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

A Novel Algicidal Bacterium, Microbulbifer sp. YX04, Triggered Oxidative Damage and Autophagic Cell Death in Phaeocystis globosa, Which Causes Har

Microbiol Spectr · 2022
L1 36/100 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
✓ What held up
  • Same input data as the authors
  • No authors-side cause for any deviation
What did not (or only partly)
  • 🟡Reported values were only indirectly comparable
  • 🟡A deviation arose in the data or preprocessing
  • 🟡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
36/100
Reproducibility score
2.2 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 3% of all assessed papers rank 1138 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, FINAL & HEALTHY. De-novo RNA-seq of Phaeocystis globosa treated with algicidal Microbulbifer sp. YX04 vs 2216E control, 6/12/24h triplicate (18 runs, PRJNA777965). The explicitly-NAMED code (github.com/jstjohn/SeqPrep) was reproduced on «our HPC»: C0 EXACT (18 runs downloaded + md5-verified 36/36, balanced 2x3x3 design); C1 PARTIAL = the named pipeline step (SeqPrep 1.3.2 + SICKLE 1.33 + seqkit on all 18 samples) reproduces the '>95% Q30' claim EXACTLY (mean 96.14%, all 18 >95%) while clean-data volume is ~30% higher than reported (mean 8.87 vs 6.78 Gb/sample; named-tool defaults removed only ~2% of bases, the closed Majorbio cloud filtered harder; min sample 6.787 ~= the reported value -> documentation gap, flagged, not fabrication). The de-novo-assembly stretch (C2 Trinity unigene/transcript counts -> C3 GO, C4/C5 edgeR DEG, C6 KEGG) was fully built and submitted but could NOT be reproduced: the Trinity job was starved by shared-account GrpCPURunMins contention across 3 submissions, and these counts are inherently not 1:1 against the undisclosed Majorbio I-Sanger cloud. Dataset PRJNA777965 profiled (grade A, delivers-promised yes, download+decompress+parse QC PASS). NOT ATTEMPTED (out of scope, wet-lab): algicidal %, ROS/MDA/SOD assays, TEM autophagy imaging, flow cytometry, qRT-PCR, 16S phylogeny.

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 57
    assessed: 2026-06-18 ⛓ 2d3860a30bf7
✎ 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

The study tests whether the newly isolated bacterium Microbulbifer sp. YX04 exerts algicidal activity against the harmful-bloom-forming alga Phaeocystis globosa via secreted metabolites, and investigates the underlying mechanism (oxidative stress, photosynthetic damage, transcriptomic response, and autophagic cell death).

