Integrity Review: Genome-wide linkage mapping of yield-related traits in three Chinese bread wheat populations using high-density SNP markers

Academic fraud report · Geng Detector

Summary

Verdict: No evidence of academic misconduct detected. This 2018 paper in Theoretical and Applied Genetics (DOI: 10.1007/s00122-018-3122-6) by Faji Li et al. was assessed for image manipulation, statistical integrity, and methodological consistency. The study performs QTL mapping across three wheat populations and 12 traits, producing linkage maps and extensive tabular data rather than image-heavy outputs. Visual inspection found no splicing or fabrication in the linkage maps, and no risky blot/gel images to manipulate. Statistical review confirmed plausible environment-dependent variation in R² values (e.g., 22.0–59.0% for QSN.caas-4BS; 9.9–43.1% for QKNS.caas-4BS), consistent with complex quantitative traits rather than fabricated uniformity. The experimental timeline (field trials 2012–2015; publication 2018) and use of the 90K iSelect array and IWGSC RefSeq v1.0 reference genome align chronologically with available resources. Confidence in this verdict is moderate-to-high, though absence of misconduct cannot be fully proven; raw data and replicate records were not independently audited.

Verdict

No actionable findings of academic fraud. The paper appears methodologically sound, statistically plausible, and chronologically consistent.

Key findings

No image manipulation indicators: The paper relies on linkage maps (Fig. 1) and data tables; no Western blots, gels, or flow cytometry images that are common targets for fabrication. Linkage maps show no obvious splicing or assembly artifacts.
Statistically realistic QTL effects: Reported R² ranges (e.g., 22.0–59.0% for QSN.caas-4BS; 9.9–43.1% for QKNS.caas-4BS) reflect genuine environment-dependent variability across E1–E10, consistent with complex quantitative traits rather than artificially uniform outputs.
Logical congruence with known loci: QTL positions correspond appropriately with established markers such as Rht-B1 and Rht-D1, as discussed by the authors.
Consistent methodological timeline: Field trials (2012–2013, 2013–2014, 2014–2015 cropping seasons) precede the 2018 publication; the wheat 90K iSelect SNP array (Wang et al. 2014) and IWGSC RefSeq v1.0 genome assembly were available prior to manuscript preparation.
No anachronistic resources detected: All referenced tools and databases predate the study period.

Evidence highlights

DOI: 10.1007/s00122-018-3122-6
R² ranges reported: QSN.caas-4BS 22.0–59.0%; QKNS.caas-4BS 9.9–43.1%
Reference genome: IWGSC RefSeq v1.0 (Chinese Spring)
SNP platform: wheat 90K iSelect SNP array
Field trial seasons: 2012–2013, 2013–2014, 2014–2015 (10 environments total: E1–E10)
Known loci cross-referenced: Rht-B1, Rht-D1

Notes

Limitations: This review did not access the original raw phenotypic or genotypic datasets; therefore, independent verification of underlying numerical values was not possible.
The absence of high-risk image content reduces (but does not eliminate) the probability of certain manipulation types.
Statistical review was based on reported values, not re-analysis of raw data.
A formal institutional investigation would be required for any definitive declaration of integrity.