[Genetic diversity detected by DNA markers and phenotypes in Upland cotton]

Genetic diversity of Upland cotton (Gossypium hirsutum L.) cultivars was evaluated by SSR, ISSR, and RAPD. The genomic DNAs of 36 domestic and exotic cultivars were amplified with 81 SSR primer pairs; 7 ISSR primers and 53 RAPD primers, which yielded 282 polymorphic bands. The range of pair similarity coefficient (Nei, M. And W. Li 1979) between cultivars is from 0.5745 to 0.9291. The result indicated that although the present cultivars have genetic variation to a certain extent, the genetic basis of them is relatively narrow, and it is essential to find some new germplasm for the development of new cultivars. As expected, the classification of the cotton cultivars using the Unweighted Pair-group Method with Arithmetic Average based on the pair similarity coefficients is mostly consistent with known pedigree information and breeding experience. The 36 cotton cultivars have been mainly classed into exotic cultivar group, Chinese Xinjiang ecological region cultivar group, early-mature domestic cutivar group and mediate-mature domestic cultivar group. Carmen and Colorado cultivars from Spain have the lowest similarity with other cultivars, while TM-1 (genetic standard line), MD5lne and PD9364 cultivars from American have higher similarity with domestic cultivars. The reason may be that the parents of the domestic cultivars were mainly introduced from America. However, clusters of domestic cultivars are not all in accordance with their ecological regions, which may result from the pedigree complexity of the cultivars. Meanwhile the genetic diversity of 36 cultivars was investigated at phenotypic level with agronomic and fiber quality traits in the field experiments in 1999 and 2000, and was measured by Euclidean distance on the genotypes. Although grouping result of cluster analysis based on the Euclidean distance is similar than that of molecular data with a significant negative correlation of -0.3350 between the pair similarity coefficients and the Euclidean distance, the subgroup classification of the two methods gives different result. The molecular markers can be effectively applied in classifying cultivars within the group. The study provides useful information for the selection of parents in the development of cotton cultivars and hybrids as well.