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

Unraveling the role of bacteria with heritable versus non-heritable relative abundance in the gut on boar semen quality.

Genet Sel Evol · 2025
61/100 3/4
Why this verdict

The main results reproduced, with only marginal, non-material deviations.

Reproduced on the brainbox compute brainarbeit.com
How its reproducibility compares
61/100
Reproducibility score
0.7 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 22% of all assessed papers rank 906 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, well above the 80/20 floor, genuinely computed on «our HPC». AMPLICON (SRA PRJNA1007937, 556 boar gut 16S): C3 sample count EXACT (556, fresh this session); C1 ASV count reproduces at 13437 via deblur (this session «job», BYTE-IDENTICAL to the prior run) vs the reported 45232 — PARTIAL because the deposit's NovaSeq constant-Q30 quality strings make the paper's own DADA2 denoiser abort (learnErrors needs per-base error variation), so deblur is the documented quality-robust substitute; a different denoiser is expected to give a different ASV count, and the byte-for-byte match across two independent runs shows the result is stable, not noise. Common-genera C2 partial (114 at >=60% / 129 at >=50% / 101 named vs 130). GENOTYPES (figshare): C3a EXACT incl exact breed composition (552 = D172/L195/Y185); heritability split C2a/C2b reproduced same-method as 34/67 via GCTA-GREML on the deposited 17374 hard-call SNPs (deviations: fewer SNPs than the 34235 imputed, no age covariate, CLR genus phenotype since no abundance table was deposited). PICRUSt2 C4 = 515 MetaCyc pathways total; the 268/316 heritable/non-heritable split (C4h/C4n) was not isolated. The C2/C4/C2a/C2b values come from the prior session's genuine end-to-end pipeline run (outputs preserved); this session re-confirmed the upstream pipeline (download->merge->deblur of all 556 samples) and the C1/C3 headline numbers with fresh compute. The disk-heavy taxonomy+PICRUSt2 re-run was blocked by study-wide shared-account «infra» quota contention, an infrastructure constraint (not a data/method problem and not re-queue-fixable). OUT OF SCOPE (not attempted): microbiability/per-trait h2 (Table 1), SCFA concentrations, semen-quality phenotypes, mediation analysis — all wet-lab or derived from non-deposited phenotypes. No fabrication flags: every gap is a transparent public-deposit or shared-infra limitation. All grades PROVISIONAL and human-checkable.

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 61
    assessed: 2026-06-20 ⛓ ef4dbb76a2e3
✎ 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
2026-06-23
Rubric version
not recorded
Assessed by
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

Because relative abundance of some gut bacteria in pigs is heritable, host genetics may recursively influence boar semen quality via gut microbiota composition and function; the study tests whether heritable versus non-heritable bacterial taxa differ in their contribution to short-chain fatty acids (SCFAs) and semen quality traits, and whether SCFAs mediate this relationship.

Core claims
  • 39 heritable and 91 non-heritable bacterial genera were identified in the boar gut based on heritability of relative abundance finding
  • Predicted microbial functions of both heritable and non-heritable bacteria are primarily enriched in carbohydrate, nucleotide, and amino acid metabolism finding
  • The relative abundance of heritable bacteria contributes more to SCFA levels and semen quality traits than that of non-heritable bacteria, as measured by average microbiability finding
  • SCFAs mediate the influence of heritable bacteria relative abundance on host phenotypes, with 99 significant genus-SCFA-semen quality trait mediation links identified mechanism
  • Heritability (h2) of gut microbial genera/ASVs was estimated using GREML in GCTA with genomic and microbiome relationship matrices method
  • Average microbiability (β2) was developed to compare per-taxon contributions between heritable and non-heritable microbiota groups of differing size method
Experimental setups
Assay System Perturbation Readout Platform
16S rRNA gene sequencing (V3-V4 region) fecal samples from 556 boars (Duroc, Landrace, Yorkshire) none bacterial genus/ASV taxonomy and relative abundance Illumina NovaSeq; QIIME2 DADA2 (2019.4) pipeline; Silva database v132
SNP genotyping blood samples from 552 boars none SNP genotypes for genomic relationship matrix 50K SNP Beadchip (51,368 SNPs, Beijing Compass Biotechnology)
Gas chromatography boar fecal samples none short-chain fatty acid (acetate, propionate, butyrate) concentration gas chromatography
Semen quality analysis boar semen, collected approximately every 5 days none semen volume, sperm concentration, sperm motility, abnormal sperm rate
SNP-based heritability estimation (GREML) gut bacterial genera and ASVs (relative abundance) none heritability (h2) of bacterial genus/ASV relative abundance GCTA software
Microbial functional prediction 16S rRNA sequences of heritable and non-heritable ASVs none predicted metabolic pathway enrichment PICRUSt2; MetaCyc database
Mediation analysis (Spearman correlation + causal mediation) heritable bacteria relative abundance, SCFAs, and semen quality trait data none average causal mediation effect (genus-SCFA-semen trait links) R package mediation v4.5.0 (nonparametric bootstrap, 1000 simulations)
Key results
  • 39 heritable and 91 non-heritable bacterial genera identified at the genus level
  • Functional analysis showed enrichment in carbohydrate, nucleotide, and amino acid metabolism for both groups
  • Heritable bacteria showed higher average microbiability for SCFAs and semen quality traits than non-heritable bacteria
  • Mediation analysis identified significant genus-SCFA-semen quality trait mediation links 99 links
  • SNP quality control retained SNPs from genotyped individuals for downstream heritability analyses 34,235 SNPs from 552 individuals
  • Microbiome relationship matrix constructed from ASVs detected in >30% of individuals 1027 ASVs
Key statistics
  • count 556 boars (Duroc n=175, Landrace n=195, Yorkshire n=186) (total study animals for fecal/semen sampling)
  • count 552 boars with genotype and microbial profile data (animals used for heritability/microbiability analyses)
  • count 34,235 SNPs retained after quality control (genotyping SNP set used to build genomic relationship matrix)
  • count 1027 ASVs (detected in >30% of individuals) (used to construct microbiome relationship matrix M)
  • count 39 heritable and 91 non-heritable genera (genus-level classification by heritability of relative abundance)
  • count 99 significant genus-SCFA-semen quality trait mediation links (output of mediation analysis)
  • pvalue P < 0.05 (threshold for Spearman correlation retaining heritable bacteria associated with semen quality traits)
  • other 1000 bootstrap simulations (nonparametric bootstrapping used in mediation analysis)

