Deployment of SSR markers for assessment of genetic diversity has been widely adopted for screening of taro germplasm (Verma et al. 2023). Genetic diversity studies of the accessions were carried out with four SSR markers. A total of 38 scorable alleles were generated with a range of 3–15 alleles per marker and an average of 9.50. This results is in agreement with Khatemenla et al. (2019), who recorded 61 scorable amplicons using ten SSR markers subjected to 22 cultivars of taro and average of 6.1. The genetic variation analysis was performed by estimating the allele frequency, gene diversity, observed heterozygosity, fixation index and the polymorphic information content (PIC) for each locus. The allele frequency expressed as a percentage or decimal describes how often an allele occurs in a population. Knowledge of the allele frequencies can be used to measure the expected heterozygosity as well as the PIC of the markers (Kumar et al., 2020).
The markers XUQTEM 73 recorded the lowest number of the major allele occurrence in the population (0.23) and the highest of 0.69 by XUQTEM 97 (Table 4). According to Otoo et al. (2015), the number of alleles recorded and its frequency can be used to measure PIC, thus for this study, the highest allele number (15), recorded the highest PIC (0.89) and highest gene diversity (0.92).
Since PIC measures the efficiency of the markers, markers which recorded a PIC value greater than 0.50 (Table 4) is an indication that those markers have a better discriminatory power. A high significant association was found between the number of alleles, PIC and gene diversity.
Again, information on the allele frequencies of loci can also be used to measure the expected heterozygosity. Expected heterozygosity or gene diversity is required to ascertain the informativeness of a locus. In this present study, high gene diversity was recorded (0.59 to 0.92) with a mean of 0.79, indicating a high genetic variability among the accessions. Obidiegwu et al. (2009) in their study of determining the diversity of guinea yam landraces with 13 SSR markers, recorded gene diversity values ranging from 0.24 to 0.77. Again, Mulualem et al. (2018) in their study of genetic variation of Ethiopian yam landrace collection using SSR, observed a high level of gene diversity ranging from 0.03–0.71. The high gene diversity of 0.59–0.92 reported for this study can be attributed to the outcrossing nature of taro.
PIC is a quantitative measure of the informativeness of a marker (Torrelo et al., 1998), and it can be used to differentiate makers within a population. The PIC values are useful for genetic mapping and linkage analysis (Elston, 2005). The value of PIC is affected by several factors, such as the breeding behaviour of the species, genetic diversity, sample size, sensitivity of the genotyping method, and the locations of markers in the genome (Chen et al., 2017). A high PIC mean of 0.73 was recorded with a range of 0.44–0.89, thus an indicator of the SSR markers being efficient in differentiating the accessions. Avval (2017) also stated that the bigger the PIC, the greater the variability among alleles within the study. Khatemenla et al. (2019) also found the SSR markers to be adequately discriminatory in the diversity studies of taro with a PIC of 0.41–0.82 and a mean of 0.68.
The unweighted pair group method with arithmetic mean algorithm (UPGMA) and cluster analysis was used in generating the dendrogram (Fig. 3). Eight distinct clusters (A, B, C, D, E, F, G and H) were created according to the genetic distance. Cluster B grouped KNUST 14 and 17 at a distance of 0.50, an indication of how closely related they are. Again, KNUST 15 and 20, 21 and 22 and 1 and 24 also grouped together at a distance of 0.50 in clusters E, G and H respectively, thus the accessions been 50% similar. KNUST 16 and 2 appeared as outliers as they clustered individually (Fig. 3). The dendrogram generated depicts how the 26 accessions of taro are related. It also to provide useful information to breeders, germplasm curators, as well as scientists and researchers in the field of agriculture and related disciplines. An inter-cluster comparison shows that a cross between two clusters will reduce inbreeding depression and also increase heterosis. This can be achieved by crossing KNUST 1 and KNUST 3. Also, a cross between KNUST 5 and KNUST 15 will yield offspring with heterotic ability. This was supported by the statement made by Trimanto (2011) that information about the genetic diversity of taro is needed for plant breeding and improvement to obtain superior varieties. No two accessions analysed in the present study showed 100% similarity coefficient thereby indicating their distinctiveness (Table S2). This observed results indicated a low level of genetic similarity between the accessions, which suggests a high probability of identifying unique accessions for breeding and conservational purposes. Generally, a fairly high level of diversity was observed in the study, with a similarity coefficient ranging from 0.50 to 0.10.
This observation was in agreement with the results of previous studies carried out by Mabhaudi and Modi, (2013), who characterized three landraces of taro and obtained a fairly high level of diversity (0.65-1.0) among the taro landraces. In addition, Furini and Wunder, (2004) and Polignano et al. (2010), attributed the fairly high levels of variations within the accessions to the fact that analysis of DNA variability is dependent on the type and number of markers used.
Singh et al. (2013) indicated that in order to obtain the utmost information from the data generated from molecular markers, a combination of ordination methods (PCoA) and cluster analysis can be used. This combination can be used to confirm groupings established by a phylogenetic tree to ascertain the true genetic relatedness among accessions considered for a study.
The PCoA plot clustered the accessions into four groups. Comparable to the cluster analysis, PCoA grouped accessions with 50% similarity together. The grouping pattern of the accession s in the PCoA plot indicates a fairly high level of diversity among them.
In practical terms, the study conducted via molecular characterization using SSR revealed that a cross between genetically distant related accessions such as KNUST 15 and 16 and KNUST 1 and 12 are highly recommended for future breeding programs. The data obtained from the study can also assist germplasm curators to select a core collection to be maintained at the University for future use.