Reconstruction of 2,965 Microbial Genomes from Mangrove Sediments across Guangxi, China.
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
FINAL. Deposited-genome validation (P16) of the 2965-MAG figshare set (0MDM_drep99_rename.zip) by re-running the paper's own QC/taxonomy tools on «our HPC»; raw-read reassembly was out of scope (TB-scale). RESULTS vs paper: C1 genome count 2965 = EXACT. C3a bacterial MAGs 2383 = EXACT. C3b archaeal MAGs 582 = EXACT. C2b MIMAG-MQ (CheckM2 v1.1.0, comp>=50&cont<10) 2940 vs 2938 = within-tol (+2). C3c phyla 76 vs 78 = within-tol (-2; GTDB-Tk v2.4.0+r226 matching the paper). C2a MIMAG-HQ 18 vs 27 = partial (23S rRNA detected in only 24/332 HQ candidates via Barrnap; tRNA never limiting). So every headline resource claim -- total genomes and the bacteria/archaea domain split -- reproduced EXACTLY, and the quality-tier + phylum counts reproduced within 2; only the MIMAG-HQ rRNA-refined subset diverges, for a well-understood rRNA-detection reason. Key compute lessons (written to kartei): (1) GTDB-Tk 2.4.0 needs numpy=1.23 + pydantic1.10 + py3.11 (numpy>=1.24 removed ndarray.tostring()); (2) extract the ~150G r226 DB to node-local /tmp, NOT «infra»; (3) on «infra» std nodes (192cpu/750G, MaxMemPerCPU=4000) the DB in tmpfs counts against the cgroup, so request the WHOLE node (--cpus-per-task=192 -> 750G) or bacterial pplacer OOMs at the default 250G. compute_ran=true; classify_wf completed cleanly («job», rc=0).
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 58assessed: 2026-06-16 ⛓ c0655217d9f9
✎ 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
- 2026-06-29
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-16no 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 compositional architecture and metabolic potential of microbial communities across different mangrove sediment ecosystems remain poorly characterized, motivating a genome-resolved metagenomic survey of six mangrove sites in Guangxi Province, China.
- ★ A standardized assembly, binning, and dereplication pipeline was used to reconstruct 2,965 non-redundant MAGs from 38 mangrove sediment samples across six sites in Guangxi resource
- ★ The MAG dataset comprises 2,383 bacterial and 582 archaeal genomes spanning 78 microbial phyla finding
- ★ 27 MAGs meet MIMAG high-quality criteria and 2,938 meet MIMAG medium-quality criteria finding
- ★ Bacterial MAGs are dominated by Chloroflexota, Desulfobacterota, Pseudomonadota, Acidobacteriota, Bacteroidota, Planctomycetota, and Nitrospirota, while archaeal MAGs are dominated by Thermoproteota, Thermoplasmatota, Asgardarchaeota, Nanobdellota, and Halobacteriota finding
- MAG phylogeny was reconstructed from concatenation of 41 (30 used downstream) single-copy marker genes with archaea as outgroup method
- RPKM was chosen over TPM for MAG abundance quantification because the study is DNA-only and RPKM/TPM profiles are strongly correlated method
- ★ The dataset provides a genomic resource for studying microbial adaptation and biogeochemical cycling in blue carbon mangrove ecosystems resource
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| whole-genome shotgun metagenomic sequencing | mangrove sediment (6 sites, Guangxi, surface 0-5cm and core up to 90cm) | none | paired-end sequence reads / metagenomic data volume | DNBSEQ-T1/MGISEQ-2000RS |
| physicochemical multiparameter measurement | sediment samples | none | temperature, pH, salinity, moisture | GT-TRJCY soil multiparameter recorder |
| laser particle size analysis | sediment samples | none | particle size distribution | Mastersizer 3000 |
| elemental and isotope ratio analysis | sediment samples | none | TOC, TN, δ13C, δ15N | Vario EL-II CHNOS analyzer; Delta XL Plus IRMS |
| phosphorus fractionation analysis | sediment samples | none | total, inorganic, and organic phosphorus (TP, IP, OP) | — |
| metagenome assembly and genome binning (MAG reconstruction) | sediment metagenomic DNA | none | metagenome-assembled genomes (completeness, contamination) | MEGAHIT, MaxBin2, MetaBAT2, CONCOCT, SemiBin2, Vamb, DAS_Tool, BASALT |
| taxonomic profiling | quality-filtered MAGs | none | taxonomic classification (phylum to species) | GTDB-Tk v2.4.0 |
| phylogenomic tree reconstruction | bacterial and archaeal MAGs | none | maximum-likelihood phylogenomic tree topology | IQ-TREE v3.0.0 |
- – 2,965 non-redundant MAGs reconstructed from 38 mangrove sediment samples
- – MAG set comprises 2,383 bacterial and 582 archaeal genomes spanning 78 phyla
- – 27 MAGs satisfy MIMAG high-quality criteria (>90% completeness, <5% contamination, full rRNA set, ≥18 tRNAs) 27/2965
- – 2,938 MAGs satisfy MIMAG medium-quality criteria (50-90% completeness, <10% contamination) 2938/2965
- – Sequencing generated 10.3 TB of metagenomic data across 38 samples 10.3 TB
- ▲ RPKM and TPM abundance profiles showed strong linear correlation, supporting RPKM use R2=0.98
- – Low-abundance phyla grouped as 'Others' collectively account for a minor fraction of most samples' community ~10%
- – 38 samples consisted of 11 surface (0-5cm) and 27 core (0-90cm) sediment samples from six mangrove regions
- count 2,965 MAGs (total non-redundant metagenome-assembled genomes reconstructed)
- count 2,383 bacterial MAGs (bacterial genome count within MAG dataset)
- count 582 archaeal MAGs (archaeal genome count within MAG dataset)
- count 78 phyla (phylum-level taxonomic diversity spanned by MAGs)
- count 27 high-quality MAGs (MAGs meeting MIMAG high-quality standard)
- count 2,938 medium-quality MAGs (MAGs meeting MIMAG medium-quality standard)
- correlation R2=0.98 (linear correlation between RPKM and TPM abundance metrics)
- other 10.3 TB (total metagenomic sequencing data volume generated across samples)
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.
