A PC SAS COMPUTER-PROGRAM TO GENERATE A DISSIMILARITY MATRIX FOR CLUSTER-ANALYSIS

Cluster analysis is being used increasingly in plant breeding research to reveal relationships among inbreds, hybrids, or varieties. However , few software programs popular in agricultural research feature a pro cedure for generating the dissimilarity matrix, which often serves as input to the cluster analysis. The elements of this matrix are estimat es of relationship for all possible pairs of subjects, computed using the raw data and expressed in terms of disparity or distance. A comput er program we have written in personal computer SAS code is designed t o generate a matrix of dissimilarities (distances) between all pairs o f subjects based on binary (0, 1) data recorded for each subject. The user can choose among three association coefficients and between two t ypes of distances to express dissimilarities. Data organization and pr ogram output are illustrated by example.