Microbial Community Shifts Associated With the Ongoing Stony Coral Tissue Loss Disease Outbreak on the Florida Reef Tract.
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 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
- 🟡A deviation arose in the data or preprocessing
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
1:1 REPRODUCED. SCTLD coral 16S microbiome (Meyer et al. 2019, Front Microbiol). The authors' repo ships the trimmed FASTQ + final seqtab.nochim + SILVA taxonomy + metadata + a complete R script reproducing every figure - an exceptionally reproducible deposit (grade A). Reproduced on «our HPC» (R4.5.3/dada2 1.38/phyloseq 1.54/vegan 2.7.5/CoDaSeq 0.99.7). 10 of 12 claims reproduce exact or within-tolerance, including bit-identical headline numbers: total/cleaned ASV & sample counts (C1-C6 exact), PERMANOVA condition R2=0.074 p=0.001 (C8 exact), the reads-vs-PC1 Pearson correlations -0.5887986 / -0.4521599 (C9 exact to 7 decimals), and beta-dispersion Coral & Condition both p<0.001 (C10 exact). The stochastic DADA2 re-run from shipped FASTQ landed at 12299 ASVs vs the published 12300 (1-ASV difference) and reproduced the archaeal ASV count (128) exactly (C12). Only ANCOM differential-abundance family COUNT is a partial (detected_0.7=31 vs reported 21; detected_0.8=20 ~= 21) - the W-statistic threshold in ANCOM v1 is version-sensitive, but every headline disease-enriched family (Vibrionaceae, Rhodobacteraceae, Cryomorphaceae, etc.) is recovered. NOT attempted (out of scope): wet-lab steps, the 5 universal disease-enriched ASVs (per-species intersection), and figure aesthetics. Key version gotcha documented: zCompositions>=1.6 cmultRepl z.delete=TRUE silently drops high-zero samples (must set z.delete=FALSE to match). No fabrication concerns - every reported number is derivable from the shipped data/code.
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 93assessed: 2026-06-21 ⛓ d6470778c842
✎ 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
- 2026-06-21
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-21no 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: sonnetThe study aims to characterize shifts in the coral microbiome associated with stony coral tissue loss disease (SCTLD) lesions versus healthy tissue across four Florida Reef Tract coral species, as an initial step toward identifying the causative pathogen(s).
- ★ Five amplicon sequence variants (unclassified Flavobacteriales, Fusibacter, Planktotalea, Algicola, and Vibrio) were consistently enriched in disease lesions across three coral species (M. cavernosa, D. labyrinthiformis, D. stokesii) finding
- ★ Several groups of likely opportunistic or saprophytic colonizers (Epsilonbacteraeota, Patescibacteria, Clostridiales, Bacteroidetes, Rhodobacterales) were enriched in SCTLD lesions finding
- ★ This represents the first microbiological characterization of SCTLD resource
- ★ Microbial community structure changes significantly with disease condition, but the effect size is small finding
- ★ Healthy coral colonies show lower microbiome variability (dispersion) than diseased or apparently healthy tissue on diseased colonies finding
- An ASV exactly matching Vibrio ishigakensis was found at high relative abundance in D. stokesii disease lesions finding
- Differential abundance of ASVs was determined using DESeq2 comparing lesion samples to tissue farthest from lesions within each coral species method
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| 16S rRNA gene amplicon sequencing (V4 region) | Coral mucus/tissue from M. cavernosa, O. faveolata, D. labyrinthiformis, D. stokesii | natural SCTLD disease state (lesion vs. healthy tissue) | microbial community composition (amplicon sequence variants) | Illumina MiSeq (2x150 bp v.2 cycle) |
| Differential abundance analysis (DESeq2) | Coral tissue samples (lesion vs. tissue farthest from lesion) within M. cavernosa, D. labyrinthiformis, D. stokesii | none/other (naturally diseased vs healthy) | differentially abundant ASVs enriched in disease lesions | DESeq2 v.1.20.0 |
| Differential abundance analysis (ANCOM) of microbial families | Coral tissue across four sample types (lesion, near-lesion healthy, far healthy, undiseased colonies) in four coral species | none/other (disease state comparison) | differentially abundant microbial families, adjusted for coral species | ANCOM |
| Principal component analysis / PERMANOVA / ANOSIM of Aitchison distance | 62 coral microbiome samples across four species | none/other (disease state comparison) | beta diversity / community structure differences by condition and species | vegan (R), CoDaSeq |
| Beta diversity dispersion analysis (betadisper) | M. cavernosa, D. labyrinthiformis, D. stokesii microbiome samples | none/other (health condition comparison) | distance to centroid as measure of microbiome variability | vegan (R) |
