Endogenous Fatty Acid Synthesis Drives Brown Adipose Tissue Involution
Part of the results reproduced; minor but material deviations remained.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- Nothing in this column.
- 🔴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
This paper has a computational component, but its primary data is legally or ethically access-restricted — identifiable patient cohorts, rare-disease genomes, or controlled-access biobanks that cannot be openly shared. The reproduction therefore could not be attempted. That is a neutral verdict: it does not mean the result is wrong or that the authors fell short — only that, for legitimate privacy reasons, it cannot be independently checked from public data. We deliberately do NOT assign a 0–100 score here, because a low number would wrongly read as a failed reproduction.
▸Reproduction agent’s raw note
DROP / data_restricted. Schlein et al. 2021 (Cell Reports 34:108624) is overwhelmingly a wet-lab / in-vivo physiology + targeted-lipidomics paper (ChREBP/MLXIPL-KO mice, histology, TaqMan qPCR, TAG/phospholipid/cardiolipin lipidomics, tracer uptake, respirometry) with NO shipped code and NO data accession ('the published article includes all datasets'). The ONLY pipeline-derived computational result is Fig 1G-J: Pearson correlations of human BAT log(MLXIPL) vs log(ACLY/ACACA/FASN/SCD), n=31 (R=0.36/0.55/0.60/0.48; P=0.045/0.0013/0.0004/0.0055 — pinned exactly from the figure). Its input RNA-seq is REUSED from Jespersen et al. 2020 (bioRxiv 2020.05.07.082057, corr. C. Scheele), which is NOT publicly deposited: no GEO/SRA/ENA/ArrayExpress/BioStudies accession exists, the source is a cc_no preprint with no peer-reviewed published version (published=NA in 2026), so the expression matrix is obtainable only on request -> a true from-raw-data reproduction is impossible. As a good-faith substitute I digitized the four scatter panels (the only data the paper ships) and recomputed R/P: qualitatively consistent (all positive; H and I near-exact at 0.54 and 0.40-0.60), but quantitatively UNRELIABLE (overlapping ~8px open circles on a 150-DPI JPG + drawn regression lines make recovered R parameter-unstable: G 0.70-0.85, J 0.0-0.36) -> can neither confirm nor refute the exact values; NO fabrication signal. NOT attempted: all mouse/lipidomics/qPCR/tracer/respirometry results (wet-lab, out of scope). «our HPC» reachability verified but no job needed (no public dataset to process).
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 assessmentassessed: 2026-06-18 ⛓ 4dcbd3db7d7b
✎ 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
- 2026-06-18
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-18no 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 paper tests whether endogenous fatty acid synthesis (de novo lipogenesis), transcriptionally controlled by ChREBP, is the molecular driver of brown adipose tissue (BAT) whitening/involution during adaptation to thermoneutrality.
- ★ Endogenous fatty acid synthesis (DNL), regulated by ChREBP, is the central driver of BAT whitening/involution at thermoneutrality. finding
- ★ ChREBP deficiency preserves BAT mitochondrial mass and thermogenic capacity during thermoneutral adaptation, independently of mitochondrial biogenesis. finding
- ★ Pharmacological inhibition of lipogenesis (ACC inhibitor CP-640186) mimics ChREBP deficiency, reducing lipid accumulation and preserving UCP1/thermogenic capacity. finding
- ★ ChREBP controls the abundance and composition of both storage and membrane lipids (including cardiolipins), which regulate organelle turnover and function. mechanism
- ★ Mitochondrial breakdown via autophagy, rather than reduced biogenesis, underlies BAT whitening and is ChREBP-dependent. mechanism
- ★ MLXIPL (ChREBP) expression correlates with lipogenic gene expression (ACLY, ACACA, FASN, SCD) in human BAT. finding
- ★ Reduced BAT lipid accumulation in ChREBP-KO mice is not caused by reduced dietary lipid uptake; triolein uptake is actually increased. finding
- Total ChREBP-KO mice express a truncated, DNA-binding-deficient Chrebp-alpha transcript, used as the genetic loss-of-function model throughout the study. resource
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| Fatty acid profiling (lipogenic index, 16:1/18:2 ratio) | BAT (TAG and phospholipid fractions), wild-type and ChREBP-KO mice | thermoneutral housing (30°C) vs room temperature (22°C) | ratio of palmitoleic to linoleic acid as marker of endogenous vs dietary fatty acid contribution | gas chromatography |
