Development of a Bioinformatics Pipeline for Detection of Genome-Wide SNPs

Our aim is establishment of a high-throughput plant breeding method by using Next-Generation Sequencing and genomic selection. With the high-throughput breeding method, we are also trying to improve the ability of biofuel production in Sorghum bicolor. To this end, we have developed a bioinformatics pipeline for detection of genome-wide single nucleotide polymorphisms (SNP) markers with large-scale short-read data. For high-throughput genotyping, we performed Restriction-site-Associated DNA Sequencing (RAD-Seq) and read mapping to identify SNPs among about 500 cultivars of Sorghum. At first, we sampled genomic DNA digested by two kinds of restriction enzymes, BglII and MseI, from the 500 cultivars. The DNA samples were sequenced by HiSeq 2000 (illumina). To correctly identify SNPs, we selected high-quality reads by preprocessing steps for raw read data. We detected genotypes in each cultivar with mapping about 855 million high-quality reads onto the genome sequences of Sorghum (JGI Sbi1) by a program BWA. With a program Stacks, 91,894 SNPs among about 500 cultivars were identified. By using the large-scale SNP information together with phenotypic data for the cultivars, genomic selection will serve an important driving force for establishment of biofuel crops in a short period.
This work is partially supported by JST, CREST and Research Funding for Computational Software Supporting Program from Meiji University.