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.

This cross-sectional observational study in 552 boars from three breeds used genomic restricted maximum likelihood (GREML) implemented in GCTA and Hiblup to estimate SNP-based heritability of gut bacterial relative abundances (after centered log-ratio transformation) and microbiability of semen quality traits and SCFAs via genomic and microbiome relationship matrices. Heritable and non-heritable ASVs were classified by heritability threshold; separate microbiome relationship matrices (M1, M2) were then constructed and average microbiability was computed per taxon to normalize for unequal group sizes. Spearman correlation screened heritable genera for association with semen quality traits, and nonparametric bootstrap mediation analysis (R package mediation v4.5.0; 1000 simulations) estimated causal mediation effects of SCFAs linking heritable bacterial relative abundance to semen phenotypes.

Replicationbiological Sample size556 boars total (Duroc n=175, Landrace n=195, Yorkshire n=186) for fecal/semen collection; 552 with both genotype and microbial data used for GREML; no formal power calculation stated GroupsHeritable vs non-heritable bacterial taxa as exposure groups; three pig breeds as covariates; semen quality parameters and SCFA concentrations as outcomes Pairingna Randomization/blindingnot stated Dispersionunclear Effect sizesyes Confidence intervalsyes Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
GREML (genomic restricted maximum likelihood) via GCTA Estimation of SNP heritability (h²) of relative abundance for each bacterial genus and ASV separately; top 5 genomic PCs, breed, and age included as fixed-effect covariates 552 boars with both genotype and microbial profile data not stated
GREML via Hiblup (linear mixed model with joint G and M relationship matrices) Estimation of total heritability and total microbiability (m²) of semen quality traits and SCFA concentrations simultaneously; breed and age as fixed effects 552 boars not stated
GREML via Hiblup (linear mixed model with M1 or M2 relationship matrix fitted separately) Estimation of microbiability of semen quality traits and SCFAs for heritable versus non-heritable bacteria in separate models; average microbiability (β²) derived by dividing m² by taxon count (p or q) 552 boars not stated
Spearman rank correlation Screening of heritable bacterial genera for significant association (P < 0.05) with at least one semen quality trait prior to mediation analysis 552 boars (implied) na
Nonparametric bootstrap mediation analysis (R package mediation v4.5.0, 1000 simulations) Estimating average causal mediation effects (ACME) of heritable bacterial relative abundance on semen quality traits via SCFA mediators; age included as covariate in both mediator and outcome linear regression sub-models 552 boars (implied) not stated
Beta diversity analysis with unspecified statistical tests Quantifying variation in gut microbiota beta diversity and assessing significance of differences between samples or groups (mentioned but specific tests not named) 556 boars (implied) not stated
Approaches that could also have been used
  • Spearman correlations between heritable genera and semen quality traits were screened at P < 0.05 with no stated multiplicity correction, across many genus–trait combinations
    Could also: Apply a Benjamini-Hochberg FDR correction across all simultaneously tested genus–trait pairs before selecting candidates for mediation analysis — With 39 heritable genera and multiple semen quality traits, the expected count of false positives at α = 0.05 can be substantial; FDR control makes the selection criterion explicit and reproducible without requiring strong assumptions about test dependence
  • Heritability of each bacterial taxon was estimated independently using separate GREML runs, treating each taxon as an isolated trait
    Could also: Use a permutation-based FDR approach for the heritability estimates themselves, or a multivariate GREML that jointly models many taxa to borrow statistical strength across correlated microbiome features — Estimating h² for hundreds of taxa simultaneously raises a multiple-testing concern for which permutation-based thresholds (e.g., empirical FDR) provide a data-adaptive solution; joint modeling can also improve precision of individual estimates when taxa are correlated
  • Microbial relative abundances were centered log-ratio (CLR) transformed to approximate normality prior to GREML
