Data Integrity Review: ANXA7/JNK knockdown and hepatocellular carcinoma lymphatic metastasis (Deng et al., Transl Cancer Res, 2021; DOI: 10.21037/tcr-20-2111)

Academic fraud report · Geng Detector

Summary

This review assesses a 2021 paper in Translational Cancer Research for potential data fabrication and methodological inconsistencies. The most serious finding concerns the cell invasion assay (Figure 1C), in which two independent shRNA treatment groups (shRNA-ANXA7 and shRNA-JNK), with means of 13 and 23 cells/field respectively, are reported with identical standard deviations of 0.816 to three decimal places—an outcome statistically implausible for real biological replicates and suggestive of fabricated or copied data. Additional concerns include suspiciously low variance in Western blot densitometry (Figure 2C), an internal contradiction between frozen-section and freeze-dried tissue methods, duplicated phrasing in the culture conditions, and mis-attribution of Leica microscopes to Japan. Original high-resolution blot images were not available, so image-forensic conclusions cannot be drawn. Confidence in the SD-duplication finding is high; other issues are supporting but individually non-determinative. Overall verdict: strong indicators of data manipulation warranting institutional investigation.

Verdict

🔴 Strong suspicion of data fabrication. The principal evidence is the cell-invasion counts in Figure 1C, where two biologically distinct treatment groups share an identical standard deviation (0.816) to three decimal places. Combined with additional irregularities in Western blot quantification, methods inconsistency, and equipment mis-attribution, the cumulative weight of evidence supports a formal investigation. No image-forensic analysis of the original blots was possible because high-resolution files were not supplied.

Key findings

Identical SDs in independent treatment groups (Figure 1C, invasion assay): 13±0.816 (shRNA-ANXA7) and 23±0.816 (shRNA-JNK) versus 41±3.266 (control). Identical SD values across groups with a nearly 2-fold difference in means are statistically implausible for counted replicates; 0.816 ≈ √(2/3) and 3.266 ≈ √(32/3), suggesting possibly generated or copied figures.
Western blot densitometry (Figure 2C) appears over-fitted: reported values 0.066±0.014, 0.027±0.002, and 0.030±0.007 show unusually small SDs (notably 0.002), consistent with data adjusted to suppress apparent significance between knockdown groups.
Methods–figure contradiction: Methods describe "frozen sections," whereas Figure 1D legend refers to "freeze-dried lymphatic tissue." These are different preparation techniques.
Copy-paste artefact in Methods: the phrase "placed in 37 ℃ incubator … with 5% CO2 atmosphere at 37 ℃" repeats the temperature, indicating careless editing consistent with template reuse.
Manufacturer mis-attribution: Leica microscopes are listed as "Leica, Japan"; Leica is a German manufacturer and is conventionally cited as Germany.

Evidence highlights

Quantitative (Figure 1C, invasion assay):
shRNA-ANXA7: 13 ± 0.816 cells/field
shRNA-JNK: 23 ± 0.816 cells/field
control-shRNA: 41 ± 3.266 cells/field
Numerical note: 0.816 ≈ √(2/3); 3.266 ≈ √(32/3).
Quantitative (Figure 2C, ATF2 densitometry):
control-shRNA: 0.066 ± 0.014
shRNA-ANXA7: 0.027 ± 0.002
shRNA-JNK: 0.030 ± 0.007
Textual (Methods, Cell lines and cell culture): duplication of "37 ℃" within a single sentence.
Textual (Methods vs Figure 1D legend): "frozen sections" vs "freeze-dried lymphatic tissue."
Equipment attribution (Methods): "inverted light microscope (Leica, Japan)" and "inverted epifluorescence microscope (Olympus, Japan)."
DOI preserved: 10.21037/tcr-20-2111.

Notes

Original Western blot and microscopy images were not provided; conclusions about splicing, background noise, or other image-based manipulation cannot be drawn and are explicitly out of scope.
The SD-duplication finding is, on its own, strong circumstantial evidence; however, ultimate determination of misconduct requires review of raw counting data and laboratory records by the authors' institution (First Affiliated Hospital of Dalian Medical University, as cited).
Findings 3–5 are individually minor but collectively reinforce concerns about drafting care and possible template-based authorship practices.
Recommended follow-up: request raw invasion-count files and uncropped blots from the corresponding author; post a PubPeer comment highlighting the SD duplication; notify the journal editor (Translational Cancer Research).