Hybrid rice breeding in Russia

Research on hybrid rice problems has been carried out for a long time in Russia, but it is mainly theoretical. Hybrid rice research faces several problems in Russia: (1) the lack of suitable restorers in japonica rice, which is mainly grown in Russian conditions, and (2) the low level of allogamy does not allow using hybrid seed for direct sowing, and (3) combining ability evaluations and screening for hybrid combinations with high productivity or heterosis. During the first period of research, we studied available CMS sources for possible use. Wild abortive (WA) cytoplasmic male sterility was transferred into our local lines. All tested elite lines appeared to be maintainers. So, several CMS lines were successfully developed and a restorer line breeding program started. Rf genes were transferred from IR54. Studying the genetics of fertility restoration showed that at least three different Rf genes are found in the japonica background: two genes described in the literature and the last one with regulation activity - the dominant allele of this gene allows other Rf genes to act. We think that indica WA-CMS lines have this dominant allele too. The two level of cross-pollination in rice is a serious problem in Russia. Only direct sowing is used in commercial rice production. Therefore, we studied floral behavior and allogamy-connected traits. Studying the inheritance of allogamy-connected traits showed that the inheritance of floral and flag-leaf traits was determined by polygenes in many cases. Additive variance prevailed in a population for all traits except for two characteristics (stigma exsertion and stigma diameter). Gene interaction (complementary epistasis) was also noticed for all traits studied. For stigma diameter, dominance was higher: for all other traits, dominance was lower. As proved by ANOVA, the number of recessive genes in varieties is correlated with the character value for all traits except for flag-leaf length. There was overdominance in the locus of floral traits. Length and breadth of the flag were partially dominant. We analyzed the correlation between all studied allogamy-connected traits and only stigma traits were closely correlated with each other. Therefore, when breeding lines with increased stigma characters, it is enough to measure only one parameter because of the high correlation between them.