Integrity Review of Garg et al., 'Systemic Delivery of MeCP2 Rescues Behavioral and Cellular Deficits in Female Mouse Models of Rett Syndrome', Journal of Neuroscience 2013 (DOI: 10.1523/JNEUROSCI.1854-13.2013)

Academic fraud report · Geng Detector

Summary

This text-level review of Garg et al. (J. Neurosci., 2013; DOI: 10.1523/JNEUROSCI.1854-13.2013) returns a verdict of 'Questionable.' Three issues raise concern. First, Figure 6E reports near-perfect recapitulation of endogenous MeCP2 fluorescence peak positions in wild-type (Mecp2+/y) versus AAV-treated (Mecp2Bnull/y) brainstem cells (171–174 AU and 202–206 AU versus 172–174 AU and 202–206 AU), a degree of quantitative agreement unlikely between independent biological samples. Second, key mechanistic claims rest on n=2 mice per genotype, yet Figure 6B reports ANOVA results of p<0.001, which is statistically implausible without zero within-group variance. Third, Results text reports inconsistent apnoea outcomes in n=5 treated mice (2 improved, 2 worsened, 1 unchanged), undermining the treatment narrative. Image-level analysis was not performed owing to lack of high-resolution source files. Confidence is moderate; pixel-level verification and author data release are required for a definitive determination.

Verdict

Questionable (🟡). Text-level evidence indicates improbable quantitative replication, inadequate sample sizes for the statistics reported, and internally inconsistent therapeutic outcomes. Pixel-level image forensics were not performed.

Key findings

Impossibly precise recapitulation of fluorescence peaks (Figure 6E): Wild-type brainstem cells reportedly show MeCP2 fluorescence peaks at 171–174 AU and 202–206 AU, while AAV-injected *Mecp2*<sup>Bnull/y</sup> cells reproduce these at 172–174 AU and 202–206 AU. This near-identical peak positioning across genetically and procedurally distinct samples is biologically implausible.
Sample size inconsistent with reported statistics (Figure 6, Figure 7): The legend states n = 2 mice per genotype, 5 fields per mouse, yet Figure 6B reports ANOVA significance of *p < 0.001*. Achieving this threshold with n = 2 requires effectively zero within-group variance, which is implausible in vivo.
Internally contradictory rescue narrative (Results, paragraph 6): Of 5 AAV9/MeCP2-treated female mice, 2 showed reduced apnoea frequency (1.3- and 3-fold), 2 worsened (1.6- and 10-fold), and 1 was unchanged. The authors nevertheless conclude the experiment supports continued development.
Image analysis not performed: Original high-resolution images were not available to this review; pixel-level splicing or duplication cannot be assessed.

Evidence highlights

DOI: 10.1523/JNEUROSCI.1854-13.2013
Figure 6E peak positions: WT = 171–174 AU and 202–206 AU; treated = 172–174 AU and 202–206 AU
Sample size: n = 2 mice per genotype, 5 fields per mouse
Reported Figure 6B significance: *p < 0.001*
Female treatment cohort: 2 improved (1.3×, 3×), 2 worsened (1.6×, 10×), 1 unchanged

Notes

Confidence is moderate; findings are derived from published text and figure legends rather than raw data or high-resolution images.
The transparent reporting of adverse outcomes (worsened apnoea in 2 of 5 mice) argues against wholesale fabrication but is consistent with selective emphasis of favourable behavioural results.
Recommended next steps: (1) request raw ImageJ quantification files for Figure 6E and the statistical source files supporting Figure 6B; (2) post structured queries on PubPeer regarding peak-position duplication and n = 2 vs. p < 0.001; (3) request uncropped Western blots from the corresponding author (Gail Mandel) for full image-level review.
A determination of misconduct requires institutional or editorial investigation; this report flags concerns, not a finding of fraud.