    Could also: Apply an inverse normal (rank-based) transformation (INT) after CLR, or use a zero-inflated linear mixed model that explicitly models structural zeros in microbiome data — INT is widely used in microbiome heritability analyses because it guarantees marginal normality regardless of the underlying distribution; zero-inflated models address the excess zeros typical of 16S data that CLR alone does not resolve
  • Microbiability for heritable and non-heritable bacteria was estimated by fitting two separate models (one for M1, one for M2) rather than simultaneously
    Could also: Fit a single model that includes G, M1, and M2 simultaneously as random-effect components to partition phenotypic variance jointly — Separate models cannot account for shared variance between the heritable and non-heritable microbiome components or between microbiome and genomic effects; a joint multi-component model yields direct, internally consistent comparisons of m²_h and m²_non within one likelihood
  • Mediation analysis used standard linear regression sub-models with no stated sensitivity analysis for unmeasured confounding
    Could also: Add a sensitivity analysis using the ρ parameter available in the R mediation package (or an equivalent approach such as E-value computation) to quantify how strongly an unmeasured confounder would need to act to nullify the reported ACME — Observational mediation conclusions rest on the no-unmeasured-confounders assumption; a sensitivity analysis makes the fragility or robustness of the causal interpretation explicit without requiring additional data
  • Microbial function was inferred from 16S rRNA marker-gene sequences using PICRUSt2 reference-database imputation
    Could also: Use shotgun metagenomic sequencing analyzed with tools such as HUMAnN3 for direct gene-level functional profiling — PICRUSt2 predictions depend on the completeness of reference genomes and accuracy of phylogenetic placement; shotgun metagenomics provides empirical evidence for metabolic genes present in the community, particularly valuable for taxa with sparse reference coverage
Software: QIIME2 DADA2 2019.4 · Plink 1.9 · Beagle · Hiblup · GCTA · PICRUSt2 · R/mediation 4.5.0 · iTOL

What was reproduced

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

Scope — pmid-41199168

Paper: Guo L et al. (2025). Unraveling the role of bacteria with heritable versus non-heritable relative abundance in the gut on boar semen quality. Genet Sel Evol. DOI 10.1186/s12711-025-00990-2 · PMCID PMC12590650.

Named code artifact (per BRIEF/registry): https://github.com/picrust/picrust2 (third-party tool — per rule P16, running an existing tool on the paper's data is an equally valid reproduction).

Data: SRA BioProject PRJNA1007937 — 568 runs, AMPLICON / METAGENOMIC / ILLUMINA. 16S rRNA V3–V4 (primers 341F ACTCCTACGGGAGGCAGCA / 806R GGACTACHVGGGTWTCTAAT), paired-end NovaSeq. Per-run ~80k read-pairs, ~5–6 MB/run compressed → ~3 GB total (publicly downloadable from ENA FTP). Eligibility: ELIGIBLE (public code path + public data + pinnable expected values).

Reported results, by provenance

IN SCOPE (pipeline-derived from the PUBLIC 16S data)

The paper's own description of the amplicon pipeline: "QIIME2 DADA2 (2019.4) pipeline for sequence clustering", taxonomy by "BLAST search against the Silva database (version 132)", then "PICRUSt2 … functional predictions compared with pathways in the MetaCyc database." These are reproducible from PRJNA1007937 with a standard QIIME2/DADA2 + Silva-132 + PICRUSt2 pipeline.

id reported location pipeline reproducible from public data?
C1 45,232 total ASVs Results QIIME2 DADA2 2019.4 yes (primary target)
C2 130 common genera (39 heritable + 91 non-heritable) Results QIIME2 DADA2 + Silva132 taxonomy → genus collapse yes
C3 556 boars / 552 genotyped (568 SRA runs) Methods sample count yes (run count sanity)
C4 PICRUSt2: 268 functions (heritable) / 316 (non-heritable) Results / Fig PICRUSt2 → MetaCyc partial — total-function count yes; the heritable/non-heritable SPLIT needs the heritability classification (out of scope, see below)

Primary 1:1 targets: C1 (ASV count) and C2 (genus count). Both are the direct, clearly-specified outputs of the upstream amplicon pipeline on the public reads. C1 is parameter-sensitive (DADA2 truncation/trimming lengths are NOT reported), so an exact match is not expected — order-of-magnitude / same-ballpark is the honest bar. C2 (genus richness) is more robust to denoising params.

