Multiple Alignment of Optical Mapping Facilitates Typing in Complex Regions

Structural variations (SV) play key roles in genetic diversity in evolution and in disease. Optical mapping has proven useful in improving genome assembly and SV detection, by offering reads lengths on hundred-kilobase scale, of sequence-motif-dependent signal patterns.

Long-range information from optical mapping allows easy identification of insertions and deletions, gap sizing, and contig scaffolding. However, while de novooptical map assembly is available, downstream analysis usually involves a reference alignment step, which makes comparison between dozens of samples difficult, especially for genomic regions frequently confounded by repetitive elements, inversions or other large-scale rearrangements.

We introduce a systematic analysis using reference-free method based on multiple alignment of optical map contigs. Segments of DNA molecules with signal patterns conserved among samples are grouped as collinear blocks. Samples with higher similarity as sharing more conserved regions are clustered together, providing hints for phylogeny. Blocks contributing to most variations between sample clusters are found by statistical analysis. Further annotation reveals the identities of these blocks as the most representative features.

We have demonstrated that this method is useful in typing samples with high degree of genomic rearrangement, and has a high resolving power in sample clustering. We anticipate this method to facilitate comparative genomics work on non-model animal and plants.