Exploring heritability, correlation, path coefficient, clustering and PCA for morphological traits in Persian walnut (Juglans regia L.)

The current study aimed to elucidate the genetic variability existing in walnut germplasm and to determine the potential traits for selecting elite genotypes. Ultimately, an ample amount of genetic variability was reported in the existing seedling-originated walnut population for observed morphological traits. The computed GCV and PCV were high for KW (31.63% & 32.75%) and moderate for NW (25.45% & 26.80%) whereas, the remaining nut and kernel traits recorded lower values. All vegetative and floral traits exhibited moderate GCV and PCV values, except for LA (32.10% & 34.10%), which showed comparatively higher values. High heritability coupled with high genetic gain was observed for KW (93.28% & 62.94%), NW (90.14% & 49.76%), LA (88.60% & 62.24%), DMF (99.32% & 49.33%), and DFF (99.29% & 51.84%). A positive correlation was observed between NW and KW (r g =0.9685, r p =0.9235). However, KP was positively and significantly correlated with KW (r g =0.7880, r p =0.7070) and NW (r g =0.6223, r p =0.4453). Several vegetative and floral traits were also significantly positively correlated with nut and kernel traits at both the genotypic and phenotypic levels. Notably, nut and kernel traits showed strong interdependencies with each other, majorly NW and KW were found to be effective traits for direct selection in further improvement studies in walnuts. Path coefficient analysis revealed that NDTS has highest positive direct effect (0.4781) on KP, followed by CL (0.1964), NPF (0.1584), DMF (0.0826), LA (0.0824), and LW (0.0601). Clustering resulted in the formation of four clusters; The maximum intra-cluster distance was observed in Cluster IV (233.95 D 2 ), likewise the highest inter-cluster distance was recorded between Clusters II and IV (884.35 D 2 ) holding 25 and 33 individuals, respectively. PCA provided a visualization of the complete dataset in reduced-dimensional plots. The first two principal components (PCs), PC1 (38.73%) and PC2 (12.39%), explained 51.12% of cumulative variance, while a total of 77.30% of variance was explained by first six PCs. The study confirmed NW, KW and KP as the most discriminative and dependable selection indices for identifying superior genotypes within a genetically diverse walnut population. However, CL, NPF, DMF, DFF, LA, LW emerged as important secondary selection criteria influencing kernel yield via their direct and indirect effects.