Association Analysis Between SSR Markers And Major Agronomic Traits In Cultivated Soybean (Glycine Max L. Merr.)

Along with the full utilization of heterosis in rice and maize, it brings considerable economic benefit. However, there are many problems to explore the potential of heterosis in soybean. At present in China the soybean yield level is very low and overall soybean production situation is very serious, so its high-yield breeding needs to be further addressed. The first thing is to mine the elite alleles for yield related traits. The cultivated soybean (Glycine max L. Merr.) is an important oil crop and there are very abundant germplasm resources in China. This provides firm material foundation in the high-yield breeding of soybean. As for the utilization of germplasm resources, association analysis is a new way in the recent past years. However, the current association analysis approach is to use a single marker analysis, such as TASSEL software, with a higher false positive rate under the situation of lower significant level.In this study the molecular marker information of 135 SSR markers in 215 soybean cultivars selected from six ecological zones of China using stratified random sampling was used to describe population genetic characteristics, including population genetic structure, genetic diversity and linkage disequilibrium between two markers. The phenotypic observations of major agronomic traits measured in Jiangpu experimental station in 2008, along with the above molecular marker information, were used to carry out association analysis between SSR markers and major agronomic traits by using E-BAYES approach and TASSEL software. Based on the results of association analysis, elite alleles were mined and superior parent combinations were predicted. The results were as follows.First, the population under study was partitioned into 4,3 and 6 sub-populations by means of the STRUCTURE and PowerMarker softwares and geographic eco-types, respectively. Which one is the best? To answer the question, some parameters, such as Fst, the number of loci, the gene diversity, polymorphism information content (PIC), number of alleles, population differentiation coefficient and Hs, were calculated. As a result, maximum differences among sub-populations and minimum differences within sub-population were achieved by the STRUCTURE software. This indicates that the STRUCTURE is the best to obtain population genetic structure.135 SSR markers generated a total of 890 alleles. Average number of effective alleles per locus(Ac). PIC, gene diversity and observed heterozygosity (Ho) were 3.46,0.5140.0.5540 and 0.013. respectively. In the third subpopulation obtained by the STRUCTURE software, the largest estimates for population genetic parameters indicate that its genetic diversity is the most abundant. In the analysis of linkage disequilibrium (LD) with TASSEL software,22.49% of total marker pairs were in LD while 34.16% of marker pairs from the same chromosome and 21.88% of marker pairs from different chromosomes were in LD; D'values were mainly between 0.2-0.4 and r2 values change mainly between 0.0-0.2; and the attenuation trend of LD in the analysis of both overall population and each subpopulation was not obvious while the genetic distance between markers is more than 20 cM.Then, in the detection od QTL for plant height, node number on main stem, branch number, seed stem ratio, apparent harvest index and maturity date, a total of 38 and 177 main-effect QTL were identified by the E-BAYES approach and TASSEL software, respectively.26 common main-effect QTL between the two approaches were found. In the analysis of eBayes. the LOD values for detected QTL ranged from 2.54 to 77.36, the estimate of heritability ranged from 0.02 to 0.25, and the proportions of the total phenotypic variance explained by all detected QTL for the above traits were 18%,40%,42%,36%,39%,46% and 39%, respectively. Meanwhile.177 epistatic QTL for the above traits were identified by the E-BAYES approach. Among these interactions,40 loci were detected as main-effect QTL by TASSEL software, and these interactions explained 81.67%,59.51%,58.44%,0.00%, 53.69%,54.42% and 61.46% of the total phenotypic variances for the above traits, respectively.The allele effects of detected QTL were estimated by maximum likelihood method. Therefore, the elite alleles and its representative carrier materials were found. These novel alleles for plant height, node number on main stem, branch number, apparent harvest index and maturity date were satt440 (179 bp) and satt440 (251 bp), satt632 (305 bp) and satt160 (285 bp), sat₂67 (223 bp) and satt244 (177/165 bp), satt354 (300 bp) and satt354 (309 bp), sat₂67 (287/255 bp) and satt352 (181 bp), and satt514 (299 bp) and sat₂67 (255/231 bp), respectively; and the corresponding carrier cultivars for the above traits were Hunanqiudou No.1 and Youbian 790, Jindou No.2 and Huai Huang No.1, Hefeiliangtangjiaoshuangqingdou and Jiashanhongmao- jia, Tianandou and Qi588-8, Qinyan No.l and Changtingxidou, and Pingguohuangdou and Wenfeng No.6. respectively. For seed stem ratio, elite alleles were satt382 (295 bp) and satt382 (325 bp) in positive effect direction and satt683 (224 bp) in negative effect direction, and the corresponding carrier cultivars were Qinyan No.1, Bayueqing and Wandou No.1, respectively.Finally, based on the results of association mapping, excellent hybrid parent combinations were predicted, i.e., Weiqingdou, Rongdou21, Kexi No.8 and Wuyi-qingdou, Jiaxiangniumaohuang, Dalihuangdou may be available in the breeding of plant height, node number on main stem, branch number, stem diameter and maturity date, respectively. In these combinations, we found that some cultivars may be used to simultaneously improve multiple traits, for example, Qinyan No.1 may be available in the improvement of both seed stem ratio and apparent harvest index. These results need to be further confirmed in the future.