- ▲ Five ASVs (unclassified Flavobacteriales, Fusibacter, Planktotalea, Algicola, Vibrio) enriched in disease lesions of all three tested coral species
- ▲ In M. cavernosa, 119 of 121 differentially abundant ASVs were enriched in disease lesions 119/121
- ▲ In D. labyrinthiformis, 124 of 132 differentially abundant ASVs were enriched in disease lesions 124/132
- – In D. stokesii, 69 of 161 differentially abundant ASVs were enriched in disease lesions 69/161
- ▲ 30 ASVs were enriched in disease lesions of at least two coral species 30 ASVs
- – Microbial community structure differed significantly by sample condition but with small effect size PERMANOVA R2=0.074, p=0.001
- – Beta diversity dispersion differed significantly by coral species and by condition p<0.001 (both factors)
- ▲ ASV exactly matching Vibrio ishigakensis found at high relative abundance in D. stokesii disease lesions and lower abundance in apparently healthy tissue
- pvalue PERMANOVA R2 = 0.074, p = 0.001 (microbial community structure vs. sample condition)
- pvalue p < 0.001 (beta diversity dispersion differed by coral species)
- pvalue p < 0.001 (beta diversity dispersion differed by condition)
- count 119 of 121 ASVs enriched (differentially abundant ASVs in M. cavernosa disease lesions)
- count 124 of 132 ASVs enriched (differentially abundant ASVs in D. labyrinthiformis disease lesions)
- count 69 of 161 ASVs enriched (differentially abundant ASVs in D. stokesii disease lesions)
- count 30 ASVs enriched in ≥2 coral species; 5 ASVs enriched in all 3 species (cross-species disease-enriched ASVs)
- count 62 coral samples; average 42,560 reads/sample (range 2,358–233,271) (total microbiome samples sequenced across four coral species)
Statistical methods review
Model: sonnetA 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 used amplicon sequencing (16S rRNA V4 region) processed with DADA2 into amplicon sequence variants (ASVs), followed by compositional-data-aware multivariate analyses (centered log-ratio transformation, Aitchison distance) to compare microbial community structure across coral species and disease states (disease lesion, apparently healthy tissue near/far from lesion, undiseased neighbor). Community-level differences were tested with PERMANOVA and ANOSIM, and dispersion differences with betadisper plus a linear model/ANOVA. Differential abundance of individual ASVs was tested per coral species with DESeq2 (Wald test), and differential abundance of microbial families across the four tissue types was tested with ANCOM, adjusting for coral species. Results were reported primarily as p-values, an R² effect size for PERMANOVA, and ASV/family counts meeting significance thresholds, without confidence intervals or measures of spread such as SD/SEM.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| PERMANOVA (permutational multivariate analysis of variance) on Aitchison distance, 999 permutations, vegan package | overall microbial community structure by coral species and sample condition (Figure 2) | 62 coral microbiome samples | not stated |
| ANOSIM (analysis of similarity) on Aitchison distance, vegan package | overall microbial community structure by coral species and sample condition | 62 coral microbiome samples | not stated |
| Multivariate dispersion (betadisper) fitted to a linear model / ANOVA | beta-diversity dispersion (distance to centroid) by coral species and health condition (Figure 4) | 62 coral microbiome samples | not stated |
| DESeq2 parametric Wald test | differential abundance of individual ASVs between disease lesion and tissue farthest from the lesion, analyzed separately within M. cavernosa, D. labyrinthiformis, and D. stokesii (Table 1, Supplementary Figures S4–S6) | per-species subsets of the 62 samples; ASVs with mean count <5 across samples filtered before testing | not stated |
| ANCOM (ANOVA-based compositional test), significance level 0.05 | differential abundance of microbial families across four tissue types (disease lesion, healthy-near, healthy-far, undiseased neighbor), adjusting for coral species as a covariate | 62 microbiome samples across four coral species; families detected in <70% of samples excluded, families with zero counts in ≥90% of samples removed | not stated |
-
Community-level comparisons used PERMANOVA and ANOSIM on a CLR-transformed Aitchison distance matrix.↳ Could also: Bray-Curtis dissimilarity is another widely used beta-diversity metric for microbiome community comparisons. — Bray-Curtis is a long-standing convention in microbiome ecology and would let the results be compared directly with the many prior coral-microbiome studies that used it, alongside the compositional (Aitchison) approach already applied here.
