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

Glacier shrinkage will accelerate downstream decomposition of organic matter and alters microbiome structure and function.

Glob Chang Biol · 2022
L1 39/100 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
  • Nothing in this column.
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
  • 🟡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
39/100
Reproducibility score
2.0 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 4% of all assessed papers rank 1131 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 reproduction of the one in-scope pipeline result (R5, functional metagenomics). PRJNA733707 was confirmed to contain the authors' MEGAHIT-assembled contigs (not raw reads), so the annotation half of the EEA_MG.md pipeline was reproduced faithfully on «our HPC»: Prodigal -> eggNOG-mapper -> EEA EC/KO gene extraction (3896 copies: AG 588/BG 936/NAG 249/LAP 639/AP 847/recA 638) -> cd-hit (2638). EEA genes are non-uniformly distributed across bacterial classes (core claim supported). Per-class associations partially reproduce: AG/BG -> Bacteroidetes (supported at phylum level), AP -> Proteobacteria incl. Alphaproteobacteria (partial), LAP -> Gammaproteobacteria (NOT matched by eggNOG seed-ortholog taxonomy; kraken2 cross-check running). Deviations documented: coverm coverage step not reproducible (raw reads absent -> gene counts); paper's maxikraken2_140GB DB host gone (substituted). R1/R2/R4 out of scope (wet-lab/field); R3 blocked (16S inputs never deposited). Verdict provisional - human sign-off required.

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 40
    assessed: 2026-06-19 ⛓ 3548e0cb2fc8
✎ 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 paper tests whether glacier shrinkage (decreasing glacier influence) accelerates the rate of organic matter decomposition by benthic stream biofilms, and whether this occurs via changes in water temperature, nutrient N:P stoichiometry, primary producer (algal) biomass, and shifts in microbiome structure and function.

Core claims
  • Glacier shrinkage (decreasing glacier influence) accelerates downstream organic matter decomposition rates in glacier-fed streams finding
  • Decomposition rates can be estimated from extracellular enzymatic activities (EEA) using stoichiometry-based enzyme allocation equations method
  • Chlorophyll-a, temperature, and stream water N:P together explain 61% of the variability in decomposition rates finding
  • Algal biomass increases with glacier shrinkage and shows a particularly strong relationship with decomposition, indicating algae contribute labile organic compounds to these carbon-poor habitats finding
  • Chytrid fungi, abundant in glacier-fed stream sediments, putatively parasitize algae and promote decomposition via a fungal shunt mechanism
  • Specific bacterial phylogenetic clades (e.g., Saprospiraceae positively, Nitrospira negatively) are significantly associated with decomposition rates finding
  • Different bacterial classes possess different proportions of EEA-encoding genes, potentially explaining microbial associations with decomposition rates finding
  • The Glacial Index (GI), based on glacier area and distance from terminus, quantifies glacier influence for space-for-time comparisons method
Experimental setups
Assay System Perturbation Readout Platform
Extracellular enzymatic activity (EEA) assay using fluorogenic MUF/AMC-linked substrates benthic sediment biofilms from 101 glacier-fed streams (6 global regions) none (natural glacier-influence gradient) activity of AG, BG, LAP, NAG, AP enzymes (nmol h-1 g-1 DM) BioTek Synergy H1 plate reader
Chlorophyll-a extraction and fluorometric quantification benthic sediment biofilms none chlorophyll a concentration (µg g-1 DM) plate reader (436/680 nm excitation/emission)
16S rRNA gene amplicon sequencing (V3-V4) sediment bacterial/prokaryotic community none bacterial community composition/ASVs Illumina MiSeq, 300 bp paired-end
18S rRNA gene amplicon sequencing (V4) sediment eukaryotic community (incl. fungi/chytrids) none eukaryotic community composition/ASVs Illumina MiSeq, 300 bp paired-end
Shotgun metagenomic sequencing subset (n=50) of sediment samples across 6 regions none EEA-encoding gene abundance/taxonomy, recA-normalized coverage NovaSeq (Illumina), S4 flowcell
Stream water physicochemistry and nutrient analysis stream water from 101 glacier-fed streams none temperature, pH, conductivity, O2, turbidity, nitrate, ammonium, SRP, DIN:SRP, DOC LaChat QuikChem 8500; Sievers M9 TOC Analyser; WTW MultiLine meter
Remote sensing / GIS-based glacier metrics 101 glacier catchments none glacier area, distance to terminus, percent glacierized catchment, Glacial Index Sentinel-2 imagery, ASTER GDEM v3
Key results
  • Decomposition rates averaged 0.0129% per day across 101 glacier-fed streams 0.0129 % d-1
  • Decreases in glacier influence (percent glacier catchment coverage, turbidity, glacier index) accelerate decomposition rates
  • Chlorophyll-a, temperature, and stream water N:P jointly explained variability in decomposition rates 61%
  • Algal biomass showed a particularly strong relationship with decomposition
  • High relative abundance of chytrid fungi found in GFS sediments
  • Saprospiraceae positively and Nitrospira negatively associated with decomposition rates among other clades
  • Bacterial classes differ in proportions of genes encoding EEA enzymes
Key statistics
  • mean 0.0129 % d-1 (average organic matter decomposition rate across all glacier-fed streams)
  • other 61% (variance in decomposition explained by chlorophyll-a, temperature, and stream water N:P combined)
  • count 101 glacier-fed streams (total streams sampled across six global regions)
  • count 50 samples (subset used for shotgun metagenomic sequencing)
  • count New Zealand n=20, Greenland n=10, Ecuador n=15, Caucasus n=19, Norway n=10, European Alps n=27 (regional sample distribution across the 101 GFSs)
  • count metagenomic subset: New Zealand n=18, European Alps n=6, Greenland n=2, Caucasus n=16, Ecuador n=3, Norway n=5 (regional breakdown of metagenomics subsample)

