Genetic dissimilarity in cassava clones determined by multivariate techniques

Abstract

The genetic dissimilarity in 15 cassava clones was evaluated based on seven morpho-agronomic traits and multivariate procedures, aiming to select the most divergent and best in an approach to future hybridizations. The experiment was carried out in 2003/2005, in random blocks, with three replications. Based on the matrix of Mahalanobis’ generalized distance, the clones were grouped as related to similarity, using the algorithms of Tocher and Nearest neighbor. The relative importance of the traits was also evaluated. The multivariate techniques were concordant in the study of genetic dissimilarity and efficient in grouping clones with short genetic distance. The following traits contributed most to the genetic diversity: total weight of the aerial part, root diameter, root weight, length of tuber roots and harvest index.