This is a descriptive metagenomics data descriptor reporting reconstruction of 2,965 non-redundant MAGs from 38 mangrove sediment samples across six sites in Guangxi, China. The analytical approach centers on a standardized bioinformatics pipeline (quality control, assembly, multi-tool binning with ensemble consolidation, dereplication, taxonomic classification, and phylogenomics) rather than formal hypothesis testing. Abundance was quantified as RPKM, with a comparison to TPM summarized as R² = 0.98. Phylogenomic relationships were inferred by maximum likelihood (IQ-TREE v3.0.0) with ultrafast bootstrap support (1000 replicates) on concatenated single-copy marker genes.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| Maximum likelihood phylogenomic inference with ultrafast bootstrap (IQ-TREE v3.0.0, -m MFP -wbt -bb 1000) | Phylogenomic tree of 2,965 bacterial and archaeal MAGs based on concatenation of 41 single-copy marker genes (Fig. 3) | 2,965 MAGs | not stated |
| Linear correlation (R²) | Comparison of RPKM vs. TPM abundance metrics to justify choice of RPKM | — | not stated |
-
Abundance was quantified as RPKM (reads per kilobase per million mapped reads), with TPM noted as an alternative and R² = 0.98 between them reported↳ Could also: Mean coverage depth (mean depth per base across the MAG) or TPM could also be used; alternatively, relative abundance as a proportion of total mapped reads is common in MAG studies — Coverage depth is model-free and widely reported in MAG-centric studies; proportion-based relative abundance facilitates direct comparison across samples of varying sequencing depth; the paper's own R² = 0.98 demonstrates near-identical rank ordering between RPKM and TPM, suggesting either metric would support the same ecological inferences
-
Dereplication was performed at 99% average nucleotide identity (ANI) to define a non-redundant MAG catalog↳ Could also: A 95% ANI threshold, widely adopted as a prokaryotic species-level boundary, could also be used for dereplication — 95% ANI dereplication yields species-level clusters and produces smaller, more broadly comparable catalogs; it is the most common threshold in cross-study MAG collections (e.g., UHGG), which would facilitate integration of this dataset with other mangrove or marine metagenome catalogs
-
Multiple independent binning tools (MaxBin2, MetaBAT2, CONCOCT, SemiBin2, Vamb) were each run with varied parameters and their outputs consolidated via DAS_Tool↳ Could also: A subset of two or three complementary binners (e.g., MetaBAT2 + SemiBin2 + Vamb) consolidated by DAS_Tool, or a single recent deep-learning binner (e.g., GraphMB), could also be used — Benchmarks show that ensemble binning with DAS_Tool recovers more high-quality MAGs than any single binner; however, published comparisons also indicate diminishing returns beyond two to three complementary tools, and reducing the ensemble simplifies parameter-sensitivity documentation
-
Phylogenomic inference used maximum likelihood with model selection (IQ-TREE -m MFP) and ultrafast bootstrap (1000 replicates) on concatenated marker genes↳ Could also: Bayesian phylogenetic inference (e.g., PhyloBayes-MPI with the CAT-GTR model) could also be applied to the same concatenated alignment — Bayesian methods provide posterior probability branch support rather than bootstrap frequencies and the CAT model can better accommodate compositional heterogeneity across deeply divergent prokaryotic lineages, which is relevant when integrating bacterial and archaeal MAGs spanning 78 phyla
-
The RPKM vs. TPM comparison was summarized with a single R² value (0.98) to justify the choice of RPKM↳ Could also: A Bland-Altman agreement analysis, or a Spearman/Pearson correlation with confidence interval, could also characterize metric agreement — R² quantifies shared variance but does not reveal systematic bias or heteroscedastic disagreement at extreme abundance values; Bland-Altman plots are the standard method-comparison approach and would show whether any samples or highly abundant MAGs deviate meaningfully between the two metrics
-
MAG quality thresholds followed MIMAG standards (≥50% completeness, <10% contamination for medium quality); GUNC chimera scores were computed but no explicit GUNC exclusion threshold is stated↳ Could also: An explicit GUNC clade separation score cutoff (e.g., CSS > 0.45) applied as a pre-catalog filter could also be used alongside MIMAG thresholds — GUNC detects chimerism that CheckM2 completeness/contamination metrics do not capture; stating and applying a GUNC exclusion threshold explicitly would allow readers to reproduce the exact catalog boundary and assess what proportion of bins were removed for chimerism
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-41419779
Title: Reconstruction of 2,965 Microbial Genomes from Mangrove Sediments across Guangxi, China
Venue: Scientific Data (2025), DOI 10.1038/s41597-025-06438-y
Type: Data descriptor (MAG resource paper).
