Comparação de coeficientes de similaridade usados em análises de agrupamento com dados de marcadores moleculares dominantes. / Comparison of similarity coefficients used in cluster analysis with dominant markers data.

AUTOR(ES)
DATA DE PUBLICAÇÃO

2002

RESUMO

With the recent advent of the molecular markers, studies of divergence and phylogenetic relationships between and within vegetable species of agricultural interest have been received greater attention. In these studies, the aim is to group similar individuals looking for bigger differences among the groups. Statistical methods of analysis such as cluster analysis, factor analysis and principal components analysis can be used in this kind of study. However, before to employ some method, the similarity matrix between genotypes must be obtained using one of the several coefficients proposed in the concerning literature. The aim of this study was to evaluate if different similarity coefficients can influence the results of cluster analysis with dominant markers. Data from 18 inbred lines of maize from two different populations, BR-105 and BR-106, were analyzed by AFLP and RAPD markers and eight similarity coefficients (Jaccard, Sorensen-Dice, Anderberg, Ochiai, Simple-matching, Rogers and Tanimoto, Ochiai II and Russel and Rao) were obtained. The similarity matrices were compared by Pearson s and Spearman s correlations, cluster analysis (with dendrograms, correlations, distortion and stress between the similarity and cofenetical matrices, consensus fork index between all pairs of dendrograms), Tocher´s optimization procedure and with the projection in two-dimensional space of the similarity matrices. The results showed that for almost all of the methodologies and both markers, the coefficients of Jaccard, Sorensen-Dice, Anderberg and Ochiai, gave similar results, due to the fact that all of them excludes negative co-occurences. It was also observed that the Simple Matching, Rogers and Tanimoto, and Ochiai II, probably due to the fact of all including the negative co-occurences. The Russel and Rao coefficient presented results very different from the others, because it excludes the negative co-occurences in the numerator and include it in the denominator of its expression, which is a reason for not recommending it. Due the fact of the negative co-occurences does not mean, necessarily, that the regions of the DNA are identical, it is suggested to choose one those coefficients that do not include it.

ASSUNTO(S)

molecular genetics of plants milho maize estatística aplicada genética molecular vegetal statistical cluster analysis análise de conglomerados

ACESSO AO ARTIGO

Documentos Relacionados