TURKISH JOURNAL OF AGRICULTURE AND FORESTRY, vol.40, no.4, pp.613-620, 2016 (SCI-Expanded)
Genetic studies of watermelons [Citrullus lanatus (Thunb.) Matsum. & Nakai] sampled from Turkey, the second leading producer, may provide valuable information for breeding and research programs. The objectives of this study were to estimate the level of genetic diversity, population structure, and optimum genome sampling size among the Turkish and several introduced watermelons. From the collection, 256 watermelon lines representing all watermelon-growing areas of Turkey along with two accessions of a related genus (Praecitrullus fistulosus (Stocks) Pangalo) and a few popular cultivars were genotyped using sequence-related amplified polymorphism (SRAP) markers. Twenty-seven primers generated 210 molecular markers for genetic analyses. The unweighted pair group method with arithmetic average analysis produced no clear-cut pattern, while principal component analysis indicated three subgroups among the lines studied. They had a narrow genetic base within C. lanatus var. lanatus and were mostly distinguished from each other. Model-based structure analysis indicated that the number of subpopulations in watermelons was four. Among the 258 lines, only 20% (51 lines) had 0.80 or greater membership coefficients to one subpopulation, and therefore were not admixed. The remaining 207 were admixed by at least two subpopulations. The similarity matrix of genome sampling size of 40 randomly selected markers was highly correlated (r = 0.915) with the matrix of simple matching coefficients based on 100 markers, implying that 40 markers were detected to be sufficient for this species. Overall, this study concluded that the majority of the Turkish watermelons were distinguished from each other, difficult to classify, and admixed, with narrow genetic variation.