Comparison of methods for selection of castor beans lineages

Abstract

Article AbstractThe choice of the most appropriate method is determined by the precision desired by the researcher, by the ease of the analysis, as well as by the way of obtaining the data. In order to select lineages of low size and high productivity this study aimed to evaluate different methods of cluster analysis in the representation of genetic divergence, compared to univariate methods. The analyzed variables were grain yield, plant size and oil yield of 24 lineages of castor beans cultivated in the years 2014 and 2015. The Single and Average methods presented similar results in the formation of groups and different from the Complete. Evaluating the purpose of this research the Complete method and principal components analysis, together with the discriminant analysis, were considered the most appropriate methods to evaluate the genetic divergence of the castor bean crop. Lineages 18, 19 and 20 showed average grain yields above 1555 kg.ha -1 , high oil content (above 46.9%), and low size plants (below 116 cm).

Principal Component Analysis for Identification of Superior Castor Bean Hybrids

“…Despite the importance of such a tool, application of multivariate methods in castor breeding programs are lacking. One of the few studies which address the topic was Sartori, Silva, and Zanotto (2018), who investigated the efficiency of multivariate methods such as clustering methods (complete link, Euclidian distance and the nearest neighbor method) and PCA to select short-height, high-yielding castor bean cultivars. While dendrograms in clustering methods yielded contrasting results, PCA allowed to efficiently select genotypes based on all desired traits concomitantly.…”