| Malus sieversii, the main progenitor of domesticated apple (Malus×domestica Borkh.), is rich in genetic diversity. M. sieversii is mainly distributed in Tianshan Mountains in Central Asia including Gongliu population, Xinyuan population, Huocheng population and Yumin population in China. Recently, M. sieversii is suffering serious destruction and is sharply decreasing in area. The genetic diversity and population genetic structure for four M. sieversii groups were analyzed using principles of molecular systematics and SSR marker and SRAP marker techniques and the methods of constructing core collection were studied by molecular marker and morphology traits, respectively, in order to provide science evidence for conservation and utilization, to construct core collection of M. sieversii. The main results are as follows:1. The genetic diversity of M. sieversii was analyzed using SSR marker. The results showed that the total 128 alleles were amplified by 8 pair of SSR primers on 109 M. sieversii accessions. The percentage of polymorphic loci (P = 100%), Nei's gene diversity (H = 0.2619) and Shannon information index (I = 0.4082). The genetic diversity of M. sieversii by SRAP marker showed that the total 209 bands were amplified by 10 pair of SRAP primers. The percentage of polymorphic loci (P) was 98.56%. The results showed that genetic diversity of M. sieversii was very abundant.2. The genetic diversity of four M. sieversii populations were studied using SSR marker and SRAP marker. SSR marker showed that the genetic diversity of Gongliu population was the richest in the four populations, with 113 for amplified bands, 88.28% for the percentage of polymorphic loci, 0.2538 for Nei's gene diversity and 0.3912 for Shannon information index. Followed by Huocheng population (A = 112; P = 87.5%; H = 0.2501; I = 0.388), Xinyuan population (A = 108; P = 84.38%; H = 0.245; I = 0.377), Yunmin population (A = 100; P = 78.12%; H = 0.2273; I = 0.3482). SRAP marker showed that Gongliu population (hs = 0.304) was the richest in the diversity, followed by Huocheng population (hs = 0.287), Yumin population (hs = 0.274) and Xinyuan population (hs = 0.260). On the basis of the highest genetic diversity, Gongliu population should be given a high priority consideration in M. sieversii population's in situ germplasm conservation.3. Both SSR marker and SRAP marker showed that genetic variation of M. sieversii was mainly within population. For SSR marker, genetic variation within groups accounted for 85.3% of total variations, genetic variation within population accounted for 85.3% of total variations, genetic variation between the populations accounted for 6.4% of total variations. For SRAP marker, genetic variation within population accounted for 77.9% of total variations, genetic variation between the populations accounted for 22.1% of total variations. Gene flow of GST was 7.265 according to SSR marker showed that there were partly gene exchanges among four populations. It is suggested that the main way of gene exchanges could be transferred by pollen or by seed.4. The results from UPGMA cluster analysis for four M. sieversii populations showed that the similarity between the Gongliu population and Xinyuan population was the highest (D = 0.0147; I = 0.9854), then between Huocheng population and Gongliu and Xinyuan populations, Yumnin population was the lowest with the other three population. UPGMA cluster analysis from 109 M. sieversii accessions showed all the accessions were clustered into ten groups. The accessions from the same population were clustered together, which demonstrated that the four populations were relatively independent populations, but there were partly gene exchanges5. Diagram among 109 M. sieversii from the four populations based on principal coordinates analysis showed that Huocheng and Yumin accessions had partly overlapped and Gongliu and Xinyuan accessions also had partly overlapped for SRAP marker. It is suggested that Gongliu population, Xinyuan population and Huocheng population located in Ily Valley was the primitive center of origin of M. sieversii and M. sieversii seeds and accessions of Huocheng population spread to the North and formed Yumin population. Yumin population was secondary center of origin of M. sieversii.6. For M. sieversii, SSR showed the highest range of genetic variation among the accessions and among the accessions within population, SRAP showed the highest range of genetic variation among the population. Therefore, SSR is a good choice to assess genetic variation among the accessions and among the accessions within population and SRAP is a good choice to assess genetic variation among the population.7. Mantel matrix correspondence test was used to compare the similarity matrices among SSR data, SRAP data and combined data SSR and SRAP. The results showed the correlation coefficients were statistically significant for two markers and their combination. The higher correspondence was found between SRAP and combined data (r = 0.929), which showed that analyses using SSR and SRAP data together do not seem to be the most efficient manner of assessing genetic diversity and population genetic structure of M. sieversii because the result was similar to using SRAP alone.8. The method for constructing core collection of M. sieversii based on molecular markers data was proposed, according to SSR marker, using 109 M. sieversii accessions. Compared with the random sampling strategy, allele preferred sampling could construct more representative core collections. When 25 M. sieversii accessions was selected, allele preferred sampling strategy combined with SM, Jaccard or Nei&Li genetic distances using stepwise clustering was the suitable method for constructing M. sieversii core collection.9. The genetic diversity of 10 traits from 300 M. sieversii accessions was used to study method for constructing M. sieversii core collection using morphology. The results showed that Mahalanobis distance was the much better than Euclidean distance, UPGMA, Ward's method and Complete linkage was better than Single linkage and Median method, and preferred sampling was more suitable than random sampling and deviation sampling for constructing core collection. When 20% accessions were selected, Mahalanobis distance and Ward's method using stepwise clustering combining with preferred sampling can construct a most reprehensive core collection and was the most suitable method for constructing M. sieversii core collection.
|
PDF Full Text Request