Abstract: Seven genetic similarity coefficients were selected to calculate pairwise genetic similarity of rose (Rosa) varieties based on microsatellite data, and the 7 corresponding dendrograms were constructed by Unweighted Pair Group Method with Arithmetic Mean. The applicability of different genetic similarity coefficients in analyzing genetic relationships based on the microsatellite in rose varieties was investigated by means of correlation analysis between genetic similar matrixes, consistency analysis between cluster trees, and test of the goodness of fit. The result showed that correlation coefficients between different similar matrixes ranged from 0.726 to 1.000. Cophenetic correlation coefficients ranged from 0.85 to 0.95, indicating that there was a good representation of similarity matrixes in the form of dendrograms. Index of CI_c between pairwise dendrograms ranged from 0.468 to 1.000, which indicated that dendrograms were dependent on selection of different genetic similarity coefficients. STRESS values among the 7 genetic similarity coefficients ranged from 16.24% to 29.90%. The STRESS values of Simple Matching, Roger and Tanimoto, and Hamann coefficients were all more than 20%. Comprehensive consideration of molecular marker characteristics, species heterozygosis, STRESS value, and combined with clustering result and variety lineage, Dice coefficient and Jaccard coefficient were found to be the most appropriate for rose variety genetic analysis, followed by Simple Matching coefficient.

Key words: Rosa, variety, polyploidy, microsatellite, genetic similarity coefficient, cluster analysis