Maizura, A. S.¹, Rahim, F. A. ¹, Sentoor, K. G ¹. and Umikalsum, M. B².

Abstract
Coconut (Cocos nucifera L.) is a vital economic crop supplying food, oil, and raw materials. Breeding programmes require superior varieties with improved fruit composition, which depend on understanding trait variation among genotypes. This study evaluated 30 coconut genotypes at MARDI Bagan Datuk, Perak, using multivariate techniques principal component analysis (PCA), cluster analysis, and biplot analysis to assess variation in fruit components. The analysis showed significant differences in fruit weight (483.0 – 1765.0 g) and nut weight (208.2 – 1248.3 g). Kernel and copra weights were highest in RIL (518.6 g) and TAGT (274.7 g), respectively. Oil content averaged 62.8%, peaking at 66.2% in LAGT. PCA identified three principal components that explain 98.39% of the variation, with split nut weight, kernel weight and fruit weight being the major contributors. Biplot analysis revealed strong positive correlations among fruit weight, nut weight and kernel weight, supporting trait-based selection. Cluster analysis grouped genotypes into three main clusters reflecting trait similarities rather than geographic origin. These results highlight the importance of multivariate analysis for identifying key traits for coconut improvement. The findings provide valuable insights for selecting promising genotypes. Integrating molecular tools into future research could further enhance the precision of selection and breeding efficiency.

Keywords: biplot, coconut germplasm, fruit component analysis, multivariate analysis, PCA

Full text (481KB)