关键词: 甘蔗, 种质资源, 分级数据, 广义遗传力, 主成分分析, 聚类分析

Abstract:

【Objective】Evaluation of sugarcane germplasm via rapid rating their important agronomic traits could provide reference for their further evaluation hybridization.【Method】Based on the field experiment, 317 sugarcane germplasm resources (2 replicates) were rated and evaluated by selecting five important agronomic traits at maturity stage, which were plant height, stalk diameter, millable stalks, leaf disease degree (comprehensive reflection of natural leaf diseases such as mosaic, rust, brown stripe disease) and general vigor. Each trait was divided into five grades, in which plant height was highest in grade 1, and the shortest in grade 5; similarly, stalk diameter was thickest in grade 1 and finest in grade 5; millable stalk was the most in grade 1 and the least in grade 5; general vigor was the best in grade 1 and the worst in grade 5; leaf diseases were the best in grade 1 and the worst in grade 5. The rating of each trait was accomplished by the members of the rating group. Based on the rating data of agronomic traits, 317 germplasm resources were comprehensively evaluated by statistical methods such as variance, broad-sense heritability, principal component analysis, clustering and discriminant analysis, and the germplasm resources with excellent agronomic traits were screened. Further, the sucrose content of the selected excellent germplasm was evaluated by primary rotation.【Result】The results of variance and broad-sense heritability analysis showed significant differences among 317 sugarcane germplasm in the classification data of agronomic traits (P<0.01) and the broad-sense heritability ranged from 0.61 to 0.72. The stalk diameter performed the highest broad-sense heritability (0.72) and the plant height performed the lowest (0.61). The results of principal component analysis showed that the five agronomic traits could be simplified to one principal component factor, i.e. the comprehensive superposition of five agronomic traits. The information provided by the principal component accounted for 63.57% of the total information, and the characteristic roots were significantly higher than the sum of the characteristic roots of other principal components. Based on the results of principal component analysis, 317 germplasms were classified into five categories, including 20 in the first category, 97 in the second category, 82 in the third category, 80 in the fourth category and 38 in the fifth category. 117 germplasm in the first and the second categories were the best in agronomic traits (36.91% of the total germplasms). The stepwise discriminant analysis of clustering results showed that the average probability of pairing is 95.85%. The clustering results fully reflected the actual performance of germplasm in field experiments. Based on cluster analysis of agronomic traits, the sucrose content of 117 germplasm in Class I and Class II was evaluated. The results showed that the average sucrose content of 117 germplasms ranged from 5.0% to 17.10% from November to next March, the highest was YZ09-1601 (17.1%) and the lowest was CP79-318 (5.0%). There were 43 germplasm with an average sugar content of more than 15.0%, of which 11 were more than 16.0%.【Conclusion】Plant height, stalk diameter, millable stalks, leaf disease and general vigor are important agronomic traits reflecting the field performance of sugarcane germplasm resources. The rapid classification method based on important agronomic traits is a promising alternative for evaluating large number of sugarcane germplasm resources.

Key words: sugarcane, germplasm, rating data, broad sense heritability, principal component analysis, cluster analysis