Genomic Resequencing and Copy Number Determination by Digital PCR Reveal Varying Size Deletions in a Repetitive Region Controlling Seed Color in Soybean

Determining structural variation in repetitive regions is not straightforward. Here, we determined the extent of naturally occurring deletions at the I locus which specifies seed color. The dominant ii allele is composed of an inverted-repeat cluster of chalcone synthase (CHS) genes whose unique arrangement generates short-interfering RNAs (siRNAs) that downregulate target CHS genes resulting in yellow seed coats. Both versions of the Williams reference genome (Wm82.a1 and Wm82.a2) have inversions or gaps and do not accurately represent this repetitive CHS gene region on chromosome 8. Therefore, we aligned NGS data to the reference genome which was improved at the I locus using the sequence of a BAC. We observed that black-seeded mutant lines with the recessive i allele have at least two types of deletions. The first type spanned 122 kb and included 19 coding regions, six of which are deleted CHS genes as confirmed by digital PCR. The large deletion appears to be a homologous recombination event between one of the CHS genes in the CHS I cluster region and a repeated CHS gene in a region 122 kb away. The second type was a shorter 17 kb deletion that appeared to be a homologous recombination event between the CHS genes within the CHS I cluster region. Both deletion types destroy part of the 27-kb inverted repeat CHS cluster resulting in lack of CHS siRNAs which leads to black seed coats. Our study demonstrates the importance of correct representation of the target region for determining structural variation.