Research misconduct report: Screening of high-yield antimicrobial peptide mutant strain of Bacillus subtilis HS-A38 and study of its antibacterial mechanism (Industrial Microbiology, 2018)

Academic fraud report · Geng Detector

Summary

Verdict: strong evidence (实锤) of data fabrication. The report identifies three serious integrity issues in Meng et al. (2018), DOI 10.3969/j.issn.1001-6678.2018.02.005. First, Figure 2 (UV mutagenesis lethality curve) and Figure 3 (NTG chemical mutagenesis lethality curve) are reported to be near-pixel-identical, despite describing fundamentally different physical vs. chemical mutagenesis treatments on the same strain. Second, time-course leakage assays (Figure 7/8) show unusually smooth, step-like increases with near-uniform error bars, lacking the stochastic noise expected of real biological measurements. Third, Table 1 standard deviations for 8 strains fall in an implausibly narrow band (1.1–1.4 mm). Confidence in the image-duplication finding is high based on visual inspection; the data-irregularity findings are suggestive but rely on visual judgement rather than raw values. Final determination requires institutional or editorial investigation with access to source data.

Verdict

🔴 Substantiated (实锤). Multiple independent integrity issues identified in the same paper, with the strongest evidence concerning apparent duplication of mutagenesis lethality curves.

Key findings

Figure 2 vs. Figure 3 duplication: The UV-mutagenesis lethality curve and the NTG (nitrosoguanidine) chemical-mutagenesis lethality curve are visually near-identical, despite the two treatments acting through fundamentally different mechanisms. Cross-treatment identity of lethality kinetics is a hallmark of fabricated or copy-pasted data.
Implausibly smooth time-course data (Figure 7/8): Extracellular leakage/OD260 or extracellular protein measurements across time points show a textbook-perfect monotonic stepwise increase with highly uniform error bars, inconsistent with the stochastic noise expected from real biological assays.
Suspiciously narrow SD range in Table 1: Standard deviations for inhibition-zone diameters across 8 strains (parent + 7 mutants) fall within 1.1–1.4 mm (e.g., 1.3, 1.2, 1.3, 1.2, 1.4, 1.2, 1.1), violating the expected variability between biological replicates of strains with differing antimicrobial activity.

Evidence highlights

Figure 2 and Figure 3 (page 3): two curves claimed to represent UV and NTG mutagenesis on the same strain show overlapping data-point positions and trajectory.
Figure 7/8 (page 5): step-wise monotonic increases with uniform error-bar orientation and magnitude across time points.
Table 1 (parent strain and 7 mutants, initial screen and subculture): inhibition-zone diameters reported with SDs clustered in a ~0.3 mm band (1.1–1.4 mm).
DOI: 10.3969/j.issn.1001-6678.2018.02.005
Publication: 工业微生物 / Industrial Microbiology, Vol. 48 No. 2, April 2018.
Authors: 孟兆丽, 丛丽娜, 曾国鸿, 马艺萌.

Notes

The image-duplication finding (Figure 2 vs. Figure 3) carries the highest evidentiary weight; the other two findings (smoothness of time-course data; narrow SD range) are consistent with fabrication but are visual judgements and would be strengthened by inspection of the raw numerical data.
Confidence is high for Finding 1; moderate for Findings 2 and 3.
The authors have not, to public knowledge, been contacted during this report generation. Recommended actions include contacting authors for raw colony counts and raw absorbance/protein readings, posting concerns on PubPeer, and notifying the journal editorial office (《工业微生物》); reporting to the authors' institution (Dalian Polytechnic University, College of Biological Engineering) is suggested but not yet initiated.
This report is AI-assisted and intended for academic-discussion purposes only; formal determination of misconduct requires an institutional or editorial investigation with access to source data.