A set of 148 F9 recombinant inbred lines (RILs) was developed from the cross of an indica cultivar 93-11 and japonica cultivar DT713, showing strong F1 heterosis. Subsequently, two backcross F1 (BCF1) populations were constructed by backcrossing these 148 RILs to two parents, 93-11 and DT713. These three related populations (281BCF1 lines, 148 RILs) were phenotyped for six yield-related traits in two locations. Significant inbreeding depression was detected in the population of RILS and a high level of heterosis was observed in the two BCF1 populations. A total of 42 main-effect quantitative trait loci (M-QTLs) and 109 epistatic effect QTL pairs (E-QTLs) were detected in the three related populations using the mixed model approach. By comparing the genetic effects of these QTLs detected in the RILs, BCF1 performance and mid parental heterosis (HMP), we found that, in both BCF1 populations, the QTLs detected could be classified into two predominant types: additive and over dominant loci, which indicated that the additive and over-dominant effect were more important than complete or partially dominance for M-QTLs and E-QTLs. Further, we found that the E-QTLs detected collectively explained a larger portion of the total phenotypic variation than the M-QTLs in both RILs and BCF1 populations. All of these results suggest that additive and over dominance resulting from epistatic loci might be the primary genetic basis of heterosis in rice.
Luo XJ, Fu YC, Zhang PJ, Wu S, Tian F, Liu JY, Zhu ZF, Yang JS, Sun CQ (2009) Additive and over-dominant effects resulting from epistatic loci are the primary genetic basis of heterosis in rice. J. Integr. Plant Biol. 51(4), 393-408.