Integrity Assessment Report: GPAT4 sustains endoplasmic reticulum homeostasis in endocardial cells and safeguards heart development

Academic fraud report · Geng Detector

Summary

This report evaluates a 2025 Nature Communications paper (DOI: 10.1038/s41467-025-58722-5) by Zhao et al. on GPAT4 in cardiac development. The verdict is “doubtful” (yellow flag). The primary concern is statistical misuse: in Figure 6 and other rescue experiments comparing WT, Gpat4 knockout (KO), and double-knockout (DKO) groups, the authors repeatedly used two-tailed unpaired Student's t-tests rather than one-way ANOVA with appropriate post-hoc tests, inflating Type I error risk and undermining the validity of rescue claims. Secondary issues include minor labeling irregularities in Figure 1H/2H Western blots (e.g., a ‘KOWT’ textual artifact) and tight reference timelines involving AlphaFold 3 (DeepMind, May 2024) and OMM-NEMOs-mKate-ER (Nat. Chem. Biol. 2024) cited after the March 2024 submission date. Pixel-level image forensics were not possible due to PDF text-only input. Confidence is moderate; recommendations include requesting raw blots and Prism files and prompting authors to publish an erratum with corrected ANOVA analyses.

Verdict

⚪ Doubtful (yellow flag). The paper contains a substantive but not necessarily fatal statistical flaw in its core rescue experiments, alongside minor editorial/timeline anomalies. No fabricated data or image manipulation has been demonstrated at this stage.

Key findings

Misuse of Student's t-test for three-group comparisons in Figure 6 (and elsewhere), where one-way ANOVA with post-hoc tests is the appropriate method.
Western blot panel labeling anomaly: Figure 1H text contains a ‘KOWT’ string and apparent lane-order inconsistency between Figure 1H and other figures (KO/WT positioning differs).
Reference timeline concerns: AlphaFold 3 (DeepMind, released May 2024) and the Zhang et al. Nat. Chem. Biol. 20, 699-709 (2024) OMM-NEMOs-mKate-ER probe are cited in a manuscript received 19 March 2024.
Pixel-level image forensics could not be performed because only PDF-extracted text (with molecular weight tokens such as 52kDa 21kDa 28kDa 42kDa) was available.

Evidence highlights

Figure 6G, 6H, 6O, 6P: three-arm comparisons (WT, KO, DKO such as Gpat4-/-; Ifnar1-/- or Gpat4-/-; Sting-/-) reported with two-tailed unpaired Student's t-tests and a mixture of “ns” and significance asterisks rather than ANOVA.
Figure 1H and Figure 2H: Western blot lane annotations show ‘KO WT’ ordering that contrasts with the WT-first convention used elsewhere; the OCR token ‘KOWT’ suggests either crowding or mislabeling.
Methods/References: AlphaFold 3 usage cited via a 2024 Nature reference (Ref 26); OMM-NEMOs-mKate-ER usage cited as Zhang et al., Nat. Chem. Biol. 20, 699-709 (2024) (Ref 35); received date 19 March 2024.
DOI: 10.1038/s41467-025-58722-5 (Nature Communications, 2025).
Authors: Tianyang Zhao, Kuipei Jin, Xiaodong Wang, Xiong Su, Youjun Wang, Mingming Gao, Wen Luo, Hongyuan Yang, Zhongzhou Yang.

Notes

The t-test misuse pattern (multiple t-tests across three groups without ANOVA) is a recurring methodological deficiency rather than an isolated slip, directly affecting whether the genetic rescue of the Gpat4 phenotype is statistically supported.
The ‘KOWT’ textual artifact and lane-order discrepancy are currently low-severity; resolution requires the original high-resolution Western blot files to determine whether it is a typesetting error or a substantive mislabeling that could invert the reported direction of effect.
The post-submission citations to AF3 and the 2024 OMM-NEMOs-mKate-ER probe are plausibly attributable to revisions during peer review; however, they warrant verification that the corresponding experiments were performed or re-analyzed before citation.
No image-duplication, splicing, or exposure-manipulation findings can be asserted from PDF text alone; raw blot/microscopy images and Prism source files should be requested to complete the assessment.