Integrity Review of "Robo1/2 regulate follicle atresia through manipulating granulosa cell apoptosis in mice" (Sci. Rep. 5, 9720; DOI: 10.1038/srep09720)

Academic fraud report · Geng Detector

Summary

This text-based integrity review of a 2015 Scientific Reports paper evaluates logical consistency, statistical reporting, and figure-text alignment. The verdict is highly suspicious (orange). No pixel-level image or Western blot analysis was performed; all conclusions derive from textual and quantitative inconsistencies. Key concerns include a reported sample size of n=1 for the Robo1/2 secondary follicle group in Figure 7 IOD analysis, which makes SD/SEM calculation and significance testing against n=11 biologically and mathematically impossible. A second notable issue is a clear mismatch between Results text and the Figure 7 legend regarding CD31 (text cites 7I/7J; legend assigns CD31 to 7K–7L and SMA to 7H–I/7J). Additional concerns include inappropriate use of Student's t-test across multi-factor designs and an inadequate follicle-counting protocol (average of only three sections, five apart). No timeline or equipment anomalies were found. Confidence is moderate; final judgment requires institutional investigation.

Verdict

🟠 Highly suspicious. Multiple textual and statistical inconsistencies undermine the credibility of the reported findings. The review is text-based only and cannot independently verify image integrity or raw data.

Key findings

Statistical impossibility (n=1 with error bars). Figure 7 legend specifies IOD analysis of CD34 expression with "secondary (WT n=11, Robo1/2 n=1)" and "mature follicles (WT n=5, Robo1/2 n=5)". An n=1 sample cannot yield SD/SEM, nor allow a valid significance test against n=11.
Figure–text misalignment in Figure 7. The Results section describes CD31 immunostaining as "Fig. 7J compared with the wild-type ovary (Fig. 7I)." However, the legend assigns 7K–L to CD31 and 7H–I (and 7J as a quantification panel) to SMA. The textual citation appears to be a leftover from a draft using different panel assignments.
Statistical method mismatch. The Methods section states Student's t-test for real-time PCR and Chi-square for IF. Given multi-factor designs (age: 4 vs 10 weeks; genotype: WT vs Robo1/2 KO; stage: primary/secondary/mature), ANOVA with post-hoc tests would be the appropriate baseline; reliance on pairwise t-tests raises p-hacking concerns.
Substandard follicle quantification. Methods state follicle counts are derived from "the average of the counts from three sections (five sections apart)" along the long axis of the ovary. This sparse sampling is inconsistent with accepted stereological practice in reproductive biology and may explain the highly variable data in Figure 3D–E.
No timeline or equipment anomalies. Lipofectamine 2000, Olympus IX51, and ethics-approval statements are appropriate for the 2014–2015 timeframe.

Evidence highlights

Direct quote from Figure 7 legend: "secondary (WT n=11, Robo1/2 n=1), and mature follicles (WT n=5, Robo1/2 n=5)" — DOI: 10.1038/srep09720.
Results text: "The immunofluorescent staining for CD31... (Fig. 7J compared with the wild-type ovary (Fig. 7I)."
Figure 7 legend: "(H–I): The fluorescent immunostaining against SMA... (K–L): The fluorescent immunostaining against CD31..."
Methods (Follicle categorization and counting): "The total number of follicles per ovary was determined by taking the average of the counts from three sections (five sections apart) cut along the long axis of the whole ovary."
Methods (Statistical Analysis): Student's t-test for real-time PCR; Chi-square for IF.

Notes

All findings are derived from logical and methodological reading of the published article; no image forensics were performed.
Verdict should be treated as a signal warranting institutional review, not a definitive finding of misconduct.
Recommended follow-ups: verify Figure 7 bar-graph error bars on the n=1 group; request uncropped blots and FCS files from the corresponding author; raise concerns via PubPeer and notify the editorial office of Scientific Reports.