Study on the method of constructing a primary core collection of Jiangxi wild Actinidia eriantha based on fruit traits

Abstract:【Objective】Actinidia eriantha Benth. is the unique kiwifruit germplasm resource in China, andJiangxi province is one of the main distribution areas. The researches about the fruit traits and the geneticdiversity of germplasm resources of wild A. eriantha have been previously reported, but there have been al-most no relevant reports about the construction of core collection resources of A. eriantha published so far.In this study, 118 wild A. eriantha resources were collected from Nancheng and Yihuang counties in Ji-angxi province, and proper methods were determined utilizing 20 fruit phenotypic traits to construct theprimary core collection resources of wild A. eriantha.【Methods】A total of 118 germplasm were collectedfrom 3 scattered areas with similar growing environments around the Magu mountain nature protection ar-ea of Nancheng county in Jiangxi province(N27°31′46″, E116°32′52″ to N27°32′00″, E116°32′16″) Theexperiment was carried out from September 2013 to December 2014, and the experimental materials werecollected by stages(Nov. 1-Nov. 10). The standard time for collection was achieved when the soluble sol-id contents(SSC) reached 6.0%. The materials ripened and softened naturally in the laboratory at roomtemperature, then 20 fruit traits were chosen to be observed and analyzed according to DUS testing stan-dards(Actinidia). The 20 fruit traits included fruit weight, fruit shape index, edible rate, relative core size,peel strength, dry matter content, flesh color, the viscosity of pulp, peel hardness, ascorbic acid(As A), ti-tratable acid, soluble sugar, soluble solid contents(SSC), color of core, cross section of fruit shape, crosssection of core shape, fruit shape, rostral end shape, fruit shoulder shape, and fruit hair density. All thetraits were measured when fruits reached an edible state. The quality traits were assigned according to thestandards of description. The quantity trait was standardized with the distance of 2/3 standard deviation.In using a sampling ratio of 30%, take the stepwise cluster with the Euclidean distance being the geneticdistance, three sampling strategies(random sampling, deviation sampling and preferential sampling), fourkinds of system clustering methods(UPGMA, Ward's method, longest distance method and nearest dis-tance method) were compared to select the best construction strategy to further compare the constructioneffect of different sampling ratios(15%, 20%, 25%, 30%, 35% and 40%) and to determine an optimalsampling ratio. Select the coincidence rate of range(CR), variation coefficient ratio(VR), ratio of pheno-type retained(RPR), and the Shannon diversity index(H) as the evaluation parameters for the representa-tion of the evaluation of the core germplasm. The more greater the parameter values, then the core germ-plasm will be more representative of the original population genetic diversity.【Results】Fruit traits of thewild A. eriantha germplasm resources for our experiment occurred in a large degree of variations, 50% ofthe variation coefficient up to more than 30%, the fruit shape was the biggest which was up to 81.70%; thesmallest was an edible rate of 6.09%, the average coefficient of variation was 27.86%. The average valueof the Shannon information index of wild A. eriantha germplasm resources fruit traits was 1.636, more than80% of the traits diversity index was greater than 1. The shape of the cross section of the fruit was the larg-est(1.953); the smallest was the fruit pubescence density of 0.771. We chose 103 germplasm from the Ma-gu mountain core region of Nancheng county in Jiangxi province as materials, standardizing 20 phenotypictraits records to explore the method of constructing the primary core collection of wild A. eriantha. With asampling ratio of 30%, we take the Euclidean distance to be the genetic distance, and then compare threesampling strategies and four kinds of system clustering methods. From the perspective of the samplingmethod, each clustering method with 4 evaluation parameters for constructing core collection preferredsampling method was higher than the other sampling methods. From the perspective of the clusteringmethod, the 4 constructions of the core collection using the preferred sampling method were combinedwith 4 kinds of the clustering method, and the evaluation of the 4 parameters of the core collection con-structed by the average method was the highest. The Euclidean distance and group-average method wereused in the cluster analysis along with the priority sampling method to perform the sample. The sampleswere subjected to analysis at various sample proportions: 15%, 20%, 25%, 30%, 35% and 40%, respec-tively. The ratio of the retained phenotype(RPR), and the Shannon diversity index(H) from 15%-30%was significantly increased, and then remained unchanged. With the sampling proportion gradually in-creased, the coincidence rate of range(CR) showed a very significant increase from 15% to 20%, thereaf-ter, the trend showed a slowdown. The rate of change of the coefficient of variation(VR) showed a slow-down after the first rapid decrease. The basic principle of constructing a core collection was a great genet-ic representative. Therefore, the sampling ratio that has a higher Shannon diversity index should be select-ed as the priority. Considering kiwifruit is a perennial vine and its hybridization breeding program has along period, the lack of a phenotype requires at least several cycles. So, the ratio of phenotype retained al-so should be considered a priority. Above these factors, constructing a core collection from a 30% sam-pling ratio is most appropriate. In conclusion, using a 30% sampling ratio, according to the Euclidean dis-tance and group-average method to cluster analysis and priority sampling method to sample, the construc-tion of a core collection is more representative. This combination strategy is the optimum scheme of wildA. eriantha core germplasm in construction. Grouping and sieving the core distribution area of the MaguMountain and the germplasm of the area of the Muan in Nancheng county showed the optimal solution ob-tained by screening. Materials from the Yihuang county and the area of Xunxi in Nancheng county wereselected as the core collection, because the material number is less and retains more phenotypic traits.Thus obtained, the wild A. eriantha primary core germplasm consisted of 44 materials, and was used toevaluate the primary core collection constructed by using principal components analysis. The comparisonof primary core germplasm and original germplasm showed that the two eigenvalue and contribution rateswere almost the same. Comparing the geometric distribution of the samples, we found in the geometric dis-tribution of samples of the primary core germplasm repository retained the original features of geometryand distribution of the germplasm, and also many outside individual traits were retained.【Conclusion】The research based on the fruit characteristics of wild A. eriantha to construct the core collection showedthat the preferred sampling method is better than the random sampling method and the deviation samplingmethod. The group-average method is better than the square sum of deviations, the longest distance methodand the minimum distance method, and the 30% sampling ratio is the optimal sampling ratio. The principalcomponent analysis showed that the use of genetic screening of the primary core collection of 44 sampleswas able to establish the optimal sampling strategies which can effectively represent the original collection.