Soybean (Glycine max(L) Merr.) is the most important economic crops, originating from China.Soybean provides vegetable oil and protein for people and livestock. At present, the soybean varieties with high oil or protein content can not satisfy the demand as food or diet in China, which can be alleviated the urgent need by breed more soybean varieties with high oil or protein content than before. Molecular assistant selection (MAS) can facilitate it. To date, many studies on the genetics of oil and protein content, QTL mapping, GE effect and epistatic effect between QTL loci can help improve the quality of soybean seed. However, most results about genetics of oil content and protein content of soybean was not consistent, indicating that the oil and protein content of soybean was sensitive to environmental change. Many soybean varieties were introduced to analyze the effect of genetic background and planted in different locations and different years to clarify the environment effect. However, the GE effect and epistatic effect between QTL loci were neglected as the limitation of statistic method and no help of analysis software, nevertheless, which affected the phenotypic variation with a higher proportion. So it is necessary to clarify the effect of environmental change to GE and epistatic effect that can improve the efforts of MAS. To date, no research was reported to analyze the GE effect and epistatic effect in three years and three locations simultaneously.The population of 147 recombination inbred lines (RIL) derived from the cross of America cultivar Charleston and Chinese variety Dongnong 594 was planted in Harbin, Hongxinglong and Jiamusi in the year of 2007,2008 and 2009 simultaneously to analyze the G×E effect and epistatic effect; As the yield of oil or protein changed significantly resulted in the environmental change, QTL controlled the stability of oil and protein content was mapped using transforming method of shukla ANOVO with data of oil or protein content. The software MGDH1-4 with mixed model approach was used to analyze genetic models with data from parent 1, parent 2 and RIL population. Software QTL Maper version 1.6 was used to analyze the main effect, epistatic effect and GE effect.In current study, the genetic model of oil content of different year and location was observed as follows:multiple gene model in location 1 (Hong xinglong),2 (Jia musi) in 2007, and main genes of three pairs with multiple gene model in location 2,3 (Harbin) in 2007 and 2008, and main genes of two pairs with multiple gene model in location 1 and 3 in 2009; The genetic model of protein content of different year and location was observed as follows:multiple gene model in location 1,3 in 2007, and main genes of two pairs with multiple gene model in locations and years except location 1,3 in 2007.QTL controlling oil content and protein content of soybean was estimated by CIM and MIM which is supplementary and verified each other. Thirty-nine QTL controlling oil content were identified, which located sixteen linkage groups across 2007.2008 and 2009 in three locations. Fifteen QTL clustered to three linkage groups, B1, C2 and G, which R2 was small and no repeat in either years or locations; Twenty-three QTL controlling protein content were identified, which located fourteen linkage groups across 2007,2008 and 2009 in three locations. Twelve QTL clustered to two linkage groups, C2 and Dla, which R2 was small and no repeat in either years or locations.G×E effect was observed both for the QTL of oil content and protein content. Such as, qOIL12-1can be identified in location 3 in 2007 and in location 1 in 2009 by CIM and MIM, and in location 3 in 2009 by MIM; qOIL2-1 can be identified by CIM and MIM across 3 location in 2009. Similar result can be observed in protein.Epistatic effect can be observed by ANOVA and cumulative distribution method both oil content and protein content and interaction between epistatic effect and environment was observed. Main effect, epistatic effect and G×E were analyzed using the software QTL Maper 1.6 for the data of one location across three years, of one year across three location and three location across three years. QTL main effect can be identified across every year or every location with low consistency, which resulted from significantly environmental change. R2 of main QTL was high than epistatic QTL and much higher than interaction QTL between epistatic effect and environment. Therefore, it is necessary to analyze main QTL and epistatic QTL simultaneously for MAS.The stability of oil content and protein content was estimated using shukla ANOVO results. E-1-1 was estimated as the optimally genetic model for stability of oil content and E-1-0 was estimated as the optimally genetic model for stability of protein content. Four QTL, qOIL7-9,qOIL8-3,qOIL15-2 and qOIL12-2 located linkage group Dla,Dlb,J and O respectively and responsible for R2,5.33%,14.61%,5.83% and 5.37% respectively, were identified for stability of oil content by CIM. QTL with the highest LOD was responsible for the highest R2 (14.61%). Two QTL, qPRO1-6 and qPRO17-2, were identified for protein content, located on linkage group Al and L respectively, and accounted for 4.70% and 5.73% of the phenotyping variation. Moreover, the chromosome region controlled the stability of oil content and protein content also possess genes controlled other traits, which region can regulate the metabolic pathway more than other regions to adapt environmental change. The current study provides theoretical evidence for fine QTL mapping and MAS. |