Anyone know how Snowshoes-FTD finds junction-spanning reads?

[thondeboer]

Hi,

I am trying to figure out how the gene fusion program Snowshoes-FTD finds junction spanning reads. I have the article from NAR, but it does not mention how this is done. In Figure 1, it simply states that it aligns the not-mapped reads (not mapped with BWA to either genome or their special exon-exon junction DB) to both fusion partners.

This seems to be the most important part of a gene fusion procedure, since this will actually get you the breakpoint, but the paper does not seem to discuss at all how it does this. I assume some dynamic programming algorithm, but I can't tell.

I don't have access to the program, since it is not freely available so maybe someone here knows?

Thanks

Thon

[louis7781x]

First,
It used discordant paired-end to find fusion candidates.
Example: |---geneA---|----------------------|---geneB---|

Second,
All unmapped reads from genome and exon-exon boundary will use megablast to align geneA and geneB to find exact break point span within geneA and geneB

[sunny.ECNU]

wonderful explaination.Thanks guy from Taipei.