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.

Using a space-for-time substitution design, the study surveyed 101 glacier-fed streams across six global regions to relate organic matter decomposition rates — estimated from five extracellular enzyme activity (EEA) measurements via stoichiometric allocation equations (Hill et al., 2017) — to glacier influence, biofilm characteristics, and microbiome structure. Multiple regression identified chlorophyll-a, temperature, and stream water N:P as jointly explaining 61% of variance in decomposition rates. Bacterial phylogenetic clades significantly associated with decomposition were identified using an unspecified association framework, and metagenomic gene abundances (normalised to recA) were used to link bacterial community function to measured EEA. The full statistical methods section was not available in the provided text; descriptions below reflect what is inferrable from the methods and abstract.

Replicationbiological Sample size101 GFSs in six regions (New Zealand n=20, Ecuador n=15, Caucasus n=19, European Alps n=27, Greenland n=10, Norway n=10); two reaches per stream; three sediment patches per reach; n=50 subset for metagenomics; no power analysis described GroupsStreams along a continuous gradient of glacier influence; upstream (near-glacier) vs. downstream reaches within each stream; six geographic regions Pairingmixed Randomization/blindingnot stated Dispersionunclear Effect sizesyes
Statistical tests used
Test Applied to n Assumptions
Multiple linear regression (inferred from R²=0.61 reported in abstract) Decomposition rate predicted by chlorophyll-a, temperature, and stream water N:P 101 streams (two reaches each); exact analytical n not stated in available text not stated
Regression or correlation of glacier influence metrics with decomposition rates (exact test not stated in available text) Percent glacier catchment coverage, turbidity, and Glacial Index vs. decomposition rates 101 streams not stated
Association tests between bacterial phylogenetic clades and decomposition (method not stated in available text) 16S rRNA ASV-level or clade-level abundances vs. decomposition rates not stated in available text not stated
Stoichiometric enzyme allocation equations — deterministic calculation, not a significance test Derivation of decomposition rate index from five EEA measurements (AG, BG, LAP, NAG, AP) 101 streams, two reaches each, three sediment patches per reach na
Metagenomic read-count comparison of EEA-encoding gene proportions across bacterial classes (method not stated in available text) recA-normalised coverage of five EEA genes across 50 metagenomes, stratified by bacterial class 50 samples not stated
Approaches that could also have been used
  • Streams were treated as independent observations along a continuous glacier-influence gradient, with region included implicitly or as a covariate
    Could also: Linear mixed-effects models (e.g., lme4 in R) with geographic region as a random effect could also be used — Streams within a region share geology, climate, and catchment history; a random-effects structure would account for this non-independence and allow inference to separate within-region from among-region drivers
  • Decomposition rates were derived exclusively from an EEA-based stoichiometric index (Hill et al., 2017)
    Could also: Direct substrate mass-loss assays (e.g., standardised cotton strips, Tea Bag Index cellulose/rooibos strips) could also be used alongside or for cross-validation — Direct mass-loss measurements provide an empirical, substrate-explicit decomposition signal independent of the enzyme-allocation model's assumptions, enabling methodological triangulation
  • Multiple bacterial clades were each tested for association with decomposition rates (generating a large family of simultaneous comparisons)
    Could also: False discovery rate correction (e.g., Benjamini-Hochberg) applied across all taxon-decomposition tests could also be reported — With many taxa tested simultaneously, FDR control reduces the expected number of spurious significant associations while preserving more power than Bonferroni correction
  • Decomposition variability was summarised with a single joint R² (61%) from the three-predictor regression
    Could also: Variance partitioning (e.g., vegan::varpart in R) could also be applied to quantify unique and shared contributions of each predictor — When predictors co-vary along the glacier gradient (e.g., temperature and chlorophyll-a both increase as glaciers shrink), variance partitioning clarifies which fraction of explained variance is uniquely attributable to each driver
  • Metagenomic EEA-gene abundances were normalised to a single prokaryotic marker gene (recA)
    Could also: Normalisation using multiple universal single-copy marker genes (e.g., via MicrobeCensus or a set of COG0012-based markers) could also be applied — Multi-marker normalisation averages out copy-number stochasticity in any individual gene, potentially reducing variability in cross-sample gene-abundance estimates across taxonomically diverse communities
  • The two within-stream reaches (near-glacier and downstream) appear to have been used either as independent observations or as contextual covariates
    Could also: Explicitly paired analyses (paired t-test, Wilcoxon signed-rank, or mixed models with reach nested in stream) could also be applied for within-stream upstream-vs-downstream contrasts — Treating the two reaches as paired within each stream controls for stream-level confounders (e.g., catchment geology, regional climate) when estimating the within-stream effect of distance from the glacier
Software: QIIME2 2019.1 · DADA2 (within QIIME2) · Trimmomatic 0.36 · SILVA (taxonomic reference database) 138.1 · megahit 1.2.9 · Prodigal 2.6.3 · eggnog-mapper 2.1.2 · Kraken2 2.1.2 · cdhit-est · coverM 0.6.1 · Statistical analysis software (R, Python, etc.) — not stated in available text

