Prediction of Genome Structural Variations Between G. soja and G. max Using Both Unmapped Reads Generated By Illumina Hi-Seq

Genomic differences were observed between cultivated soybean (Glycine max L. Merr.) and wild soybean (Glycine soja Sieb. and Zucc.) which might be related to domestication syndrome. For improving the detection of the structural variations after re-sequencing of Glycine soja accessions, we retrieved the paired-end reads that were not aligned against reference genome, Williams 82, and de novo assembled them. The resulting assemblies were regarded as putative sample specific contigs that are absent in reference genome. The contents of genes on the assembled contigs were predicted and annotated using software geneid and BLAST. The predicted gene contents on assembled contigs were annotated as it follows: small-organs related proteins, transposable elements and environmental tolerance related proteins. We designed eight primers among those contigs and tested in 46 G. soja and 58 G. max accessions including reference cultivar, Williams 82 to validate the predictions. 4 primer sets were not amplified in DNA of Williams 82 while the other 4 primer sets were amplified; hence the contigs derived from unmapped reads are likely to represent the large deletion or insertion as well as un-sequenced regions of reference genome. These primer sets showed high polymorphism information contents (PIC), average value of 0.44 and 0.36 for G. soja and G. max, respectively, suggesting the usefulness of these markers for the breeding program. Especially, the primer set which show high PIC value in G. soja and low PIC value in G. max could be molecular marker for introduction of G. soja abundant useful gene.