sdRNA-D43 paper in Cell Communication and Signaling shows impossible non-parametric statistics with n=3 (DOI: 10.1186/s12964-024-01975-2)

Academic fraud report · Geng Detector

Summary

This academic-integrity review assesses the paper "sdRNA-D43 derived from small nucleolar RNA snoRD43 improves chondrocyte senescence and osteoarthritis progression by negatively regulating PINK1/Parkin-mediated mitophagy pathway via dual-targeting NRF1 and WIPI2" by Zengfa Deng et al., published in Cell Communication and Signaling (DOI: 10.1186/s12964-024-01975-2). Verdict: highly suspicious, recommended for deeper investigation. The central finding is a mathematical impossibility: figure legends (e.g., Fig. 3C/E/G, Fig. 6, Fig. 7) state n=3 per group and declare use of non-parametric tests, yet report P<0.001. With 3 vs 3 observations, a two-sided Mann-Whitney U test can produce only 20 permutations, and the smallest achievable P value is 0.1 (two-sided) or 0.05 (one-sided); P<0.001 is mathematically impossible. Either the sample size, test choice, or P values are misrepresented. Secondary concerns include templated, AI-like Methods writing, a typo ("reveled" for "revealed") in Fig. 6, and the non-existent reagent "ADAB" instead of DAB. Limitations: no raw blots were available for image-forensic analysis.

Verdict

Highly suspicious; in-depth investigation recommended. The strongest evidence concerns an internal mathematical contradiction between stated sample size, test type, and reported P values.

Key findings

Impossible non-parametric statistics with n=3. Multiple figure legends (e.g., Fig. 3C, 3E, 3G; Fig. 6; Fig. 7) state n = 3 per group and the Methods describe non-parametric testing, yet significance markers indicate *P < 0.001.

Mathematical contradiction. With two groups of n=3, the Mann-Whitney U test has only C(6,3)=20 possible allocations; the minimum two-sided P value is 0.1 (one-sided 0.05). P<0.001 cannot be produced by such a test.

Implausible alternative explanation. If a parametric t-test were silently substituted and produced P<0.001 with n=3, the implied standard deviations would be near zero, which is biologically unrealistic for primary chondrocyte cultures and Western blot quantitation.

Templated/AI-like writing in Methods. The Transfection section contains an explanatory aside ("The sdRNA-D43 inhibitor can bind complementarily to the sdRNA-D43 sequence... thereby inhibiting the function of sdRNA-D43") that reads like generated filler rather than a standard protocol.

Substantive typos. Figure 6 legend contains "reveled" (should be "revealed"). The IHC section references an "ADAB substrate kit"; no such reagent exists. The correct reagent is DAB (diaminobenzidine).

Patient-cohort inconsistencies. Abstract states n=36 patients, while Fig. 1H states n=30 per group. This can be reconciled by exclusion but is insufficiently clarified.

Evidence highlights

Figure legends cited as "Non-parametric test" with n = 3 and *P < 0.001 (Page 11, Fig. 3 and elsewhere).
Permutation count: C(6,3) = 20; minimum two-sided P for MWU ≈ 0.1; minimum one-sided P ≈ 0.05.
Typo evidence: "reveled" in Fig. 6 legend (Page 7 area); "ADAB substrate kit" in IHC section.
Ethics approvals present: [2021]334 and SYSU-IACUC-2022-001910; timeline (received 02 Feb 2024, accepted 01 Dec 2024) is consistent.
DOI: 10.1186/s12964-024-01975-2.

Notes

Image forensics not performed: raw Western blot images were not provided; pixel-level analysis of Fig. 3, Fig. 6, and Fig. 7 is needed to assess band reuse or splicing.
The statistical contradiction alone is sufficient to request raw data and SPSS/GraphPad output files from the authors.
A PubPeer post and an editorial inquiry are warranted.
Final determination of misconduct requires institutional investigation; this report flags anomalies only.