-
Differences in beta-diversity dispersion were tested by fitting betadisper distances to a linear model and using ANOVA.↳ Could also: A permutation-based test on the betadisper object (e.g., permutest.betadisper or Tukey's HSD on betadisper) could also be used. — A permutation-based approach avoids relying on normality assumptions of ANOVA and is a standard companion test to PERMANOVA/ANOSIM for dispersion effects in ecological data.
-
ASV-level differential abundance was tested per coral species using DESeq2's parametric Wald test on raw counts.↳ Could also: ALDEx2 or ANCOM-BC, which operate directly on CLR-transformed compositional data, could also be applied. — Since the rest of the paper's community analyses already use a CLR/Aitchison compositional framework, a CLR-based differential abundance tool would apply the same compositional-data philosophy consistently across all analyses.
-
Differential abundance of ASVs enriched in disease lesions was assessed separately within each of three coral species, and the five ASVs enriched in all three species were then reported.↳ Could also: A combined or meta-analytic model (e.g., a single DESeq2 model with coral species as a covariate, or a formal meta-analysis combining per-species results) could also be used to test for a shared disease-association signal across species. — Modeling species jointly can increase power to detect ASVs with a consistent cross-species pattern and provides a single overall test statistic and p-value for the shared association.
-
Family-level differential abundance across four tissue types was tested with ANCOM at a fixed ANOVA significance level of 0.05.↳ Could also: ANCOM-BC or a negative-binomial GLM framework (e.g., DESeq2 or edgeR) with an explicit false-discovery-rate correction could also be used for the family-level comparison. — These alternatives provide effect-size estimates (log-fold changes) alongside significance calls and explicit FDR control, which can complement ANCOM's declared/non-declared family classifications.
-
Results are summarized largely through p-values and a PERMANOVA R², without confidence intervals.↳ Could also: Bootstrapped or permutation-based confidence intervals around effect sizes (e.g., around the PERMANOVA R² or DESeq2 log-fold changes) could also be reported. — Confidence intervals convey the precision of an estimated effect in addition to its statistical significance, which can be informative when sample sizes per species/condition are modest.
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-31608047
Paper: Meyer et al. 2019, Microbial Community Shifts Associated With the Ongoing Stony Coral Tissue Loss Disease Outbreak on the Florida Reef Tract. Front Microbiol 10:2244. DOI 10.3389/fmicb.2019.02244.
Data: 16S rRNA V4 amplicon, NCBI SRA PRJNA521988 (also shipped in the GitHub repo as cutadapt-trimmed FASTQ). 65 sequencing samples (62 coral + 3 controls).
Code: https://github.com/meyermicrobiolab/Stony-Coral-Tissue-Loss-Disease-SCTLD-Project
commit 8039ec981235acd99a02bb7b2c915829a455976f. Single R script
FIELDSITES_script_to_reproduce_manuscript_figures.R runs the whole pipeline:
DADA2 (denoise→ASV) → SILVA v132 taxonomy → phyloseq filtering →
CoDaSeq CLR → PCA (Aitchison) → vegan ANOSIM/PERMANOVA/betadisper → ANCOM.
