Analysis Of Genetic Diversity Of The Germplasm Materials In Brassica Rapa

The species of Brassica rapa (Brassica campestris L. syn. Brassica rapa L.) belongs toBrassica genus in the Cruciferae family. The genomic composition of the species is AA(2n=2X=20) and this species has a high rate of natural cross-pollination. There are variouscultivar groups within Brassica rapa, which widely distributed in East Asia, IndiaPeninsula and many countries in Europe. Because of their global economic importance,many countries have started to collect its germplasm resources and focused on the relevantresearches of their genetic diversity. However, in many studies, it is still controversial todetermine the origin, evolutionary history, botanical classification and name assorting forthis species.In this study a suitable PCR system was constructed by using Touchdown PCR,improved rapid silver staining and PCR optimization. Based on the PCR system, 15specific SSR primer pairs were screened from the selected 28 primers and used for thePCR amplification of 83 experimental cultivars. The dendrogram of UPGMA wasproduced according to the selected 11 primer pairs which had stable patterns ofamplification. Meanwhile, combined with some data presented by Wageningen Universityand Structure2.0 computational package, the population structure of the experimentalmaterials was taken into further analyze genetic background. In addition, this study alsoevaluated the natural variation of the morphological traits among the different leaves fromvarious cultivars by using principal component analysis.The main results suggested that in 15 specific primers, the PIC values of the 13 primerpairs ranged from 0.319 to 0.787, with the average value 0.513. Each of the specificprimers could detect 1～6 alleles and the primer M8 had multi-amplified products. Theresult of UPGMA clustering showed that the selected 11 specific primer pairs coulddiscriminate the majority of materials. The genetic similarity coefficient was 0.58～0.95among the materials, and at threshold 0.58, all materials could be distinguished into two major groups: in the group I the variety of most cultivars were Chinese cabbage, whosegeographical origins were from East Asia. And the groupⅡwas a mixed cluster containedmany different types of cultivars apparently, whose origins involved East Asia, IndiaPeninsula and many countries in Europe. In this group, the cultivars of European turnipwere clustered into one classification more together. The varieties Neep greens and Mizunahad a more complex genetic background.Secondly, in the analysis of population structure, the experimental materials could bedivided into five subgroups. The main varieties in subgroup S1 were European turnip andBroccoletto and subgroups S3 and S4 mainly included Chinese cabbage, Pak choi and allcultivars of Wuta-tsai. Winter turnip rape's were classified into the last subgroup S5. All ofsubgroups had a similar range of variation. As experimental results suggested, there was noobvious genetic relationship between varieties Wuta-tsai and Pak choi or Chinese cabbage,Pak choi possibly had a close genetic relationship with Taicai. Furthermore, theevolutionary history of Italian Broccoletto was likely associated with European turnip orturnip rape.Finally, the different types of Brassica rapa usually had apparent morphologicalvariations. The leaf traits and seed coat color of the materials were used for clustering. Theanalysis showed that when the similarity coefficient was 0.48, the materials could bedivided into three groups. The range of the coefficient variation became greater.Meanwhile, this clustering result partly fitted to the SSR cluster's. Based on the results, thecultivars of Chinese cabbage had the same morphology and Broccoletto had a conversesituation. European and Chinese turnip had phenotypic difference, the cultivars of Mizunawere extremely close in morphology. Principal component analysis showed that everyvariability rate in the former four factors were greater than 10% and their cumulativecontribution rate were greater than 65% in 11 character variables, however, their respectivevariability rate was not good enough to represent the main trait variations of the materials.Otherwise, some variables had obvious relationships. The clustering of the former fourmain trait factors could better represent the clustering result of all traits from theexperimental cultivars.