The objective of this work was to select similarity coefficients to be used among sets of cotton genotypes with low genetic diversity. Sixty-five lineages and four cotton cultivars were analyzed by RAPD and SSR markers; and the genetic similarity was estimated by seven similarity coefficients: Simple Matching, Rogers & Tanimoto, Ochiai, Hamman, Jaccard, Dice and Russel & Rao. The adequacy of the use of each coefficient to the collected data was verified by correlation between the distance matrices, the consensus index between the dendrograms and the Tocher's optimization method. The coefficient of Russel & Rao was the most divergent, and its use is not recommended. Among the parameters used to estimate the quality of information provided by each coefficient, differences were observed only by the consensus index, which established two groups: one in which simultaneous absence of bands are taken into account, and other in which it is excluded. Considering the presence of only two microsatellite alleles per polymorphic locus and the higher consensus index coefficients, the Simple Matching, Hamman and Rogers & Tanimoto coefficients should be preferred when analyzing cotton elite genotypes with low genetic similarity.
Gossypium hirsutum; genetic diversity; RAPD markers; SSR markers
