Finding genomic differences from whole-genome assemblies using SyRI

Genomic differences can range from single nucleotide differences (SNPs) to large complex structural rearrangements. Current methods typically can annotate sequence differences like SNPs and large indels accurately but do not unravel the full complexity of structural rearrangements that include inversions, translocations, and duplications. Structural rearrangements involve changes in location, orientation, or copy-number between highly similar sequences and have been reported to be associated with several biological differences between organisms. However, they are still scantly studied with sequencing technologies as it is still challenging to identify them accurately. Here I present SyRI, a novel computational method for genome-wide identification of structural differences using the pairwise comparison of whole-genome chromosome-level assemblies. SyRI uses a unique approach where it first identifies all syntenic (structurally conserved) regions between two genomes. Since all non-syntenic regions are structural rearrangements by definition, this transforms the difficult problem of rearrangement identification to a comparatively easier problem of rearrangement classification. SyRI analyses the location, orientation, and copy-number of alignments between rearranged regions and selects alignments that best represent the putative rearrangements and result in the highest total alignment score between the genomes. Next, SyRI searches for sequence differences that are distinguished for residing in syntenic or rearranged regions. This distinction is important, as rearranged regions (and sequence differences within them) do not follow Mendelian Law of Segregation and are therefore inherited differently compared to syntenic regions. Using SyRI, I successfully identified rearrangements in human, A. thaliana, yeast, fruit fly, and maize genomes. Further, I also experimentally validated 92% (108/117) of the predicted translocations in A. thaliana using a genetic approach