| Lipidomics | BAT (TAG, DAG, CE, phospholipids, cardiolipins, mitochondrial lipid extracts), wild-type and ChREBP-KO mice | thermoneutral housing (30°C) vs 22°C | lipid class concentrations and fatty acyl chain composition (endogenous C14-C16 vs dietary 18:2 enrichment) | — |
| qPCR gene expression | BAT, wild-type and ChREBP-KO mice (also BAT-specific insulin receptor KO mice) | ChREBP knockout; thermoneutrality; fasting-refeeding; insulin receptor knockout | mRNA levels of lipogenic (Chrebp isoforms, Acly, Acaca, Fasn, Scd1, Elovl6, Srebf1/2) and thermogenic/mitochondrial genes (Ucp1, Dio2, Ppargc1a, Tfam, Cox4i1) | — |
| RNA-seq correlation analysis | human BAT biopsies (lean, overweight, obese subjects; published dataset) | none | correlation of MLXIPL expression with ACLY, ACACA, FASN, and SCD expression | RNA-seq |
| Histology and immunostaining (H&E, UCP1) | BAT, wild-type and ChREBP-KO mice | thermoneutral housing (30°C) | lipid droplet accumulation and UCP1 staining intensity | — |
| Western blot | BAT, wild-type and ChREBP-KO mice | thermoneutral housing | UCP1 and mitochondrial OXPHOS complex protein levels | — |
| Electron microscopy | BAT, wild-type and ChREBP-KO mice | prolonged thermoneutral acclimation (12 weeks) | mitochondrial number and ultrastructure | transmission electron microscopy |
| LC3 immunostaining and AAV-LC3 fluorescent reporter imaging | BAT, wild-type and ChREBP-KO mice | thermoneutral housing | number of LC3 punctae/mature autophagosomes | — |
- ▲ Lipogenic index increased in both TAGs and phospholipids of BAT after 1 or 4 weeks of thermoneutral housing
- ▼ ChREBP-KO mice showed a strongly reduced lipogenesis index in both TAGs and phospholipids after 12 weeks at thermoneutrality compared with wild-type
- ▲ BAT of ChREBP-KO mice showed higher 14C-triolein uptake than wild-type at both 22°C and 30°C, while glucose uptake was reduced 2-fold
- ▲ UCP1 and mitochondrial OXPHOS complex protein levels were higher in thermoneutral BAT of ChREBP-KO mice than wild-type
- ▲ Electron microscopy showed higher mitochondrial numbers in BAT of ChREBP-KO mice after prolonged thermoneutral acclimation
- ▲ Isolated BAT mitochondria from ChREBP-KO mice showed higher maximal (CCCP-induced) and UCP1-dependent (GDP-sensitive) respiration than wild-type
- – ACC inhibitor CP-640186 reduced BAT lipid accumulation and partially preserved UCP1 protein levels during thermoneutral adaptation
- ▼ Wild-type BAT showed significantly more LC3 punctae (autophagosomes) than ChREBP-KO BAT at thermoneutrality
- fold_change 2-fold (Absolute Chrebp-alpha mRNA expression compared with Chrebp-beta isoform in BAT)
- fold_change 2-fold higher (14C-triolein uptake into BAT of ChREBP-KO vs wild-type mice at 22°C and 30°C)
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 combines a genetic loss-of-function model (ChREBP-KO vs wild-type littermate mice) with pharmacological intervention (ACC inhibitor CP-640186) and housing at 22°C vs 30°C to investigate de novo lipogenesis in brown adipose tissue involution. Readouts include gene expression, radiotracer uptake, lipidomics, histological quantification (LC3 punctae, UCP1 immunostaining, electron microscopy), and mitochondrial respiration. A published human BAT RNA-seq cohort (Jespersen et al., 2020) is used to assess correlations between ChREBP/MLXIPL and lipogenic gene expression. The statistical methods section is not included in the provided text excerpt, so test names, sample sizes, and correction procedures are largely unstated.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| correlation (type not stated; Pearson or Spearman inferred from context) | MLXIPL vs ACLY, ACACA, FASN, SCD in human BAT RNA-seq dataset (Figures 1G–1J) | — | not stated |
| correlation (type not stated) | glucose vs lipid uptake in WT and ChREBP-KO BAT (Figure 2C) | — | not stated |
| test not stated | LC3 punctae counts between WT and ChREBP-KO at thermoneutrality (Figures 3D–3G); described as 'significantly more' in WT | — | not stated |
| test not stated | qPCR gene expression comparisons between genotypes and/or temperatures (Figures 1D, 3A, S1A–S1D); 'not significantly affected' noted for Srebf1/Srebf2/Srebp1c | — | not stated |
| test not stated | lipogenic index (16:1/18:2 ratio) in TAGs and phospholipids across temperature and genotype (Figures 1B, 1C, 1E, 1F) | — | not stated |
