Structural Variation Analysis with Long Single Molecule Reads

Structural Variants (SVs), which include deletions, insertions, duplications, inversions and chromosomal rearrangements, have been shown to effect organism phenotypes, including changing gene expression, increasing disease risk, and playing an important role in evolution. Still it remains challenging to detect all types of SVs from high throughput sequencing data and it is even harder to detect more complex SVs such as a duplication nested within an inversion. 

To overcome these challenges, we have been applying long third generation sequencing reads from Pacific Biosciences and Oxford Nanopore to analyze SVs in a range of samples from small microbial genomes, through mid-sized plant and animal genomes, to large mammalian genomes. The increased read lengths, which currently average over 10kbp and some approach 100kbp, allow us to span more complex SVs and accurately assess SVs in repetitive regions, two of the major limitations when using short reads. To support this analysis, we have developed a new suite of tools for detecting and analyzing SVs from long reads. First, our new read mapping algorithm NGM-LR is able to accurately align the long noisy reads, especially near SVs. Second, our algorithm Sniffles scans the alignments for split-reads and high-mismatch regions to precisely detect SV breakpoints. In complementary work, we use whole-genome alignment to detect certain types of variants directly from a de novo assembly of the long reads. Finally, our algorithm SplitThreader uses a network flow algorithm with the breakpoints and alignment coverage to reassemble the underlying genome structure, even when there are complex nested SVs.