ABSTRACT
Bitter melon (Momordica charantia L.) is a plant species recommended by the Brazilian Health Regulatory Agency (Anvisa) as hypoglycemiant. The characterization of plants is an essential step in any breeding program. The objective of the present work was to organize and characterize a bitter melon germplasm collection, based on morpho-agronomic characters, to assess its genetic diversity and identify genotypes of agronomic interest. Eighty-eight genotypes were characterized for 38 descriptors. Redundant descriptors were identified through Principal Component Analysis (PCA); after their exclusions, a new PCA was carried out to verify the dispersion among the genotypes. Groups in the PCA were defined using the kmeans clustering method. The groups were studied for phenotype pattern using radar chart. Populational diversity was estimated through Shannon and Pielou indexes. Intra group diversity was estimated through analysis of similarity (anosim). The relative importance of variables for diversity was also estimated. Seventeen variables were redundant. The genotypes were grouped into 5 groups. Groups G1 and G5 were antagonist regarding fruit and seed productions and fruit, leaf, and seed sizes. A trend of decrease in fruit, leaf, and seed sizes was found in groups from G1 to G5. The diversity was high. Intra group diversity was high among small fruit genotypes, and low for medium-sized fruit genotypes. The variable number of male flowers (NMFL) was identified as that presented the greatest contribution to estimation of diversity. The genotypes UFRRJ MSC072, 042, 028, and 087 stood out with the highest number of fruits produced.
Keywords
Momordica charantia L.; Principal Component Analysis; K-means clustering analysis; Diversity indexes; Grouping patterns
Thumbnail
Thumbnail
Thumbnail
Thumbnail
Thumbnail
