Abstract
We have developed fifty thousands insertion lines using the endogenous retrotransposon Tos17. Large scale phenotyping for mutant lines shows that half of mutant lines have at least one phenotypic change. However, co-segregation analysis between genotype of Tos17 and phenotype indicates relatively low efficient for tagging by Tos17. We have been analyzing the whole genome sequence of regenerated rice plants from last year to detect cause of the somaclonal variation. We sequenced two lines of regenerated rice at M4 stage using a next-generation sequencer (Solexa). Coverage of each line is 5-times of rice genome. In the initial read data of the sequencer, about 13% of read have at lease one mismatch to the corresponding rice genomic sequence. Almost mismatches are read error of the sequencer or miss-alignment to repetitive loci. To obtain true mutation, we have developed filter program for Solexa reads. Finally, we obtained from 100 to 200 mutations in regenerated plant. At least, 1 mutation / Mb in the regenerated plant at M1 stage is estimated. Improvement of our method enables direct identification of mutation or mapping of target loci for phenotype.