Core claims
  • Microbulbifer sp. YX04 shows high algicidal activity against P. globosa, peaking at 93.2% in the declining growth phase finding
  • YX04 kills algae indirectly by secreting active metabolites into the cell-free supernatant rather than by direct cell-cell contact mechanism
  • The algicidal activity of the YX04 supernatant is stable across a wide range of temperatures (-80 to 100°C) and pH values (4 to 12) for 2 h finding
  • YX04 supernatant induces ROS overproduction, H2O2 accumulation, elevated SOD activity, and MDA accumulation (lipid peroxidation) in P. globosa finding
  • YX04 supernatant damages the photosynthetic system of P. globosa, shown by decreased Fv/Fm and rETR and by downregulation of photosynthesis-related genes finding
  • Transcriptome analysis reveals extensive differential gene expression in P. globosa affecting photosynthesis, respiration, cytoskeleton/microtubule, and autophagosome formation/fusion pathways finding
  • Cellular damage triggers autophagosome formation and large-scale autophagic flux in P. globosa, indicating autophagic cell death mechanism
  • YX04 supernatant is algicidal against multiple other HAB species (Thalassiosira pseudonana, T. weissflogii, Skeletonema costatum, Heterosigma akashiwo, Prorocentrum donghaiense) but not against tested Chlorophyta finding
Experimental setups
Assay System Perturbation Readout Platform
algicidal activity bioassay Phaeocystis globosa culture YX04 bacterial culture / cell-free supernatant / bacterial cells at varying concentrations and times algicidal rate (%)
16S rRNA gene sequencing and phylogenetics strain YX04 none species/genus identification (sequence similarity) PCR/sequencing
scanning electron microscopy (SEM) strain YX04 cells none bacterial cell morphology SEM
thermal and pH stability assay YX04 cell-free supernatant exposure to temperatures -80 to 100°C and pH 4 to 12 for 2 h retention of algicidal activity
DCFH-DA fluorescence assay and H2O2/SOD/MDA biochemical assays P. globosa 5% YX04 supernatant treatment (time course) ROS content, H2O2 content, SOD activity, MDA content fluorescence/biochemical kits
transmission electron microscopy (TEM) P. globosa YX04 supernatant treatment morphological/ultrastructural changes, thylakoid membrane damage, autophagosome formation TEM
pulse-amplitude modulation (PAM) chlorophyll fluorometry P. globosa 5% YX04 supernatant treatment (6, 12, 24 h) maximum photochemical quantum yield (Fv/Fm) and relative electron transfer rate (rETR) PAM fluorometer
transcriptome (RNA-seq) analysis P. globosa 5% YX04 supernatant treatment (6, 12, 24 h) differentially expressed genes and KEGG pathway enrichment
immunofluorescence and immunoblot (Western blot) P. globosa YX04 supernatant treatment autophagosome formation and autophagic flux markers
Key results
  • Algicidal activity of YX04 culture peaked at 93.2% in the declining growth phase 93.2%
  • Bacterial cells alone had no algicidal effect; cell-free supernatant exhibited higher activity than the whole culture
  • Algicidal activity of the supernatant was concentration- and time-dependent, reaching 93-97% with 10% supernatant after 72 h 93-97%
  • ROS (DCFH-DA) fluorescence significantly increased after 6 h of treatment, peaking at 24 h 2.03-fold
  • H2O2 content significantly increased after 24 h of treatment, peaking at 179.9 mmol/g prot 5.07-fold
  • MDA content markedly increased after 12 and 24 h of treatment, indicating lipid peroxidation 7.5-fold (24 h)
  • Fv/Fm decreased progressively during 24 h of YX04 supernatant exposure 0.583 to 0.267
  • rETR markedly decreased after 6 and 12 h of treatment relative to stable control values 65% and 84% decrease
Key statistics
  • other 99.36% 16S rRNA sequence similarity to Microbulbifer maritimus TF-17 (species identification of strain YX04)
  • fold_change 2.03-fold increase in ROS fluorescence at 24 h vs control (oxidative stress in P. globosa after YX04 supernatant treatment)
  • fold_change 5.07-fold increase in H2O2 (179.9 mmol/g prot) at 24 h vs control (H2O2 accumulation in P. globosa)
  • fold_change 1.7-fold increase in SOD activity (281.6 U/mg prot) at 9 h vs control (antioxidant enzyme response in P. globosa)
  • fold_change 2.39-fold (12 h) and 7.5-fold (24 h) increase in MDA content vs control (lipid peroxidation/oxidative damage in P. globosa)
  • count 3,019 / 2,308 / 3,257 genes upregulated after 6 / 12 / 24 h treatment; 692 upregulated at all time points (transcriptome DEG analysis of P. globosa)
  • count 3,273 / 3,262 / 3,629 genes downregulated after 6 / 12 / 24 h treatment; 2,005 downregulated at all time points (transcriptome DEG analysis of P. globosa)
  • other Fv/Fm decreased from 0.583 to 0.267 over 24 h; rETR decreased by 65% and 84% at 6 and 12 h (photosynthetic system damage in P. globosa)

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 employs a repeated-measures experimental design with three biological replicates per condition, comparing a YX04 supernatant treatment group to an untreated control across multiple time points (6, 12, and 24 h) and concentrations (1–10%). Biochemical and photosynthetic endpoints are reported with standard error bars and categorical significance symbols; transcriptome DEGs are identified using an adjusted P-value threshold (< 0.05) combined with a |log2 fold-change| filter. Descriptive fold-change values are provided for key biochemical results in the text, but the underlying statistical tests are not named.