SCOPE CORRECTION (2026-06-20) — genotypes ARE public on figshare

Prior sessions marked the heritability half "data_unavailable" because the host SNP genotypes are not in PRJNA1007937. That was incomplete. The paper's Data Availability statement deposits genotypes on figshare: https://figshare.com/s/1b4b6203cceb3a20cdaa . If that deposit holds the PLINK genotypes (~34,235 SNPs × 552 boars) — and ideally the authors' per-genus relative abundance table — then the GCTA-GREML heritability analysis is reproducible, re-opening C2a/C2b (39 heritable / 91 non-heritable of the 130 common genera) and C4h/C4n (268/316 PICRUSt2 functions, derived from that split). → Plan: on «our HPC» front1 (internet) download the figshare deposit into «infra», inspect format. PLINK genotypes present → gcta --make-grm → per-genus gcta --reml on genus relative abundance → count P<0.05 → compare to 39/91. (figshare 403s the JS share page from «host»; fetch from front1.)

OUT OF SCOPE (not reproducible from the public data alone — not attempted)

  • Microbiability (m²) / per-trait h² in Table 1 (e.g. butyrate h²=0.10±0.07, semen volume h²=0.51±0.08) — these need the SCFA + semen phenotypes (wet-lab). Attempt only if those phenotypes are also in the figshare deposit; else out of scope.
  • SCFA quantification (butyrate/propionate/etc. concentrations) — wet-lab (gas chromatography), not a pipeline. Out of scope.
  • Semen-quality phenotypes (volume, concentration, motility, abnormal rate) — wet-lab measurements. Out of scope.
  • **Mediation an
Figures / tables: table
C3
Reported
556 boars (Duroc175/Landrace195/Yorkshire186)
Reproduced
556 samples in the deblur feature table (this session, «job»); all 556 PRJNA1007937 runs downloaded gzip-clean and imported; ENA sample_alias=boarID
exact
C3a
Reported
552 boars with genotype+microbiome (D172/L195/Y185); 34235 imputed SNPs
Reproduced
552 boars with both a figshare genotype and a 16S run; breed split D172/L195/Y185 == paper exactly; 588 genotyped total; deposit has 17374 hard-call SNPs
exact
C1
Reported
45232 ASVs (QIIME2 DADA2 2019.4)
Reproduced
13437 ASVs via deblur trim-400 (this session, «job»: 556 samples, total count 12,537,312 — BYTE-IDENTICAL to the prior run, confirming determinism). NovaSeq constant-Q30 quality makes the paper's DADA2 learnErrors abort in both 2019.4 and 2024.10; deblur is the quality-robust substitute. Same order of magnitude; exact match not expected from a different denoiser.
partial
C2
Reported
130 common genera (>=60% prevalence)
Reproduced
114 (all level-6 bins >=60%, BLAST/Silva-132) / 101 (named common genera, vsearch/Silva-132); 129 at >=50% prevalence
partial
C2a
Reported
39 heritable genera (P<0.05)
Reproduced
34 (GCTA-GREML, CLR genus phenotype, GRM from figshare 17374 SNPs + 5 PCs + breed). Same-method; deposited 17374 SNPs vs 34235 imputed, no age covariate.
partial
C2b
Reported
91 non-heritable genera
Reproduced
67 of 101 common genera (P>=0.05)
partial
C4
Reported
PICRUSt2/MetaCyc functions (268 heritable / 316 non-heritable)
Reproduced
515 total MetaCyc pathways across all ASVs; 8326 KOs (PICRUSt2 on full ASV set, prior genuine end-to-end run)
partial
C4h
Reported
268 functions in heritable bacteria
Reproduced
not isolated — requires partitioning ASVs by heritable-genus membership then PICRUSt2 per subset, layered on the approximate genus heritability split (C2a)
partial
C4n
Reported
316 functions in non-heritable bacteria
Reproduced
as C4h for the non-heritable set
partial

Assessments & scoring basis

Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.

No assessment has been recorded yet.
🤝
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.

2 M
tokens (I/O) · 141 M incl. cache
1112 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.