Rice is cultivated all around the world and shows numerous morphological and physiological differences in the form of phenotypic variations. Some of these variations have been used as genetic resources to improve rice plants so that they better satisfy human needs. Phenotypic variations are considered to be genetically controlled by the collective function of a large number of genes on rice genomes However, the genetic bases and biological functions of most of them are still unknown, which has prevented us from wider practical application of rice germplasms. We exploit useful phenotypic variations from a wide range of rice germplasms and clarify the genes involved with biological functions by aid of recent advances in genome information and technology. Also, we try to develop new breeding materials and propose more effective breeding methodologies.

For breeding of super rice varieties with useful new genes, it is important to promote allelic exchange in existing cultivars by crossing them with wild or genetically remote cultivars. However, hybrids are rarely obtained from such distant crosses due to multitude of reproductive barriers. Interestingly, we found a fertile tetraploid progeny derived from anther culture of Asian rice cultivar, O. sativa, and African rice cultivar, O. glaberrima. Using genomic and phenotypic analyses, we now aim to clarify mechanisms involved in recovery of seed set in these plants. Furthermore, we hope to establish a novel remote cross breeding strategy in rice that overcomes reproductive barriers by introduction of polyploidization and haploidization.

Nuclei that have very complex structures and various functions are the most important organelles in eukaryotic cells. Nuclear DNA are divided and packed into chromosomes, enabling the accurate transmission of genetic information to daughter cells. Our research group is studying the molecular structures and functions of nuclei and chromosomes, mainly in plants. Our most recent goal is the development of plant artificial chromosomes to elucidate chromosome functional elements: centromeres, telomeres, and replication origins. We are also interested in the relation between chromatin modifications and gene expression.

Along with drastically changing our lives, computers have also brought a paradigm shift in plant breeding. “Genomic breeding” is a state-of-the-art method which integrates genome-wide genotype information to model and predict phenotypic variation in populations. Although modern computer-based methods have strong potential for accelerating the development of new varieties, statistical models used in the method still need improvements to capture in detail events happening in real nature. At present, we try to integrate all useful environmental and biological data into new statistical models for genomic breeding that will allow accurate phenotype prediction within relatively small datasets. We apply a large variety of modeling methods that also include machine learning strategies. We hope that our work will boost up breeding in the future, helping to overcome the global shortage in food production experienced worldwide.