Genetic Diversity Analysis Based on Morphological Characters and Microsatellite Markers in Sugarcane (Saccharum officinarum L.)

Abstract

Genetic diversity in sugarcane was investigated using 16 quantitative, 37 qualitative morphological traits and 23 microsatellite markers. Results showed that a moderate level of genetic diversity was present in 51 evaluated sugarcane genotypes. Agromorphological traits showed high Shannon-Weaver diversity indices (>0.80) for quantitative characters and very low (0.0) to high (>0.80) for most of the qualitative traits. Mean Euclidean distance for agromorphological quantitative traits was 87.33 between pairs of genotypes for all possible pair wise combinations, and was ranged from 6 to 251. Fourteen pairs of distantly related genotypes had Euclidean distances ≥ 200 while 37 pairs of closely related genotypes had Euclidean distance values ranging from 6 to 30.Diverse genotypes based on their mean Euclidean distance values can be utilized for parent selection in hybridization program. Integrating available information on their good combining ability with other genotypes to the phenotypic distance data, as a criterion in parental selection, ensures a higher chance of generating better performing hybrids. Thus, cross combinations between genetically closely related genotypes should be avoided. Crosses between genetically distant sugarcane genotypes might produce higher variances for quantitatively inherited traits in segregating population. Cane yield was found to be positively and significantly correlated with plant height, stalk length, number of tiller per clump, leaf width, internode diameter, number of millable cane and single cane weight. Cane yield of sugarcane could be improved by selecting sugarcane genotypes having high values of length of stalk, number of tiller per clump, internode diameter, single cane weight and number of millable cane. Principal component analysis conducted based on correlation matrix of 16 agromorphological traits resulted to five principal component axes accounted for 81.31 % of total variation. The first principal component that accounted for 28.82% of total variation was mainly attributed to variation in plant height and stalk length traits. This further indicates that plant height, stalk length, number of millable cane and leaf length were among the most important traits which accounted for 81.31 % variation expressed in the evaluated sugarcane genotypes. It could be suggested that the use of these traits will save considerable amount of time, labour and cost for identification of superior sugarcane genotypes. The positive absolute values of two vectors revealed that plant height, internode length and diameter, number of internode per cane, single cane weight and pol percent had the greatest contribution to genetic divergence. Cluster analysis by UPGMA based on Euclidean distances classified 51 genotypes in to six clusters and cluster 3 was identified as the largest cluster. Fifty one genotypes were also grouped based on agromorphological traits into six clusters using Mahalanobis D2 statistic. The highest inter-cluster distance (12.358) was found between clusters I and cluster V. The distance between cluster V and cluster VI was minimum (2.628).The crosses between genotypes in cluster I with genotypes in cluster V might produce a good hybrid which would exhibit highest heterosis. The 23 microsatellite markers revealed high gene diversity (PIC) values in the 51 sugarcane genotypes. Average PIC value was 0.942. Primer pair SMC 226 CG showed the highest PIC value (0.979) that makes it the most discriminating among 23 markers used. The level of polymorphism indicates that distinction between any two genotypes is possible with appropriate SSR primer pair. This supports to the use of SSR markers, as an excellent tool, for diversity analysis and loci mapping in sugarcane. A total of 76 unique alleles were generated by 21 SSR markers. Most of the unique allele produce markers showed high PIC value. These 21 markers distinguished 88.24% sugarcane genotypes. Only two primer pairs viz. SMC 226 CG and SMC 278 CS produced unique allele in 19 genotypes i.e. 37.25% genotypes were distinguished. Cluster analysis showed that the genotype pair POJ 2878/ I 156-97 was the closet among all genotypes having the lowest Euclidean distance value (5.66).The most distant genotype pairs were I 6- 04/ I 33-97 (8.77) and I 174-93/ I 33-97 (8.77). Among the 51 genotypes studied, genotype I 33-97 was found to be the most distant genotype with mean Euclidean distance 8.01 with other 50 genotypes. The difference between the lowest and the highest Euclidean distances indicated the presence of low to moderate level of genetic diversity among the studied sugarcane genotypes at genotypic level. Dendrogram derived by UPGMA using Euclidean distance of SSR molecular genotyping data revealed two major clusters. Most of the genotypes (37) grouped in the cluster 1 and rest 14 genotypes concentrated in the cluster 2. On the other hand, five major clusters were formed when Jaccard’s similarity coefficient was used for constructing dendrogram following UPGMA method. The clustering patterns using molecular marker data were different from that of clustering of genotypes constructed from agromorphological quantitative data. This might be due to nature of data and different clustering models used. The diversity analyzed by molecular (SSR) markers data was found to be more precise because these markers are not influenced by environmental factors. Both phenotypic and SSR profile data should be considered during the selection of clone/ genotype for conserving in the active collection of germplasm or parents for hybridization program. The knowledge obtained in this study might be useful for future breeding program for increasing genetic diversity of sugarcane germplasm to meet the demand of sustainable sugarcane production in Bangladesh. To widen the genetic base of BSRI sugarcane germplasm pool, incorporation ofS. spontaneum, S. barberi, S. robustum and Erianthus spp. into the population should be initiated. Saccharum species other than officinarum can be utilized as female parents to broaden the cytoplasmic base. Phenotypic and molecular characterization of the entire collection should be done to determine the relationships among the genotypes available at “Field Gene Bank of BSRI, Ishurdi, Bangladesh.