| test not stated | mitochondrial respiration (CCCP-induced maximal; GDP-sensitive UCP1-dependent) between genotypes (Figures 1N, S1G); radiotracer uptake across tissues (Figures 2A, 2B) | — | not stated |
-
Genotype and temperature effects appear to be evaluated with pairwise comparisons based on the results language↳ Could also: A two-way ANOVA (genotype × temperature) with a post-hoc correction (e.g., Tukey HSD or Sidak) could also be used for each outcome — A factorial ANOVA explicitly estimates and tests the genotype × temperature interaction term, which directly addresses whether thermoneutrality affects WT and KO mice differently — the central biological question — while controlling the family-wise error rate across pairwise comparisons within the same model
-
Correlations between MLXIPL and individual lipogenic genes were computed in a human BAT cohort that spans lean, overweight, and obese subjects↳ Could also: Spearman rank correlation (as an alternative to Pearson) and/or partial correlation controlling for BMI or adiposity index could also be applied — If the association is nonlinear or if BMI independently drives both MLXIPL expression and lipogenic gene expression, rank-based or partial correlation would provide estimates less confounded by the BMI gradient inherent in the cohort design
-
LC3 punctae per cell were counted as a discrete, per-image outcome and compared between genotypes↳ Could also: A non-parametric test (e.g., Mann-Whitney U) or a negative-binomial model for count data could also be used, ideally with blinded image analysis — Punctae counts are non-negative integers that are often right-skewed; non-parametric or count-specific models make fewer distributional assumptions than tests designed for continuous Gaussian data, and blinded quantification reduces observer-dependent variability
-
A broad lipidomic dataset covering many lipid classes and species was compared across genotype and temperature conditions↳ Could also: Multivariate dimensionality reduction (e.g., PCA or PLS-DA) combined with per-species FDR correction (e.g., Benjamini-Hochberg) could also be applied — High-dimensional lipidomic data benefit from multivariate visualization to confirm group separation before univariate testing, and FDR correction explicitly controls the expected proportion of false discoveries when hundreds of lipid species are tested simultaneously
-
Gene expression across a panel of lipogenic and thermogenic markers was compared between groups without a stated multiple-testing correction↳ Could also: Benjamini-Hochberg FDR correction applied across all genes tested within an experiment could also be used — Testing many genes simultaneously inflates the chance of false positives; FDR correction quantifies and bounds the expected proportion of spurious findings among those declared significant, which is useful context when interpreting panels of related targets
-
Radiotracer uptake (14C-triolein and 3H-deoxyglucose) was measured in multiple tissues from the same animals and compared across genotypes↳ Could also: A linear mixed model with tissue as a repeated factor nested within animal could also be used — Uptake values from different tissues of the same animal are correlated; a mixed model explicitly accounts for within-animal dependence, using the data structure more accurately than treating each tissue-by-animal combination as an independent observation
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-33440156
Title: Endogenous Fatty Acid Synthesis Drives Brown Adipose Tissue Involution Authors: Schlein C, Fischer AW, Sass F, … Scheja L. (corresponding: Heeren J / Scheja L, UKE Hamburg) Journal: Cell Reports 2021; 34(2):108624 · DOI 10.1016/j.celrep.2020.108624 · PMCID PMC8240962 (CC BY-NC-ND, open access)
Nature of the paper
This is overwhelmingly a wet-lab / in-vivo physiology + lipidomics study of brown adipose tissue (BAT) "whitening" / involution, driven by ChREBP-dependent de-novo lipogenesis. The bulk of the evidence is:
- mouse models (wild-type vs ChREBP/MLXIPL-KO; thermoneutral 30 °C vs cold/22 °C housing),
- histology + UCP1 immunostaining, electron microscopy, western blots,
- TaqMan qPCR gene-expression (normalised to Tbp),
- targeted lipidomics (TAG / phospholipid / cardiolipin species; lipogenesis indices 16:1/18:2) — performed on a proprietary platform (BERG; Narain/Bussberg/Kiebish),
- radiotracer organ-uptake (¹⁴C-triolein, ³H-deoxyglucose), respirometry,
- a pharmacological ACC-inhibitor (ACCi) experiment.