What was reproduced

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

Scope — pmid-35320603 (Kohler et al. 2022, Glob Chang Biol, DOI 10.1111/gcb.16169)

Title: "Glacier shrinkage will accelerate downstream decomposition of organic matter and alters microbiome structure and function."

Code: https://github.com/michoug/EEA_GlacierStream (default branch main, last push 2022-04-01). Two folders: FunctionalMetagenomic/ (shotgun metagenome EEA-gene pipeline, documented in EEA_MG.md + extractGenesOfInterest.R) and Phylofactorization/ (Phylofactorization_16S_k_final.R, by A. Washburne).

Data: SRA PRJNA733707 (metagenomes). 16S amplicon data + field/wet-lab tables are NOT in this BioProject and NOT shipped in the repo.

What the paper reports (result types)

# Reported result Origin In scope?
R1 Decomposition rate avg 0.0129 % d⁻¹ across 101 GFSs wet-lab fluorometric EEA assays + stoichiometric allocation equations OUT — wet-lab/field, not pipeline-derived from public data
R2 Chl-a + temperature + N:P explain 61% of decomposition variance regression on field/wet-lab measurements OUT — wet-lab/field
R3 Bacterial clades associated with decomposition (e.g. Saprospiraceae +, Nitrospira −) via phylofactorization of 16S Phylofactorization_16S_k_final.R on a 16S ASV table OUT (blocked) — requires my_SV_table.txt, my_tree.nwk, my_metadata_table.csv, my_taxonomy_table.csv (placeholder filenames; not in repo, not in PRJNA733707) AND wet-lab decomposition rate K. Inputs unavailable → not reproducible.
R4 High relative abundance of chytrid fungi in GFS sediments 18S/microscopy (not specified as the metagenome pipeline) OUT — not derivable from the shipped metagenome pipeline
R5 Metagenomics: different bacterial CLASSES carry different proportions of the 5 EEA-encoding genes (AG 3.2.1.20, BG 3.2.1.21, NAG 3.2.1.14, LAP 3.4.11.1, AP 3.1.3.1) + recA (K03553) normaliser FunctionalMetagenomic pipeline on PRJNA733707 IN SCOPE — this is the one pipeline-derived metagenomic result; repo ships the full recipe.