Replicationbiological Sample sizeThree biological replicates stated throughout; no formal power analysis or sample-size justification described GroupsYX04 supernatant treatment vs. untreated control at multiple time points (6, 12, 24 h) and concentrations (1%, 3%, 5%, 10%); also different YX04 growth-stage fractions (Fig. 2a) Pairingunpaired Randomization/blindingnot stated DispersionSEM Exact p-valuesno Effect sizesno Confidence intervalsno Multiplicity correctionAdjusted P-value applied for transcriptome DEGs (correction method not named); no correction stated for biochemical or physiological assays
Statistical tests used
Test Applied to n Assumptions
Not named (pairwise comparisons annotated with *, P < 0.05; **, P < 0.01; ***, P < 0.001) Algicidal activity across growth stages (Fig. 1a); algicidal rate at different concentrations (Fig. 2b); ROS, H2O2, SOD, and MDA levels (Fig. 3); Fv/Fm and rETR photosynthetic parameters (Fig. 5a, 5b) 3 biological replicates not stated
Differential expression analysis with adjusted P-value (adjustment method not named) Transcriptome DEG identification at 6, 12, and 24 h treatment (Figs. 4, 5c, 6); threshold: adjusted P < 0.05 and |log2FC| threshold applied 3 biological replicates per time point (implied from study design) not stated
KEGG pathway enrichment analysis (statistical method not named) Upregulated and downregulated DEG sets at each time point (Figs. 4c, 4f) not stated
Approaches that could also have been used
  • Error bars throughout represent the standard error of the mean (SEM) with n = 3 biological replicates
    Could also: Standard deviation (SD) or 95% confidence intervals could also be used to summarize data spread — With only n = 3 replicates, SD conveys the actual variability among observations; 95% CIs additionally communicate inferential uncertainty in a form directly interpretable without significance thresholds
  • Multiple pairwise comparisons between treatment and control at several time points are reported with significance symbols, without a named statistical test or multiple-comparison correction
    Could also: A one-way or two-way ANOVA followed by a planned post-hoc test (e.g., Dunnett's test against a common control, or Tukey HSD) could also be applied — An omnibus test followed by a post-hoc correction controls the family-wise error rate across all time-point comparisons, which is a standard approach when one control is compared to several treatment levels simultaneously
  • The specific statistical test used for biochemical and photosynthetic comparisons is not named in the text
    Could also: Reporting the test name (e.g., Student's t-test, Mann-Whitney U, one-way ANOVA) along with the test statistic and degrees of freedom would also be standard practice — Naming the test and providing its statistic (t, F, U) alongside P-values allows readers to assess whether assumptions were met and to independently verify reported significance levels
  • For transcriptome DEGs, an adjusted P-value < 0.05 is used as a threshold but the multiple-testing correction method is not named
    Could also: Explicitly naming the correction method (e.g., Benjamini-Hochberg FDR) and the DEG-calling software (e.g., DESeq2, edgeR) with version numbers could also be standard — Naming the FDR method and software version enables reproduction of the DEG list and allows readers to understand the stringency of control applied across tens of thousands of simultaneous gene-level tests
  • Significance is communicated only via categorical symbols (*, **, ***) rather than exact P-values
    Could also: Reporting exact P-values (e.g., P = 0.003) alongside or in place of symbols is also common practice — Exact P-values allow readers to judge evidential strength more precisely and are increasingly expected by journals for transparent and reproducible reporting
  • No formal sample-size or power justification is described; n = 3 biological replicates is used for all assays
    Could also: An a priori power analysis based on an anticipated effect size and target power (e.g., 80%) could also be reported — Stating the basis for the chosen n helps readers assess whether the design was adequately powered to detect biologically meaningful differences, particularly when effect sizes are moderate and variability is unknown
Software: Not stated

What was reproduced

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

Scope — pmid-35019679

Paper: Zhu et al. 2022, Microbiol Spectr 10:e00934-21. "A Novel Algicidal Bacterium, Microbulbifer sp. YX04, Triggered Oxidative Damage and Autophagic Cell Death in Phaeocystis globosa." DOI: 10.1128/spectrum.00934-21 · PMID 35019679 · PMCID PMC8754136

Named code: https://github.com/jstjohn/SeqPrep — a third-party paired-end read adapter-trimming + merging tool. Per BRIEF rule P16, running this existing tool on the paper's own data is a fully valid reproduction.

