complex.
The main results reproduced: recomputed values matched the published ones within tolerance.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- ✓Same input data as the authors
- ✓Reported values were directly comparable
- ✓No relevant deviation in data/preprocessing
- ✓No authors-side cause for any deviation
- ✓Reported values are derivable from the shared data
- ✓Any deviation was negligible
- ✓The central claim held under reproduction
- ✓Overall, the reproduction was clean
- Every checked point held up.
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 to reproduce the core computational result 1:1. S1 (Med15 dN/dS molecular-evolution analysis, PAML codeml branch models) is fully self-contained in the authors' repo (shipped alignment + labelled RAxML tree + codeml control files AND the authors' own result files). Re-running codeml (bioconda PAML 4.10.10 vs paper's 4.8) on the shipped inputs reproduced all six hypotheses bit-identically to 6 decimals, including the paper's printed omega values omega_a=0.38 / omega_b=0.37 / omega_0=0.24 (= reproduced 0.37702/0.37155/0.24407 in the H3 3-ratio model) and the LRT (2dlnL=16.79, p=2.3e-4) supporting relaxed purifying selection on Med15a/Med15b. Two bonus codeml datasets (full Poaceae, complete-sequence) also bit-identical. This is a clean third-party 1:1 with no fabrication signal. NOT attempted: S2, the RNAseq DESeq2 '229 differentially expressed genes' number — its count matrix and design table are not shipped, the README's DESeq2 code is for a different (time-course) experiment, and reproducing it would need the full SRA PRJNA401266 download + a ~10 GB IWGSC RefSeq v1.0 transcriptome index with an under-specified design; recorded honestly as not-attempted rather than forced. All wet-lab, de novo assembly, haplotype-conversion, and protein-annotation results are out of scope.
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.
-
v1 current initial assessment Score 95assessed: 2026-06-16 ⛓ 5750c6b0efca
✎ 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.
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
- 2026-06-16
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-18no 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: opusDoes the dominant SuSr-D1 locus on wheat chromosome 7DL suppress race-specific stem rust resistance, and what gene underlies this suppression?
- ★ SuSr-D1 encodes Med15b.D, a subunit of the conserved Mediator complex that regulates transcription of protein-coding genes. finding
- ★ Nonsense (loss-of-function) mutations in Med15b.D abolish suppression and confer race-specific stem rust resistance. finding
- ★ Differential gene expression at 24 hpi is not required for immunity to Pgt; transcriptional reprogramming at this time point is dispensable for resistance. mechanism
- ★ Med15b.D exhibits homeolog-dominant suppression of immunity that is generally compensated by other Med15 homeologs, making Med15 a candidate target for gene editing. finding
- ★ SuSr-D1 was isolated via flow-sorted chromosome sequencing, EMS mutagenesis, and map-based cloning, narrowing to a 1079 kb interval on chromosome 7D. method
- A Med15 duplication into Med15a and Med15b families (six homologs, three homeologs each) occurred in the Triticeae after divergence from rice and Brachypodium. finding
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| Stem rust seedling infection/phenotyping | Wheat cultivar Canthatch (CTH), Thatcher, mutants NS1/NS2 and CTH-W series | EMS mutagenesis (loss of SuSr-D1); inoculation with Pgt races QTHJC, TTKSK/Ug99, TRTTF | Infection type / sporulation (uredinia) | — |
| Adult-plant field rust resistance test | CTH and CTH mutants, Kenya field | Pgt race TTKSK (Ug99) inoculation | Resistance/disease response | — |
| Flow-sorted chromosome sequencing / de novo assembly | Chromosome 7D from CTH and mutants NS1, NS2, and eight EMS mutants | EMS-induced SNVs | EMS-induced SNVs and chromosome deletions; nonsense mutations in Med15b.D | Flow cytometric sorting |
| Genetic/map-based cloning with DNA markers | NS1 and NS2 doubled-haploid (DH) populations; Chinese Spring reference | none | Recombinants; physical interval (1079 kb) on 7D | — |
| Homeolog-specific RNAseq expression profiling | Wheat first leaf tissue | none | Expression of six Med15 homologs | — |
| RNAseq differential expression (steady-state) | Uninoculated leaves of CTH, NS1, NS2 | med15b.D mutation (NS1, NS2) vs wild-type | Differentially expressed genes (FDR 5%) | — |
| Time-course RNAseq (Pgt and mock inoculated) | Seedlings of CTH, NS1, NS2 | Pgt vs mock inoculation, 0 and 24 hpi | Differentially expressed genes (FDR 5%) | — |
| Branch-specific selection / dN-dS (ω) analysis | Med15, Med15a, Med15b clades across Triticeae and non-Triticeae species | none | Selective pressure (ω) estimates | — |
- ▲ Pgt race QTHJC elicits resistance (no sporulation) on NS1/NS2 mutants vs susceptible large pustules on wild-type CTH.
