Genntic Diversity And Core Collection Of Wild Castor-oil Plant In South China

The genetic diversity of wild castor materials in south China were analyzed its core collection were initially constructed in this paper.1. The data of genotypic value of 13 traits, 29 SRAP markers and 30 SSR markers were used to detect the genetic diversity of 3 populations composed of 194, 258 and 258 wild castor materials in south China respectively. The results were as follows:(1) The average genetic distance of three clusterings by UPGMA based on the genotypic value data, SRAP data and SSR data were 7.144, 9.249 and 6.519 respectively and the average Shannon index was 7.486, 0.428 and 0.287 respectively, which suggested that the SRAP data could reveal the genetic diversity of population more efficiently and the genetypic value could more fully reflect the degree of variation of population.(2) Although the average number of alleles detected by SRAP markers was almost equal to SSR markers (1.996 and 2 respectively), the number of average effective alleles detected by SRAP markers was higher than SSR markers (1.367 and 1.182 respectively), which indicated that the alleles detected by SRAP markers was more well-distributed in the population. The average Nei's gene diversity value and the average Shannon index of SRAP markers (0.428, 0.287) were higher than those of SSR markers (0.263, 0.153), which indicated that SRAP markers could reflect the genetic differentiation level and the genetic variation level of the population more efficiently.(3) Three clustering results showed that the genetic relationship between materials was little related to geographic distance, but materials from the same sub-region had closer genetic relationship.(4) The genetic diversity of wild castor-oil plant was most abundant in Guangdong, followed by Guangxi and lowest in Hainan.2. Based on the mixed linear model, the variance of random effects were estimated by MINQUE (Minimum Norm Quadratic Unbiased Estimation) and the genotypic values of quantity characters were predicted by AUP (Adjusted Unbiased Prediction). After clustering by UPGMA method with genotypic value data of quantity characters, 6 primary core subsets were built with LDSS (Least Distance Stepwise Sampling) method at the sampling ratios of 5%, 10%, 15%, 20%, 25% and 30%, from which the primary core collection were selected out according to the results of representativity evaluation. On the basis of the primary core collection, the candidate core subsets was built with 6 compression methods ( genotypic values, SRAP, SSR, genotypic value + SRAP, genotypic value + SSR, genotypic value + SRAP + SSR) combining with 3 sampling strategies (LDSRS (Least Distance Stepwise Random Sampling), LDSPS (Least Distance Stepwise Preferred Sampling), LDSDS (Least Distance Stepwise Deviation Sampling)), then the candidate core collections were determined by the evaluation to the above candidate core subsets , so did the core collection from the candidate core collections. The results were as follows:(1) The most suitable sampling ratio was 25% for constructing the primary core collection.(2) The average Shannon ( I ) index of primary core collection reached 72.96% of the original population. The average Simpson ( H ) index of primary core collection reached 98.49% of the original population. Not significant at F values and t values, all the correlation coefficients of quantity traits variance, mean and Rj between the primary core collection and the original population were over 0.98, the correlation coefficients of CV was 0.5496, the mean difference rate was 0.00, the variance difference percentage was 100%, the Rj coincidence rate was 94.39%, the change rate of the coefficient of variation was 120.59%, the primary core collection was considered to be a good representation of the original population.(3) The compression method "genotype value + SRAP + SSR"combining with LDSPS was the optimal way for core collection construction.(4) The core collection of wild castor-oil plants of 194 in south China was initially constructed upon the results of evaluation parameters and genetic diversity analysis, it was composed of 12 materials, namely HN-HK10, HN-LV55, HN-LV58, GX21, GX29, GX73, DHD66, ZJ95, LZ23, LZ36, YJ2 and Y1, accounting for 6.19% of original population.(5) The average Shannon ( I ) index of primary core collection was 0.648. The average Simpson ( H ) index of primary core collection was 2.013. Not significant at F values and t values, all the correlation coefficients of quantity traits variance, mean and Rj between the primary core collection and the original population were over 0.99, the correlation coefficients of CV was 0.8946, the mean difference rate was 0.00, the variance difference percentage was 100%, the Rj coincidence rate was 91.37%, the change rate of the coefficient of variation was 120.51%, the core collection was considered to be a good representation of the original population.