Interpreting Tigra-Sv Output

Question

[Mark Dunning]

I'm interesting in using the tigra-sv package to filter down my candidate list of breakdancer calls, but somewhat stuck with what to do with the output files.

For every putative breakdancer call it gives me a contigs.fa file and a het.fa file but I don't really know what to do with these! Do I need to re-align them? Some of the contigs.fa files can be of size zero.

I was hoping for a nice output table that would list all the candidate SVs that I provided to tigra-sv with some assessment of whether tigra-sv thinks it is legitimate or not. It seems like there should be an extra parsing step.

Does anyone have advice on how to deal with the output?

Regards,

Mark

Answer 1

[Chris Miller]

Hi Mark, I ran this by a few of the guys downstairs that do some of the maintenance on tigra-sv (I'm not very familiar with the program myself).

There are 3 output files from running tigra-sv:

1. *.fa.contigs.fa => basic contigs reported by TIGRA
2. *.fa.contigs.het.fa => alternative contigs constructed from the contig graphs
3. *.ref.fa => local reference file covering the sv region

The sv alleles could be in either file 1 or 2. The cross_match of 1 vs 3 and 2 vs 3 are run. The cross_match results are further parsed based on some logic to decide which of 1 or 2 (or neither, aka, failed the filter) is the right one and its sv type (ins, del, itx.ctx).type (ins, del, itx.ctx).

If you have further questions, you might try contacting one of the authors - Ken Chen ([e-mail]) or Lei Chen ([e-mail])

[Mark Dunning]

Thanks for the reply. It sounds like the cross-matching is done automatically? I've run tigra-sv and get the 'AllDone' message printed, but no useful output other than the 3 file types you list.

I'll get in touch with the authors if you don't happen to know.

[Chris Miller]

Yeah, I've only used Tigra in the context of our analysis pipelines, which take care of all the details and intermediate files for me. Ken or Lei may be able to help you out, though.