Abstract
In the study of genetic divergence, through cluster analysis, it is important that the measure of dissimilarity (or similarity) between genotypes allows for the grouping of genotypes, with homogeneity within the group and heterogeneity between groups. Thus, the objective here was to compare the consistency of the clustering pattern of wheat genotypes based on Mahalanobis (D2) and modified Mahalanobis (D) distances, in six clustering methods. Data on grain yield, thousand-grain weight, hectoliter weight, and grain protein obtained from 191 trials were used. The number of genotypes ranged from 14 to 44. For each trial, analyses of variance, correlation, multicollinearity diagnosis, and clustering were performed. Mahalanobis (D2) and modified Mahalanobis (D) distance matrices were determined between the genotypes. Genotype clustering was performed using the following hierarchical methods: single linkage (the nearest neighbor) (SL), complete linkage (the farthest neighbor) (CL), unweighted pair-group method using arithmetic average (UPGMA), weighted pair-group method using arithmetic average (WPGMA), unweighted pair-group method using centroids (UPGMC), and weighted pair-group method using centroids (WPGMC). The consistency of the clustering pattern was evaluated using the cophenetic correlation coefficient. The highest consistency was obtained with the UPGMA method from D. In the SL, CL, UPGMA, WPGMA, and UPGMC methods, the distance D shows greater consistency in the clustering pattern of wheat genotypes compared to D2.
Funding Information
  • Conselho Nacional de Desenvolvimento Científico e Tecnológico (304652/2017-2)
  • Conselho Nacional de Desenvolvimento Científico e Tecnológico (304878/2022-7)
  • Conselho Nacional de Desenvolvimento Científico e Tecnológico (001)
  • Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (304652/2017-2)
  • Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (304878/2022-7)
  • Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (001)