Integrity Assessment Report: Radically truncated MeCP2 rescues Rett syndrome-like neurological defects (Tillotson et al., Nature 2017)

Academic fraud report · Geng Detector

Summary

Verdict: No evidence of academic fraud detected (CLEARED). This is a forensic-style integrity review of a 2017 Nature paper (DOI: 10.1038/nature24058) by Tillotson, Bird and colleagues on a truncated MeCP2 protein rescuing Rett-like neurological defects. Four checks were performed and all passed. Statistical P-values showed authentic 'messiness', including marginal non-significant results such as P=0.071, P=0.055, and P=0.085, with inconsistent decimal precision (e.g., P=0.0002, P=0.333, P=0.604), inconsistent with batch-fabricated data. Methodological timeline (CRISPR plasmid Addgene #42230, GraphPad Prism 7, Zeiss LSM710, scAAV9) is internally consistent with 2015–2016 execution. Statistical tests were appropriately matched to data distributions (t-tests, KS, Friedman, Mantel–Cox). Image-level pixel analysis was not possible. Limitations: pixel-level forensic image analysis was not performed.

Verdict

✅ Cleared. No indicators of data fabrication, statistical malpractice, or methodological inconsistency were identified in the textual content of the report. The paper exhibits hallmarks of authentic experimental data.

Key findings

P-value distribution is authentically irregular: presence of marginal non-significant values and inconsistent decimal precision argues against algorithmic fabrication.
Methodological timeline is consistent: all reagents, software, and equipment are appropriate for the 2015–2016 experimental period.
Statistical methods are well-matched to data types: no 'one-test-fits-all' misuse detected.
Image-level pixel forensics not performed: insufficient evidence to confirm or rule out image manipulation.
Authors publicly disclose source data: including full western and Southern blot scans.

Evidence highlights

Marginal P-values (strong authenticity indicator):
Extended Data Fig. 2c: P=0.071
Fig. 3f: P=0.055
Fig. 3h: P=0.055
Extended Data Fig. 4c: P=0.085
Variable P-value precision: e.g., P=0.0002, P=0.333, P=0.604 — inconsistent precision is inconsistent with batch fabrication.
Reagent provenance:
Addgene #42230 (pX330-U6-Chimeric_BB-CBh-hSpCas9), published by the Zhang lab in 2013 — appropriate for 2015–2016 use.
Software/equipment: GraphPad Prism 7, FACSDIVA v8.0.1, Zeiss LSM710 — period-appropriate.
scAAV9 viral vector for P1/P2 neonatal intracerebral injection, consistent with cited Gadalla et al. 2013.
Statistical method–data matching:
Parametric (t-test) for normally distributed Open Field data.
Kolmogorov–Smirnov test for non-normal Elevated Plus Maze and Rotarod latency data.
Friedman test (non-parametric repeated measures) for Rotarod performance over time.
Mantel–Cox test for survival curves.
Acknowledgement of statistical support to Alastair Kerr.
Transparency markers:
Non-specific bands marked with asterisks (*) in Fig. 3a/4a — a rigorous practice rarely seen in fraudulent work.
Authors state: "All data are available... Source data underlying all graphs and full scans of all western and Southern blots are included."

Notes

DOI: 10.1038/nature24058
Authors: Rebekah Tillotson, Jim Selfridge, Martha V. Koerner, Kamal K. E. Gadalla, Jacky Guy, Dina De Sousa, Ralph D. Hector, Stuart R. Cobb, Adrian Bird.
Publication: Nature, 2017 (online 19 October 2017; PMC 2018 April 11).
Limitations: This assessment was based on textual content only; pixel-level forensic analysis of western blots, Southern blots, microscopy, and behavioural tracking images was not conducted. Definitive image-integrity verification requires the original figure files.
Recommended actions: None — no concerns warrant contact with the authors, PubPeer post, or journal notification.