| Shanyou63 is an elite indica hybrid produced by Zhenshan 97A and Minghui 63 as female parent and male parent respectively, whose performance of agronomic characters and mechanism of heterosis have already been studied. However, most researchers resorted to the classical methods such as single-marker analysis, interval method, mixed composite interval mapping, etc. In this paper, based on a set of 241 recombinant inbred lines (RIL) population derived from Shanyou63, we applied a statistical genetic model with epistasis effect to detect QTL underlying 16 agronomic characters for further exploration of the genetic architecture of this hybrid, including plant height, heading date, yield of plant, etc. In a single model, we considered all the main effects and the epistatic effects of 221 mapped markers in 12 chromosomes covering a genome of 2070.9 cM.There were two steps to carry out the statistical strategy. Firstly, in order to get an appropriate statistical method for the population, we performed a series of simulation experiments applying the information of marker genotypes as the data set. And then the simulated data of an oversaturated genetic model including 221 main effects and 221×(221 ? 1) / 2 = 24310 epistatic effects were analyzed by five methods: E-BAYES, Stepwise, PENAL, LASSO and SSVS, respectively. In the simulation experiment, nine of the markers overlapped with QTL of the main effects and 5 out of the 24310 possible marker pairs had interaction effects. The heritability of the QTL was set at two levels: 80% and 60%. Each treatment was repeated 100 times. The principal statistical properties, including statistical power, precision and accuracy of estimates for QTL location and effects, were investigated. Secondly, based on the results of the simulation, the appropriate method for the population was applied to analyze the real data, which made us can get a deep understanding of the genetic architecture of the cross. The results of analysis were shown as follows: (1) The results of the simulation study indicated that the E-BAYES method provided accurate estimates of both the QTL effects and location with very high statistical power. The average statistical powers of QTL detected by the E-BAYES method were 97.9% and 88.14% for the two QTL heritability levels, respectively. Compared to the E-BAYES method, the other methods could not attain to such sound effects. Even though the power detected by the SSVS was best among the four, they only accounted for 25.78% and 24.71% with 80% and 60% of the QTL heritability. In addition, the E-BAYES method outperformed the others in precision and accuracy of estimates for QTL effects. Besides, only one false positive QTL was detected by the E-BAYES method while more false positive QTLs were detected by the others.(2) According to the results above, a total of 115 QTLs covering 12 rice chromosomes were found by the method of E-BAYES underlying 16 agronomic traits. Among these QTLs, there were 27 main effects and 46 pairs of interaction effects. The proportion of contribution from an individual QTL with main effects varied from 1.52% to 22.36%, whereas the proportion of contribution from a pair of QTL ranged from 1.10% to 7.08%. Concerning the main effects of two markers in 46 pairs of interaction effects, only one pair was with two main effects, 15 pairs with just one and 30 pairs with on main effect. The number of QTLs greatly varied with different traits. For some trait only one QTL could be found whereas at most 15 QTLs were detected. The proportion of contribution from the total QTLs for each trait ranged from 5.73% to 36.45%. The cumulative proportion of contribution from the main effects for each trait ranged from 1.68% to 25.66% with an average of 11.66%, while that of the epistatic effects ranged from 3.9% to 22.86% with an average of 11.70%. The results clearly indicated that epistasis was one of important genetic factors in heterosis of the superb hybrid Shanyou63.
|
PDF Full Text Request