Abstract: 【Objective】With the development and improvement of breeding technology in indica-japonica hybrid rice, it has been possible to explore inter-subspecific heterosis directly to develop high-yielding rice varieties. It is highly significant to be able to detect gene frequencies of indica or japonica accurately, identify indica-japonica attributes and cultivar classification, and to predict competitive heterosis by genetic distance for indica-japonica hybrid rice breeding and cultivar classification using InDel molecular mark and sophisticated technology. 【Method】Twenty-four indica-japonica hybrids or japonica-japonica hybrids, recently bred by different breeding units in Zhejiang province, were collected to analyze using eighteen pairs of primers selected from the InDel molecular marker. The gene frequencies of japonica genomics were detected by the Wang’s reported method, and tested materials were judged for different attributes. Genetic distance (GD) among tested materials was assessed by InDel molecular markers using Nei’s calculation method.. The genetic similarity (GS) was assessed using UPGMA cluster analysis of genetic similarity. Principal component analyses (PCA) were conducted using SPSS 17.0 statistical analysis software to assign InDel bands, then, on the basis of the average values of the eigenvectors of the first and second principal components, a scatterplot was created. Ten agronomic important characteristics, i.e. period duration, plant height, spikelet number per panicle, seed setting, 1000-grain weight and grain weight per plots were investigated for analysis of competitive heterosis. The relationships of competitive heterosis of japonica- japonica hybridsor indica-japonica hybrids with GD values to control cultivar were determined. 【Result】Eighteen pair of InDel primers could amplify all materials to obtain special indica bands, japonica bands and indica-japonica hybrid bands. According to calculated frequencies of japonica genes, these materials were accurately judged different to have indica-japonica attributes, and were divided into six clusters, twelve hybrids belonging to indica-japonica hybrids, and another thirteen hybrids placed into japonica-japonica hybrids. Again, using cluster analysis of InDel markers, this distribution was verified. With the results of PCA analysis, we show that all varieties classify into three non-overlapping tzones, an indica zone, typical japonica -japonica zone, and a japonica cline-intermediate type zone. Comparison of the average of plot yields in two-year trials, we found that the yieldsof indica-japonica hybrids increased by 12.5% and 14.9% over that of japonica-japonica hybrids, expressing outstanding yield advantages. Upon analysis of ten agronomic characters between the two type hybrids, total spikelet numbers and fertile spikelet numbers of indica-japonica hybrids were significantly increased by about 30%. We suggest that the attribute of larger panicles is the major contribution factors to yield increases. After comparative analyses of the genetic distances (GD) relationship of tested varieties to the control Jiayou 2, we found no significant relations the group of japonica-japonica hybrids, whereas have most significant positive relations with traits of panicle length and 1000-grain weight, but has a significant negative relations with the grain density. 【Conclusion】Using InDel molecular mark technology is powerful and effective in its ability to detect gene frequencies of japonica in various types of cultivars, and to identify their indica-japonica attributes or classify their cultivar types. The indica-japonica hybrid rice expressed significant competitive heterosis in grain yield, the major contribution of heterosis was more spikelets per panicle.

Key words: Oryza sativa L., InDel, molecular marker, heterosis, indica-japonica hybrids, japonica-japonica hybrids