Code: https://github.com/SongzeCHEN/MetaGenome-MAG-Analysis (authors' own pipeline, 7 shell scripts).
Raw data: SRA PRJNA1270782 / SRP589204 (38 metagenomes).
Deposited MAGs: figshare 10.6084/m9.figshare.29320385 → one file 0MDM_drep99_rename.zip (2.07 GB, the final
dRep-99%-dereplicated, MDMcleaner-cleaned, renamed non-redundant MAG set). Also ENA PRJEB96880.
Pipeline (from Methods + repo)
fastp v0.19.5 (QC) → BBMap v38.92 (host removal) → MEGAHIT v1.2.9 (assembly, per-sample + co-assembly) →
binning (MetaBAT2, MaxBin2, CONCOCT, SemiBin2, Vamb, BASALT, DAS_Tool) → MDMcleaner v0.8.7 (purify) →
CheckM2 (quality) → dRep v3.4.0 (-comp 50 -con 10 -sa 0.99, 99% ANI dereplication) →
GTDB-Tk v2.4.0 + GTDB release 226 (taxonomy) → (downstream: GUNC, Barrnap, tRNAscan-SE, METABOLIC, IQ-Tree, CoverM).
IN SCOPE (pipeline-derived, reproducible at 80/20 on the DEPOSITED genomes)
The deposited zip is the pipeline's final product. We re-run the validation tools the paper used on it:
- C1 — Genome count = 2,965. Count
.fain the deposited zip. Direct, deterministic. - C2 — CheckM2 quality tiers. Re-run CheckM2 on all deposited MAGs; regenerate the high-quality (completeness >90 %, contamination <5 %) and medium-quality (50–90 %, <10 %) counts → reported 27 HQ / 2,938 MQ. (HQ in the paper additionally requires full rRNA + ≥18 tRNA; the CheckM2 comp/cont part is what we reproduce — the rRNA/tRNA refinement is the optional last 20%.)
- C3 (stretch) — GTDB-Tk taxonomy split. Reproduce 2,383 bacteria / 582 archaea and 78 phyla. Heavy (GTDB r226 DB ~110 GB + pplacer); attempted only if time permits.
OUT OF SCOPE (not attempted, with reason)
- Wet-lab: sediment sampling, DNA extraction, sequencing — non-computational.
- Full de-novo reassembly of all 2,965 MAGs from raw SRA reads (~TBs, weeks of compute) — far beyond 80/20. We validate the deposited genomes (P16: applying the validation tools to the paper's own data is equally valid) rather than regenerating them from raw reads.
- Downstream METABOLIC / IQ-Tree phylogenomics / CoverM abundance — derivative figures, not core resource claims.
Drop check
Repo public ✓, MAGs publicly downloadable from figshare ✓, reported values pinnable (2,965; 27/2,938; 2,383/582; 78 phyla) ✓ → eligible, not a drop.
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 central resource claim — 2,965 non-redundant MAGs — reproduces exactly 1:1 by counting .fa files in the authors' own deposited figshare archive, with a verified sha256 and no fabrication risk. The remaining claims are not deviations but incomplete checks on our side: CheckM2 quality tiers (27 HQ / 2938 MQ) were still running at the finalize cutoff, and taxonomy (2383 bac / 582 arc / 78 phyla) was descoped due to the heavy ~110GB GTDB r226 dependency. The only authors-side note is a minor internal GTDB-Tk 2.4.0 ↔ GTDB r226 version/DB inconsistency, which affects only the not-attempted C3 and caused no observed discrepancy. Overall: a clean exact match on the headline, with secondary descriptors unverified rather than contradicted.
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.