- – Only one gene in the 1079 kb interval (Med15b.D) carried nonsense point mutations in both NS1 and NS2; six additional mutants had nonsense mutations and two had large deletions.
- – Nonsense mutations span amino acid positions 33 (W06) to 820 (W10/W11) in the 1301-aa Med15b.D protein.
- ▼ At 24 hpi, 8389 genes were differentially expressed in CTH vs 2199 in NS2 and only 55 in NS1, showing severe reduction/loss of transcriptional response in mutants.
- – 229 differentially expressed genes in uninoculated mutant vs wild-type leaves, with 87% localizing to homeologous regions on chromosomes 1A (31 downregulated) and 1D (169 upregulated); 1B unmodified.
- – Med15a and Med15b gene families share 98% and 97% protein identity within families and 90% between families. 90-98%
- – Med15a and Med15b experienced relaxed purifying selection (ωa=0.38, ωb=0.37) relative to non-Triticeae species (ω0=0.24). ω=0.37-0.38 vs 0.24
- – Mutants differed from wild-type in only one of five growth parameters, contrasting with pleiotropic defects of Arabidopsis atmed15 (nrb4-4) null allele.
- count 1079 kb physical interval on chromosome 7D (minimum interval encompassing SuSr-D1)
- count 229 differentially expressed genes (FDR 5%) (uninoculated CTH vs NS1/NS2)
- count 8389 DE genes (CTH), 2199 (NS2), 55 (NS1) (DE between Pgt and mock at 24 hpi)
- count 139, 19, 56 DE genes (Pgt vs mock at 0 hpi in CTH, NS1, NS2)
- other 87% of DE genes on chromosomes 1A and 1D (physical distribution of DE genes)
- other ω0=0.24, ωa=0.38, ωb=0.37 (branch-specific selection on Med15 clades)
- other 98% and 97% (within), 90% (between) protein identity (Med15a/Med15b family protein identity)
- count 10.6 kb gene, 10 exons, 1301 aa protein (Med15b.D gene structure)
Statistical methods review
Model: opusA 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 combines genetic mapping, flow-sorted chromosome sequencing, and mutational analysis to clone SuSr-D1, with the quantitative component relying mainly on RNAseq-based differential gene expression analysis reported at a 5% false discovery rate, plus branch-specific molecular-evolution tests (dN/dS, ω) of selective pressure on Med15 clades. Differentially expressed gene counts are summarized as numbers of genes per comparison and visualized with Venn diagrams, scatter/pair-wise plots, and histograms. RNAseq profiling was performed in triplicate, and morphological growth parameters were compared between wild-type and mutants.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| RNAseq differential gene expression analysis (specific method/tool not stated in text) | CTH vs NS1/NS2 steady-state, mock- vs Pgt-inoculated at 0 and 24 hpi (Figs. 4, 5) | RNAseq profiling performed in triplicate (per Fig. 3b); per-comparison n not otherwise stated | not stated |
| Branch-specific test for variable selective pressure (dN/dS, ω estimation) | Med15, Med15a, and Med15b clades (Supplementary Fig. 12; Supplementary Table 9) | — | not stated |
| Comparison of morphological growth parameters between wild-type and mutants (test not stated) | Five growth parameters (Supplementary Fig. 13) | — | na |
-
Differentially expressed genes were defined using a 5% FDR threshold, with the specific testing framework not named in the text.↳ Could also: A named DE framework such as DESeq2 (Wald test) or edgeR/limma-voom with Benjamini-Hochberg FDR could also be specified. — Naming the model and shrinkage/normalization approach would make the dispersion handling explicit and aid reproducibility, which can be helpful when comparing counts across studies.
-
Results were primarily summarized as counts of differentially expressed genes per comparison.↳ Could also: Reporting effect-size thresholds (e.g., log2 fold-change cutoffs) alongside adjusted p-values, or MA/volcano plots, could also accompany the counts. — Combining a fold-change criterion with the FDR threshold can convey biological magnitude in addition to statistical signal, which some readers find informative for prioritizing genes.
-
RNAseq profiling was performed in triplicate.↳ Could also: Reporting per-gene dispersion estimates or noting power considerations for n=3 could also be included. — With three replicates, summarizing variability (e.g., within-group dispersion) helps readers gauge the stability of DE calls and is often appreciated for small-n designs.
-
Branch-specific selective-pressure tests reported ω (dN/dS) point estimates for the Med15 clades.↳ Could also: Likelihood-ratio tests against null models with reported test statistics and p-values, plus confidence intervals on ω, could also be presented. — Formal model comparison and intervals on ω would quantify the strength of evidence for relaxed purifying selection alongside the point estimates.
-
Morphological growth parameters were compared between wild-type and mutants without a stated test.↳ Could also: A specified test (e.g., t-test or Mann-Whitney U per parameter) with multiplicity adjustment across the five parameters could also be reported. — Stating the test and any correction across the multiple growth traits makes the basis for 'no difference in four of five parameters' transparent.
