Interest in exploiting sweet sorghum (Sorghum bicolor (L.) Moench) as a biofuel crop is growing due to its rich stalk sugar content (Wang et al., 2009). Many breeding programs are working towards development of high-yielding varieties and hybrids with higher sugar content, resistance to diseases, drought tolerance and good agronomic traits (Klein et al., 2008). Significant breakthrough has been made in developing and releasing sorghum hybrids and varieties for commercial cultivation both in India and elsewhere (Kumar et al., 2011). Genetic similarity estimates among genotypes are important in selecting parental combinations for creating segregating populations to maintain genetic diversity in a breeding programs (Becelaere et al., 2005), develop mapping populations for detecting quantitative trait loci (QTL) (Varshney, 2011) and categorize lines into heterotic groups for hybrid crop breeding (Menz et al.,2004). Simple sequence repeats (SSR), also known as microsatellites, are based on tandem repeats of one to six core nucleotide elements. Different studies have recommended the use of SSR markers in analyses of genetic diversity due to their high degree of polymorphism (Geleta et al., 2006; Ali et al., 2008; Shehzad et al., 2009). Simple sequence repeats are co dominant markers dispersed throughout the genome, and have multiple alleles that often have conserved loci between related species (Brown et al., 1996; Schulman, 2006). The SSRs are able to discriminate among closely related individuals, and have advantage over other markers in their ability to trace pedigrees in plants (Powell et al., 1996). In sorghum, several studies have been conducted involving SSR markers either alone or in combination with other marker types (Casa et al., 2005; Ali et al., 2008; Klein et al., 2008; Murray et al., 2009). In sweet sorghum, genetic diversity has been successfully determined using amplified fragment length polymorphism (AFLP) and simple sequence repeat (SSR) markers. Due to their high values for polymorphic information content (PIC) and the Shannon diversity index (Geleta et al., 2006), these marker types were adequate for use in these species. Sweet sorghum gene pool creation has not received much attention mainly because it was not considered to be among important crops in Kenya, and the pedigree information is scarce and incomplete. The study was therefore undertaken to assay genetic diversity among a collection of sweet sorghum genotypes using SSR markers to identify specific genotypes exhibiting highest levels of polymorphism i.e. genotypes bearing a maximum of favourable alleles with the aim of improving the genetic base of the current cultivated sweet sorghum to support a breeding programme that will provide Kenyan and African farmers with high-yielding sweet sorghum varieties.

Results

Marker characterization and allele frequencies

The allele sizes among the genotypes for the 11 microsatellites varied from 98 to 244 bp (Table 1). Number of repeats in the SSR motif had strong correlation with allele number and their polymorphism information content (Table 2). The 11 SSRs revealed a total of 86 alleles with a mean of 8 alleles per marker (Table 2). The allele number ranged from 3 (mSbCIR 246) to 19 (Xtxp 012). All the markers were polymorphic (Table 2). The PIC value over the 11 SSR markers averaged 0.53, ranging from 0.09 for marker Xcup 63 to 0.89 for marker Xtxp 012. The mean level of heterozygosity per SSR marker was 0.17 (Table 2). Heterozygosity level ranged from 0.00 for marker Xtxp 145 to 0.84 for marker mSbCIR 246. Marker Xtxp 012 had the highest gene diversity of 0.9 while marker Xcup 63 had the lowest gene diversity of 0.09. The mean gene diversity per SSR marker was 0.56 (Table 2). All the loci were polymorphic with Zambian population having the highest number of private alleles (Table 3).

Table 1 The 11 SSR markers used in this study, the dyes used to label them, repeat motif, chromosome number and allele size range

 

Supplementary Table 1 86 sweet sorghum genotypes used in genetic diversity study, their genotype identification (GI), genotype name and region of collection

Table 2 Summary of allele frequency, allele number and diversity indices of 86 sweet sorghum genotypes

Table 3 Table of genetic diversity for each sweet sorghum populations analyzed in this study

Population structure

There was a clear genetic differentiation among individuals within populations, and within the individuals using significance tests based on 1,000 permutations (Table 4). The variability among the populations was low as 3 % but amounted to 22% and 75 % within individual genotypes and among individuals respectively. The F_ST value was 0.034.

Table 4 AMOVA partitioning SSR variation, among populations, among individuals within populations, and within individuals in 86 sweet sorghum genotypes

Genetic diversity within regions

The allele frequency based pair-wise genetic distances between the countries calculated using Power- Marker version 3.25 revealed the relatedness of genotypes on a country by country basis (Table 5). Genotypes from Kenya and Brazil were the most distant at 0.303 whereas genotypes from Kenya and Zambia were the closest at 0.070.

Table 5 Genetic distance matrices between countries calculated according to Nei (1987) for the 86 sweet sorghum genotypes

Cluster analysis

The pair-wise dissimilarity indices among the sweet sorghum genotypes were estimated using allelic data by simple allele matching followed by cluster analysis using unweighted neighbor-joining algorithm. All the 86 genotypes fell into two (I and II) clusters corresponding mainly to their geographic origin and pedigree. The biggest cluster, cluster I had 83 genotypes with cluster II having 3 genotypes (Z35, Z42 and K86), Figure 1.

Figure 1 Dendogram (radial axis) of 86 sweet sorghum genotypes revealed by cluster analysis of genetic similarity estimates generated by Nei coefficient based on 11 SSR markers

Discussion