Data: SRA BioProject PRJNA777965 — 18 paired-end Illumina HiSeq 4000 RNA-seq runs (SRR16796923–SRR16796940), Phaeocystis globosa transcriptome. Design = 2 groups × 3 timepoints × 3 reps:

  • Control "2216E" (sterile-medium control): C6_{1,2,3}, C12_{1,2,3}, C24_{1,2,3}
  • Treatment "YX04" (bacterium): T6_{1,2,3}, T12_{1,2,3}, T24_{1,2,3}

Reported pipeline (Methods + Majorbio I-Sanger Cloud)

  1. QC/trim: SeqPrep + SICKLE remove low-quality reads → "clean data".
  2. Assembly: Trinity (de novo, no reference genome) → unigenes/transcripts.
  3. DEG: edgeR + DESeq2, threshold P-adj < 0.05 and |log2FC| ≥ 1.
  4. Annotation/enrichment: BLASTx vs NCBI nr / STRING / SwissProt; GO + KEGG; GOseq + KOBAS, P < 0.05.

In scope (pipeline-derived → attempt)

id reported result location pipeline difficulty
C1 ~6.78 Gb clean data per sample, >95% Q30 Results/Methods SeqPrep+SICKLE (named code) LOW — MINIMUM core
C2 69,230 unigenes; 94,211 transcripts Results Trinity HIGH (de-novo, version-sensitive)
C3 11,312 unigenes GO-annotated Results BLAST+GO HIGH
C4 DEG up: 3019/2308/3257 (6/12/24h), 692 shared Fig 4 quant+edgeR/DESeq2 HIGH
C5 DEG down: 3273/3262/3629 (6/12/24h), 2005 shared Fig 4 quant+edgeR/DESeq2 HIGH
C6 60 KEGG pathways enriched (up); 85 (down) Results GOseq/KOBAS HIGH
C0 dataset N: 18 runs present & parseable SRA (profiling) LOW

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

  • Algicidal activity %, growth inhibition curves, cell counts (microscopy).
  • ROS/MDA/SOD/antioxidant enzyme assays (biochemistry).
  • TEM autophagosome imaging; flow cytometry; qRT-PCR validation (wet-lab).
  • 16S phylogeny / bacterium isolation & identification (wet-lab + manual).

Strategy

  • MINIMUM (quick floor): download 18 runs → profile → run SeqPrep+SICKLE (the exact named GitHub tool) → reproduce clean-read/Q30/clean-Gb (C1). This is a true 1:1 reproduction of the one explicitly named pipeline step.
  • Then keep going (harder): Trinity assembly (C2/C3), quant+edgeR DEG (C4/C5), enrichment (C6). De-novo assembly counts are inherently version/parameter sensitive (Majorbio cloud params not fully disclosed) — treat exact-count agreement cautiously; ballpark concordance is the honest target.

Caveats / fabrication-watch

  • Authors ran analysis on the closed Majorbio I-Sanger cloud, so exact versions/params are not in the paper → exact de-novo counts may not be 1:1 reproducible even with correct data; that is a documentation gap, not fabrication.
  • Paper cites "BioSample SAMN22896686–22896721" (36) but only 18 even-numbered biosamples carry these RNA runs — to verify during profiling.
Figures / tables: Figure 4
C0
Reported
18 paired RNA-seq runs SRR16796923-16796940; 2grp(2216E/YX04) x 3time(6/12/24h) x 3rep
Reproduced
18 runs downloaded from ENA + md5-verified (36/36); balance exactly even (3 per cell)
exact
C1
Reported
~6.78 Gb clean data/sample; >95% Q30 after SeqPrep+SICKLE
Reproduced
Q30 mean 96.14% (all 18 >95%) EXACT; clean Gb mean 8.870 (range 6.79-11.73) vs 6.78 = ~30% higher retention
partial
C2
Reported
69230 unigenes; 94211 transcripts (Trinity de novo)
Reproduced
NOT RUN - Trinity job built+submitted 3x, blocked by shared-account GrpCPURunMins contention; also version-sensitive vs closed Majorbio cloud
did not match
C4
Reported
up DEGs 3019/2308/3257 (6/12/24h), 692 shared
Reproduced
NOT RUN - downstream of blocked C2 assembly (Salmon+edgeR staged)
did not match
C5
Reported
down DEGs 3273/3262/3629 (6/12/24h), 2005 shared
Reproduced
NOT RUN - downstream of blocked C2 assembly (Salmon+edgeR staged)
did not match

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 36/100

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.

47.6 k
tokens (I/O) · 2.4 M incl. cache
8 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.