Geng Academic Fraud Report: Paper "Enhancer regulator MLL4 controls skeletal muscle metabolic efficiency by limiting AMPK-mediated fuel catabolism" (Nature Communications, 2025)

Academic fraud report · Geng Detector

Summary

This report raises two methodological and reporting concerns about the above paper (DOI: 10.1038/s41467-025-66684-x). Overall verdict: questionable. Key issues: (1) In the Figure 1 legend, two distinct physiological measurements — running time and running distance — are assigned an identical P-value series ("0.0387, 0.0027, 0.0387, 0.0027"), which is statistically highly improbable for independent continuous variables and strongly suggests a copy-paste error in legend preparation. (2) The Figure 4 RNA-seq analysis uses only n=2 independent samples per group, a design that is methodologically weak for differential expression analysis (typically requiring n≥3), although the authors disclose this transparently. Confidence: moderate for finding 1 (strong statistical anomaly) and high for finding 2 (transparent but limited design). No image manipulation, data fabrication, or ethical misconduct is alleged.

Verdict

🟡 Questionable. The paper raises no confirmed misconduct but exhibits a probable copy-paste reporting error and a weak transcriptomic sample size.

Key findings

Figure 1 legend anomaly: Identical P-value series reported for two independent measurements (running time and running distance), suggesting a copy-paste error in legend drafting.
RNA-seq sample size limitation: Figure 4 uses only two independent samples per group, which is methodologically fragile for differential expression analysis, though the authors disclose this honestly.

Evidence highlights

Figure 1 legend text contains the exact P-value sequence "*P value: 0.0387, 0.0027, 0.0387, 0.0027*" applied to both Fig. 1f (running time) and Fig. 1g (running distance). For two distinct continuous physiological variables, four identical P-values to four decimal places is statistically very unlikely.
Figure 4 legend explicitly states: "Each group is represented by RNA-seq data from two independent samples generated from muscles from indicated mice." Sample size n=2 per group is below the conventional minimum (n≥3) for reliable RNA-seq DEG analysis.
DOI: 10.1038/s41467-025-66684-x

Notes

No image manipulation, Western-blot irregularities, or signs of data fabrication are identified in this report.
Finding 1 could plausibly reflect a benign typographical error rather than deliberate falsification, but the precision to four decimal places makes coincidence implausible; raw statistical outputs should be reviewed.
Finding 2 reflects a methodological limitation, not misconduct; transparent reporting is noted.
Recommended follow-ups: request underlying statistical output for Figure 1f/g from the authors; note the RNA-seq n=2 limitation when interpreting Figure 4 conclusions.
Report is AI-assisted and intended for academic discussion only; final determination of misconduct requires institutional investigation.