All of the above are out of scope for computational reproduction: they are wet-lab / manual / instrument-derived and cannot be re-run from any shipped code or public data (no code is provided; data-availability statement is "The published article includes all datasets generated or analyzed during this study", Lead Contact L. Scheja). No GitHub/GitLab repo, no GEO/SRA/ENA/ArrayExpress/PRIDE accession is given anywhere in the article (verified against the full text via the PMC BioC API).
In-scope (pipeline-derived) result — the ONLY one
Figure 1G–J: Pearson correlation of human BAT gene expression
log(MLXIPL) (ChREBP) vs the lipogenic genes log(ACLY), log(ACACA), log(FASN),
log(SCD), across n = 31 human supraclavicular deep-neck BAT subjects.
Reported values (read directly from the panels of Fig 1, PMC8240962):
| panel | x-gene | reported R | reported P |
|---|---|---|---|
| G | ACLY | 0.36 | 0.045 |
| H | ACACA | 0.55 | 0.0013 |
| I | FASN | 0.60 | 0.0004 |
| J | SCD | 0.48 | 0.0055 |
Pipeline = (RNA-seq expression matrix) → per-gene log expression → pairwise Pearson correlation + 2-sided p-value, n = 31. (The figure axes are log10-transformed and the regression line is linear in log-log, so R is Pearson on the log-expression values.)
Input data for this result is REUSED, not generated here
The 31-subject human BAT RNA-seq is not original to this paper. The Results/Methods state it is "an already established RNA sequencing dataset … as published" — the cited source is:
Jespersen NZ, Andersen MW, Jensen VH, Stærker TW, Severinsen MC, Peijs L, Soares R, Forss I, Andersen ES, Hahn CH, Homøe P, Mandrup S, Pedersen BK, Nielsen S, Scheele C. "Thermogenic genes are blunted whereas brown adipose tissue identity is preserved in human obesity." bioRxiv 2020.05.07.082057 (2020), DOI 10.1101/2020.05.07.082057. Corresponding author Camilla Scheele (Univ. Copenhagen). License:
cc_no(all-rights-reserved). Cohort: 27 subjects + 4 with T2DM = 31 for the transcriptomic analysis (matches the n = 31 used here).
Reproducibility blocker (controlling fact)
The Jespersen 2020 RNA-seq is not publicly obtainable:
- No GEO / SRA / ENA / ArrayExpress / BioStudies accession could be found (searched NCBI GDS+SRA E-utilities, EBI ENA portal, ArrayExpress, BioStudies; the target paper itself deposits nothing).
- The source is a bioRxiv preprint under a
cc_nolicense; CrossRef + the bioRxiv API show no peer-reviewed published version even as of 2026 (published: NA), so no journal data-availability requirement ever forced a deposit. - Therefore the expression matrix needed to recompute the Fig 1G–J correlations is only
available on request from the Scheele lab → drop_reason
data_restricted.
What was done instead (good-faith partial)
Because the paper states it "includes all d
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 wet-lab/in-vivo physiology paper whose only pipeline-derived result is Fig 1G–J (Pearson correlations of human BAT log(MLXIPL) vs ACLY/ACACA/FASN/SCD, n=31). Its input RNA-seq is reused from an undeposited preprint with no public accession, so a from-raw-data reproduction is impossible — the gap is on the data-availability/authors' side (compounded by a data statement that claims all data are included). A good-faith figure-digitization shows all four correlations are genuinely positive with two of four (H, I) recovering near-exactly and no fabrication signal, but is too parameter-unstable to confirm the exact R/P. Verdict: a justified drop (data_restricted); the in-scope claim is qualitatively but not quantitatively confirmable.
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.