-
Randomization and blinding for phenotyping/infection assays were not described.↳ Could also: Reporting randomization of layout and blinded scoring of infection types could also be stated where applicable. — Documenting these design elements, when used, helps readers assess potential observer effects in qualitative rust-response scoring.
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.
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.
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-32111840 (Hiebert et al. 2020, Nat Commun)
Paper: "Stem rust resistance in wheat is suppressed by a subunit of the mediator complex." PMID 32111840 / PMC7048732 / DOI 10.1038/s41467-020-14937-2.
Repo (authors' own, P16 own-code): github.com/matthewmoscou/Canthatch @ commit 319c79fb2692fac87e0788beee75d16d9c409d18 (2020-02-03).
Data: SRA PRJNA401266 (flow-sorted chromosome 7DL DNA-seq of Canthatch/NS1/NS2; RNAseq).
The paper combines wet-lab genetics (EMS mutants, flow sorting, KASP mapping, transgenic complementation, rust phenotyping) with bioinformatic pipelines. Only pipeline-derived computational results are in scope.
In scope (pipeline-derived, attempted)
S1 — dN/dS (ω) molecular-evolution analysis of the Med15 gene family ★ PRIMARY
- Pipeline: PAML v4.8
codeml(branch models) on a PRANK codon alignment + RAxML ML tree, both shipped in the repo. Fully self-contained: inputs (.phyalignment, labelled trees,codeml_H*.ctl) AND the authors' own result files (results.H*.txt) are indata/codeml/mutation_rate_Poaceae_b90/(the b90 = ≥90%-coverage tree the paper used). - Determinism: codeml ML optimisation is deterministic given (alignment, tree, ctl, version) → expect near-exact reproduction of lnL and ω, modulo PAML version differences.
- Reported values (Results, "Med15 duplication"): background ω₀ = 0.24, Med15a ωₐ = 0.38, Med15b ωᵦ = 0.37 (all < 1 → relaxed purifying selection, no recurrent positive selection). These map to the shipped 3-ratio H3 model (ω = 0.37702, 0.37155, 0.24407).
- Compare: my codeml lnL + ω for H0,H1,H2a,H2b,H2c,H3 vs (a) the authors' shipped
results.H*.txtand (b) the paper's printed ω₀/ωₐ/ωᵦ; plus the H3-vs-H0 likelihood-ratio test. - Bonus self-contained codeml runs in the same job:
mutation_rate_Poaceae(full),mutation_rate_complete_sequence,mutation_rate_Med15ab_b90— each compared to its shipped result.
S2 — RNAseq differential expression (DESeq2), 229 DEGs (attempt, heavier)
- Pipeline: FastQC→Trimmomatic→kallisto pseudoalignment to IWGSC RefSeq v1.0 AllConf transcripts → DESeq2, FDR 5%, 3×3 design (Canthatch/NS1/NS2, 3 reps). Reads = PRJNA401266.
- Reported value: "A total of 229 differentially expressed genes ... false discovery rate of 5%".
- Requires SRA RNAseq download (9 PE libraries) + a ~10 GB wheat transcriptome index on «infra». Attempted only after S1; the exact DESeq2 contrast/design (which comparison the 229 counts) is under-specified in the README, so this is expected to be partial at best.
Out of scope (not pipeline-reproducible / wet-lab / external)
- EMS mutant generation, chromosome flow sorting, KASP genotyping of mapping populations, transgenic complementation, stem-rust infection phenotyping (all wet-lab).
- RAxML phylogeny topology/bootstrap figures: stochastic (random seeds, 10000 bootstraps); the ML tree is shipped and is an INPUT to S1, so we consume it rather than regenerate it.
- De novo Edena/Trinity assemblies, BBMap/QKgenome haplotype conversion, InterProScan/Phyre2 protein annotation, marcoils/pcoils coiled-coil prediction — large/external-DB/manual.
- Aegilops tauschii / Bromus natural-variation surveys (external accessions, manual curation).
Primary verdict basis
S1 (codeml dN/dS) is a clean, deterministic, self-contained 1:1 target and is the core result. S2 (DESeq2 229 DEGs) is the stretch goal.
Assessments & scoring basis
Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.
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.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
The core S1 Med15 dN/dS analysis reproduced 1:1, bit-identical to 6 decimals across all six codeml branch models and two bonus datasets on the authors' shipped alignment+tree, despite a PAML 4.8->4.10.10 gap — the printed omega 0.38/0.37/0.24 are exact rounds of 0.37702/0.37155/0.24407 and the LRT (16.79, p=2.26e-4) confirms relaxed purifying selection as claimed. No fabrication signal; the only deviation is cosmetic rounding on the authors' side. The single gap is the secondary S2 RNAseq '229 DEGs' number, not attempted purely because its count matrix/design were never deposited — a data-availability limitation, not a methodological or authors' defect, so it does not downgrade the otherwise pristine reproduction.
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-4-8Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.