Welcome to Chinese Agricultural Science Bulletin,

Chinese Agricultural Science Bulletin ›› 2019, Vol. 35 ›› Issue (2): 14-20.doi: 10.11924/j.issn.1000-6850.casb18020050

Special Issue: 马铃薯

Previous Articles     Next Articles

Method of Sweet Potato Core Collection Construction

  

  • Received:2018-02-09 Revised:2018-12-20 Accepted:2018-04-04 Online:2019-01-11 Published:2019-01-11

Abstract: The National Germplasm Guangzhou Sweet Potato Nursery Genebank is the only national germplasm nursery with outdoor vegetative conservation in China. Based on 1200 sweetpotato germplasms, we did dimensionality reduction on 22 phenotypic traits by factor analysis and R-cluster. Three kinds of Q-cluster methods including Two Step Cluster, K- Means Cluster and Hierarchieal Cluster were combined with five geographical resources to group germplasms. The core collections were constructed by combing randomly sampled and required germplasms. According to the above scheme, sample database with 15 sweetpotato core collections were constructed. The representativeness of quantitative traits in the 15 construction schemes was evaluated by the coefficient of variation, standard deviation conformity and extreme difference conformity. The representativeness of quality traits was evaluated with Shannon- Weaver index and variance of phenotypic frequency(χ2) test. The results indicated that using R-clustering for dimensionality reduction, Two Step Cluster combined with geographical resources for grouping, and random sampling within the group was the best approach to construct the sweetpotato core collection. By using this strategy, the genetic diversity and variation in the constructed core collection are well presented in the total resources.