1. Why is there a fundamental discrepancy between your similarity matrix and dendrogram? Table 4 shows G5 and G8 with the lowest similarity (0.14), yet in Figure 4 they are not positioned as the most distant pairs. Additionally, G1 and G3 show 0.96 similarity but Figure 6 places them in different clusters (G1 in cluster I, G3 in cluster II). This contradiction undermines your clustering interpretation.
2. Your PCA biplot (Figure 3) claims Dim1 explains 97.6% of variation, but with 12 traits and 10 genotypes, such extreme variance concentration is statistically improbable. Did you use standardized or non-standardized data? This suggests either data transformation errors or that your traits are highly redundant—yet you treat them as independent descriptors.
3. The STRUCTURE analysis (Figure 7) shows admixture patterns, but you never report the optimal K value via ΔK or LnP(D) plots. Your text mentions “three major groups” but Figure 7 appears to show K=2 or K=3 without clear justification. What was the actual optimal K, and why wasn’t the ΔK plot included?
4. Your genotype numbering in the STRUCTURE bar plot (Figure 7) does not match the 10 genotypes listed in Table 1—there appear to be 11 bars. Can you clarify this inconsistency? This suggests either a data entry error or mislabeling that compromises all downstream analyses.
5. For ISSR markers, you screened 12 primers but used only 8 polymorphic ones. However, primers with low PIC values (UBC-828: 0.116) produce unreliable clustering. Why include markers with PIC <0.20 when they contribute noise rather than meaningful genetic distances? This artificially inflates similarity coefficients.
6. Your genetic similarity range (0.14–0.96) is unusually broad for 10 genotypes of the same species collected from the same country. Could this extreme range be an artifact of using monomorphic primers in your initial screening, which you discarded, rather than genuine diversity?