The repo also ships the exact derived tables: Fieldsites_silva_otu_table.txt
(65 × 12300 ASV count matrix = seqtab.nochim), Fieldsites_silva_taxa_table.txt
(SILVA taxonomy), Fieldsites_metadata.txt. This makes the entire downstream
analysis deterministically reproducible from shipped inputs.
In scope (pipeline-derived) — what we attempt
| id | result | pipeline | determinism | source of reported value |
|---|---|---|---|---|
| C1 | total ASVs = 12300 (65 samples) | DADA2 → phyloseq | exact from shipped table | script comment "12300 taxa and 65 samples" |
| C2 | families=597 / orders=339 / kingdoms=4 (raw) | phyloseq get_taxa_unique | exact | script comments |
| C3 | after removing Mitochondria/Chloroplast/Eukaryota/NA → 11332 taxa | phyloseq subset_taxa | exact | script comment "11332 taxa and 65 samples" |
| C4 | families=593 / orders=336 / kingdoms=2 (cleaned) | phyloseq | exact | script comments; paper "out of 593" families |
| C5 | coral samples after removing controls = 62 | phyloseq subset_samples | exact | paper "A total of 62 coral samples"; script "62 samples" |
| C6 | low-abundance filter mean>5 → 693 taxa; genus 1254→279 | phyloseq filter_taxa | exact | script comments |
| C7 | bacterial+archaeal ASVs in coral = 11,189 + 128 = 11,317 | phyloseq | near (coral-only subset) | paper Results |
| C8 | PERMANOVA condition R²=0.074, p=0.001 | vegan adonis on Aitchison dist | stable (perm p) | paper "PERMANOVA R² = 0.074, p = 0.001" |
| C9 | cor(reads, PC1) = −0.589; cor(log reads, PC1) = −0.452 | CoDaSeq CLR→prcomp | exact | script comments |
| C10 | betadisper: condition p<0.001 and coral species p<0.001 | vegan betadisper+anova | stable | paper Results |
| C11 | ANCOM: 21 families significant out of 593 (ANOVA p<0.05) | ANCOM.main (script) | stable | paper "Twenty-one families (out of 593)"; script lists them |
| C12 | DADA2 re-run from shipped FASTQ yields ~12300 ASVs | DADA2 filterAndTrim→dada→mergePairs→removeBimera | STOCHASTIC (authors warn slight variation) | repo seqtab.nochim |
Out of scope (not pipeline-derived / not attempted)
- Wet-lab: DNA extraction, PCR, MiSeq sequencing, field sampling, coral disease scoring.
- Read counts per sample averages (42,560 reads/sample) — descriptive of raw reads; partially checkable via DADA2 read-tracking if re-run.
- The 5 universal disease-enriched ASVs (Flavobacteriales/Fusibacter/Planktotalea/ Algicola/Vibrio) — derived via per-species ASV intersection; attempt if time permits.
- Figure aesthetics (color palettes, bar charts) — not numeric claims.
Notes
- DADA2 ASV inference is stochastic; authors explicitly state "if the dada2 analysis is repeated, it will result in a slightly different ASV table every time." Therefore C1–C7 are graded against the shipped seqtab (exact), and C12 is a separate, looser check of the re-run.
numreads(nonchim) for C9 = rowSums of the shipped OTU table (= seqtab.nochim), blanks removed — fully reconstructable from shipped data.
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.
This is a near-perfect, grade-A reproduction: the authors' repo ships trimmed FASTQ, the final seqtab.nochim, SILVA taxonomy, metadata, and a complete figure-reproducing R script, and 10 of 12 claims reproduce exact or within-tolerance — including bit-identical headline statistics (PERMANOVA condition R2=0.074 p=0.001; reads-vs-PC1 correlations -0.5887986 / -0.4521599 to 7 d.p.). The only deviations are on our/technical side: a stochastic 1-ASV DADA2 difference (12299 vs 12300) and an ANCOM family count of 31/20 vs the reported 21, both driven by R/package-version sensitivity, while every headline disease-enriched family is recovered. The central conclusion — a disease-associated microbial community shift — holds fully, and every reported value is derivable from the shared data with no fabrication concern.
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.