Establishment of Core Collection of Sweetpotato (Ipomoea Batatas L.) Germplasm Using RAPD Markers

Sweetpotato (Ipomoea batatas LAM.) is among the world’s most important, versatile and underexploited food crop after rice, wheat, maize and cassava based on fresh-weight basis in developing countries. The presence of high level of genetic diversity in germplasm of sweetpotato is reflected in the big number of accessions being maintained at various genebanks around the world. Maintenance of these germplasm collections can be a problematic and costly activity. Germplasm rationalization can facilitate maintenance and utilization of the available germplasm. Identification of core collection is one of the essential steps in rationalization. Core collection normally established based on qualitative and quantitative characteristics of morphological data. Morphological markers, however, have some disadvantages because they are often subjected or converged according to development and environmental variation / changes. Establishment of a core collection based on genetic variation using molecular markers can be advantageous due to the absence of environmental and developmental stage variation. The main objectives of this study were to use RAPD markers to investigate the genetic diversity of sweetpotato germplasm, and consequently to establish a core collection based on the RAPD polymorphism. A total of 134 accessions of sweetpotato germplasm originated from six countries namely Malaysia (52), Indonesia (35, of which 17 were from Irian Jaya), Thailand (24), Philippines (6), AVRDC-Taiwan (7) and from CIP-Peru (10) were used in the study. They were subjected to RAPD using initially 54 primers and finally 5 primers selected mainly on their ability to produce polymorphic bands. Cluster analysis was done using Jaccard’s similarity matrix based on the RAPD data. Consequently, core collection was established based on the groups revealed by the cluster analysis. The core collection was validated through three field trials with single, two and three replications. The five primers yielded 30 – 46 amplification products and the detectable DNA bands ranged between 101 – 2939 bp in size. A total of 202 reproducible bands were amplified by the five primers and they were all polymorphics. The Jaccard’s similarity index among the 134 sweetpotato accessions ranged from 0.02 to 1.00. Cluster analysis based on matrix of pairwise genetic similarity using UPGMA grouped the 134 accessions into thirteen distinct groups at 0.124 genetic similarities indicating the presence of wide genetic diversity within the germplasm. In general, many of the accessions were grouped on the basis of their geographical origin. A core collection was established by random sampling (25 %, with a minimum of one) from each of the 13 groups. A total of 34 accessions from the entire collection were incorporated into the core collection. Accessions from all the countries were represented in the core collection except those from the Philippines. Newman-Keuls test (for difference in mean) and Levene test (for homogeneity of variance) showed no significant differences between means, except for petiole length in trial with single replicate, while Levene’s test showed homogeneous variances between the entire and core collections for all characters. Mean difference percentage (MD %) and variance difference percentage (VD %) were analyzed to compare the entire collection and core collections in the three field evaluations. The MD% average value for all three trials (8.26), (4.85) and (3.19) was less than 20% indicating the core had captured all the variability in the entire collection. The VD% average value for the single replication trial (21.3%) was higher than 20%, where it revealed that the core did not represent the entire collection, while for trial with more replications (2 and 3), the core collection was shown to be representative of the entire collection (VD% 11.36 and 7.16 respectively). Coincidence rate (CR %) and variable rate (VR %) average values increased from single replication to two and three replications trials. The high average value of CR% and VR% retained in the core collection indicated that it was representative of the entire collection. The validation generated from the twelve qualitative characters showed similar trend of frequency distribution except for immature leaf colour and petiole colour between the entire and core collections. It was noted that these two classes in the two characters occurred at low frequency (2 accessions over 134) within the entire collection. The Shannon-weaver index calculated for the twelve qualitative characters showed similar average values between the entire and core collections, 0.361 and 0.364 respectively. Results of characterization and evaluation on the sweetpotato core collection showed that the coefficient of variation values of quantitative characters generally indicated high variation. Cluster analysis used on the core collection revealed the presence of five groups. However, this grouping was found to be based on morphological and no parallelism between the geographical and genetic diversity. Phenotypic correlation on characters measured showed positive phenotypic correlations between vine length and internode length and petiole length; between vine thickness and dry matter and starch, between yield and yield components (root number and mean root weight). But negative phenotypic correlation was observed between yields (its components) and dry matter and starch content. The highest broad-sense heritability estimates were revealed by starch (89.9 %), dry matter content (87.0%), vine length (83.8%), vine thickness (75.9%) and root number (75.0%). Broad-sense heritability for other characters ranged from 72.0% for internode length character to 58.5% for mean root weight which was the lowest broad-sense heritability value. Finally this study identified five accessions namely 11080, 10209, 11079, 10057, 10082, and 10236 from the core collection with potential for breeding due to their high yield, nutritional contents and favorable root flesh colour. This study has proven that random amplified polymorphic DNA (RAPD) was suitable for assessing the genetic diversity in sweetpotato germplasm, consequently useful for core collection development by being an alternative to morphological markers. Few primers (5) used was found to be enough to reveal the genetic diversity with RAPD-PCR which can reduce cost and time. The core collection established was found to be representative and captured all the genetic diversity present in the sweetpotato entire collection. This core collection can serve as a logical and efficient starting point for projects involving screening of the germplasm collection for sources of desirable alleles. It also provides a logical subset of germplasm to examine when it is not feasible to examine the entire collection. For example this sweepotato core collection can be evaluated for high yield, important nutritional element like dry matter, starch, sugar, carbohydrates and beta-carotenes