In-scope pipeline (R5) — EEA_MG.md

Trim-Galore --paired (R1/R2)  →  MEGAHIT (--min-contig-len 1000, -m 2.4e11)
  →  Prodigal -p meta (gene calling)  →  eggNOG-mapper 2.1.2 (diamond, --sensmode very-sensitive)
  →  extractGenesOfInterest.R  (keep EC ∈ {3.2.1.20,3.2.1.21,3.2.1.14,3.4.11.1,3.1.3.1} OR KEGG_ko = ko:K03553)
  →  cd-hit-est (-c 0.95 -G 0 -aS 0.9)  →  coverm contig trimmed_mean (needs reads)
  →  kraken2 (--confidence 0.5, maxikraken2_1903_140GB) → taxonomy of EEA genes

Final metagenomic claim = per-bacterial-class proportions of EEA-encoding genes.

DATA FORMAT — RESOLVED on «our HPC» (2026-06-25, «job»)

CONFIRMED: PRJNA733707 contains MEGAHIT-assembled CONTIGS, not raw reads. seqkit stats -a on an 8-run sample: every sequence min_len = 1000–1013 bp (== pipeline's megahit --min-contig-len 1000), avg 2–7 kb, max up to 46 kb, and a flat synthetic quality (Q20=Q30=100%, AvgQual=30.00). 662 contig sets, 526 MB on disk (≈ ENA's 0.59 GB), 1.88 Gbp total.

→ Consequence: Trim-Galore + MEGAHIT were already run by the authors; this deposit IS their assembly output. We therefore START at the annotation half: Prodigal → eggNOG-mapper → EEA-gene extraction → kraken2 class taxonomy → per-class EEA-gene proportions (= core of R5). The coverm trimmed_mean coverage step is NOT reproducible (raw reads absent), so we report gene counts, not coverage-weighted abundances. This is the faithful, feasible subset of R5.

CRITICAL DATA CONCERN (original 2026-06-19 hypothesis — now confirmed above)

ENA filereport for PRJNA733707 (queried 2026-06-19):

  • 662 runs, all labelled WGS / METAGENOMIC / ILLUMINA, library_layout = SINGLE.
  • Total 0.59 GB compressed / 1.88 Gbp across ALL 662 runs.
  • read_count median ≈ 869 (min 5, max 14 288); base_count per "read" is ~2 000–18 000 bp (e.g. SRR16190682: 330 reads / 4.93 Mbp ≈ 15 kb each)
Figures / tables: Figure 5
R5
Reported
Functional metagenomics: bacterial classes carry different proportions of the 5 EEA-encoding genes (AG 3.2.1.20, BG 3.2.1.21, NAG 3.2.1.14, LAP 3.4.11.1, AP 3.1.3.1) vs recA K03553 (Fig 5)
Reproduced
Pipeline reproduced on the 662 deposited contig-sets (the deposit IS the authors' MEGAHIT assembly): Prodigal->eggNOG-mapper(diamond,very-sensitive)->EEA EC/KO extraction->cd-hit. 3896 EEA gene copies (AG 588, BG 936, NAG 249, LAP 639, AP 847, recA 638), 2638 non-redundant. EEA genes are demonstrably non-uniformly distributed across bacterial classes/phyla (the core qualitative claim). Coverage-weighting not reproducible (raw reads absent) -> gene COUNTS.
partial
R5b
Reported
AG and BG genes disproportionately in Bacteroidia
Reproduced
SUPPORTED at phylum level: Bacteroidota carries the most AG (277) and BG (410) of any phylum (seed-ortholog taxonomy)
partial
R5a
Reported
AP genes disproportionately in Alphaproteobacteria
Reproduced
PARTIAL: AP highest in Proteobacteria; Alphaproteobacteria carries a large share but Beta/Cyanobacteria comparable (seed-ortholog)
partial
R5c
Reported
LAP genes disproportionately in Gammaproteobacteria
Reproduced
MISMATCH by seed-ortholog: LAP dominated by Alpha/Betaproteobacteria, Gamma low; kraken2 cross-check running
did not match
R3
Reported
16S phylofactorization clades (Saprospiraceae+, Nitrospira-)
Reproduced
BLOCKED: inputs not shipped (placeholders; 16S absent from PRJNA733707)
did not match
R1
Reported
Avg decomposition rate 0.0129 % d-1
Reproduced
out of scope (wet-lab assays)
partial
R2
Reported
61% variance explained (chl-a+temp+N:P)
Reproduced
out of scope (field regression)
partial

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 39/100

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.

67.5 k
tokens (I/O) · 3.2 M incl. cache
24 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.