Multivariate Analysis and Genetic Diversity in Wood Apple (Feronia limonia L.) Using R-based Analytical Approaches

Background: Understanding genetic variability in wood apple is crucial for effective germplasm conservation and improvement. Traditional univariate analyses, while informative, are limited in revealing complex trait interactions. Multivariate statistical techniques such as Principal Component Analysis (PCA), hierarchical clustering, and heatmap visualization enable researchers to reduce dimensionality, uncover trait co‑variation, and classify genotypes based on multivariate similarity. Aim: The present study assesses phenotypic variability among nine wood apple (Feronia limonia L.) genotypes using comprehensive multivariate statistical approaches, with the goal of understanding genetic diversity and identifying superior types for crop improvement. Methodology: A field evaluation was conducted under uniform agro-climatic conditions with three replications in Randomized Complete Block Design (RCBD), with fourteen morphological, yield, and biochemical traits recorded for each genotype according to standard protocols. Multivariate statistical tools, including Principal Component Analysis (PCA), hierarchical clustering, and heatmap visualization, were applied using R software to classify genotypes and map trait associations. Results: Significant variability was observed across all measured traits. PCA showed that the first two principal components accounted for 92.7 % of the total variation, with PC1 alone explaining 74.4 %, indicating strong discriminatory power of key traits. Total sugar, pulp weight, fruit length, and yield per tree were identified as major contributors to genotypic divergence. Hierarchical clustering grouped the genotypes into two main clusters, clearly distinguishing elite types such as WFL-03 and WFL-08. Heatmap visualisation further confirmed inter-genotypic diversity and revealed unique expression patterns among traits. Hierarchical clustering and heatmap analysis further classified the genotypes into distinct performance groups, with WFL-03 and WFL-08 emerging as superior in yield and quality traits. These results highlight the effectiveness of multivariate approaches in uncovering hidden patterns in trait expression and guiding selection strategies. Conclusion: The substantial genetic variability uncovered by multivariate analysis highlights the effectiveness of these tools in identifying superior genotypes. These insights provide a robust foundation for future selection, breeding, and conservation efforts, promoting the genetic improvement and sustainable utilization of the wood apple as an important but underexploited fruit crop.