Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
The use of multivariate analysis in developing subsets of a Caribbean collection of maize
Abstract
The objectives of the present study were to explore methodologies which could be used to develop a subset of accessions that represent the variability of the total collection, and to identify useful accessions for breeding purposes. Data of two sets of the Caribbean collection of maize, characterized by agronomic and morphological traits were analyzed with multivariate techniques. Cluster analysis and principal component analysis (PCA) were utilized to define similarity groups. Average euclidean distance squared was used to measure dissimilarity among accessions. The coefficient of association matrix was used to produce phenograms using the average linkage clustering method. Groups were created using stratified random sampling in proportion to number of accessions. A subset of 48 accessions (11%) was defined as a representative sample based upon variability attributable to phenetic relationships. Identification of useful accessions appropriate for breeding subset was defined using three-mode PCA on six traits. Multi-trait, multi-location selection was made using three-mode PCA because data on accessions, attributes and environments was analyzed jointly. Components of two modes were displayed for each component of a third mode using joint plot diagrams. Accessions were selected by inspection of joint plots using two criteria: In the first case, attributes were the basis for selecting 26 accessions. A second group of 25 accessions was identified based on performance across environments. These two groups raised a subset of 47 accessions, 11% of the total evaluated. The phenotypic diversity subsets represent a sample of the variability considering phenetic relationships, whereas, the breeding subset is a sample represented on basis of specific combination of attributes, and averaged over environments. A combination of these three subsets would contribute to development of a more appropriate representation of the variability within the collection.
Subject Area
Agronomy
Recommended Citation
Rincon-Sanchez, Froylan, "The use of multivariate analysis in developing subsets of a Caribbean collection of maize" (1996). ETD collection for University of Nebraska-Lincoln. AAI9623637.
https://digitalcommons.unl.edu/dissertations/AAI9623637