Chemometric Multivariate Analyses of Kernel Biometric Techniques and Shell Morphological Attributes for Analyzing Pistacia vera L. Genotypes Variability

In this study, the corresponding correlations, interdependencies, and relationships of nine traits of pistachio ( Pistacia vera L.) kernels were established. Here, the percentage of dehiscent shells, shell homogeneity, shell apex, position of suture opening, suture opening, number of pistachio kernels/100 g, (length/width) pistachio kernel ratio, pistachio kernel flavour and fat content of pistachio kernel were studied. The evaluation of experimental data was performed by applying a chemometric approach: Principal Component Analysis (PCA), Bayesian Networks method and Network Meta-Analysis. In this regard, based on PCA the first principal component of shell and kernel had high loadings for shell homogeneity. By using Bayesian Networks method, dendrogram of selected pistachio varieties demonstrate a high similitude existing between varieties. Furthermore, the of shell and kernel dendrogram of studied pistachio varieties demonstrated that all studied accessions could be separated into four distinct groups. Based on cluster analysis, a detailed comparison between the varieties demonstrated that the accession named ‘KERMAN’ showed a significant similarity with the Tunisian genotype ‘ELGUETAR’. In this line, Network Meta-Analysis exhibited the different interdependencies between morphological parameters, and the hierarchical clustering via heat maps displayed relationships between studied factors. These results showed a potential use of desired characteristics at least in some infra-specific studies in Pistacia vera L. The findings of this study will help plant growers by highlighting hidden and eventual relationships between Pistacia vera L. varieties.