Integrity Review: 'Metabolome integrated with transcriptome, and genome analysis revealed higher accumulations of phytoalexins enhance resistance against Magnaporthe oryzae in new Zhefang rice variety diantun 506' (DOI: 10.1186/s12870-025-06856-5)

Verdict

Highly suspicious. Multiple statistical, methodological, and writing irregularities were identified from text alone. The most serious is the use of a Pearson correlation threshold of |r| > 0.09 to build a putative 'strong' DAM–DEG interaction network (Figure 6), which is statistically indefensible. Methodological descriptions show copy-paste artifacts. Image integrity and raw data cannot be assessed without the original figures and supplementary tables.

Key findings

• Statistically indefensible correlation threshold. Pearson |r| > 0.09 corresponds to R² ≈ 0.0081 (≈0.81% of variance explained), indistinguishable from random noise. Calling these edges 'strong connections' misrepresents the analysis.
• Self-contradictory DAM filter criteria. The methods state DAMs were identified by 'VIP > 1, P-value ≤ 0.05, fold change > 2, and FC > 0.5.' FC > 2 and FC > 0.5 cannot both be selection rules; this appears to be an unedited copy-paste error.
• Mathematically loose overlap claim. Reported 588 (D502) and 595 (D506) DAMs and '≈55% overlap' (420) only holds for Venn-diagram union (588+595−420 = 763, 420/763 ≈ 55.04%); phrasing as 'approximately' is ambiguous and suggests imprecise bookkeeping.
• Possible redundant (salami) publication. Materials, inoculation regime (13-strain mix, 48 h sampling), and D502 vs D506 comparison mirror reference [13], a 2025 BMC Plant Biology paper by the same group, indicating the transcriptome and metabolome may have been split from one experiment into two publications.
• Copy-paste artifacts in text. Phrase 'i dentif i c a t i o n of the ORF region' near the NCBI ORF finder description suggests PDF/web extraction with watermark or formatting residue, undermining confidence in the originality of the methods section.

Evidence highlights

• Correlation threshold: 'a Pearson correlation coefficient > 0.09 or <−0.09' used to define DAM–DEG interactions feeding Figure 6; result described as showing 'strong connection.'
• DAM criteria: Methods state 'DAMs were identified with VIP > 1, P-value ≤ 0.05, fold change > 2, and FC > 0.5.'
• Counts: D502 = 588 DAMs; D506 = 595 DAMs; '420 DAMs approximately (55%) were associated with both varieties.'
• Reference overlap: Reference [13] is the team's own 2025 BMC Plant Biology paper; same varieties (D502, D506), same 13-strain M. oryzae inoculum, same 48 h sampling window described in 'Previously, we performed integrated transcriptome and genome analysis... [13].'
• Typographic artifact: 'i dentif i c a t i o n of the ORF region' in Protein/sequence/3D structure analysis subsection.
• DOI preserved: 10.1186/s12870-025-06856-5.

Notes

• Image-based checks (Western blot duplication, microscopy manipulation, figure reuse) were not possible because original high-resolution figures were not supplied; the report relies solely on text, statistical logic, and methodological consistency.
• The 0.09-threshold issue alone is sufficient to invalidate the network-level conclusions of Figure 6 unless authors can justify this cutoff with permutation testing or external validation.
• The FC contradiction is a clear editorial-level methods error that should be corrected before the network claims are considered credible.
• The overlap with reference [13] warrants verification by BMC Plant Biology editors regarding sample provenance, RNA-seq reuse, and proper cross-citation.
• Confidence is moderate and limited by the absence of raw metabolomics data, correlation matrices, and figure files. Final determination of misconduct requires investigation by the journal/institution.