Detection of somatic structutal variants in paired genome data

[Jane M]

Hello everybody,

Several tools have been developed for the detection of structutal variants in genome data since 2009. I would like to focuse on the case of somatic structural variants.

The most commonly used tools are probably BreakDancer, Pindel, Delly, Novobreak, SquareDancer and SVDetect. And maybe some others.

I recently got 20 paired normal tumor genome data. I am looking for the "best tools" to use in my case.
My case: 101bp paired-ends reads. Coverage ~30x for each sample. I am working on leukemia. Almost 100% of tumor in the tumor sample but I can have tumor contamination in the normal sample (from 0 to 20%).

I intend to use BreakDancer and Pindel. They seem complementary for the size of indels they are able to detect.
Regarding my data, do you any advice/recommendations?
Do you think I should prefer other tools?

Thank you in advance,
Jane

[m_two]

Based on this result https://www.synapse.org/#!Synapse:syn312572/wiki/63089

I would consider adding Delly since Novobreak has not been released yet. Sensitivity is still an issue. Given the coverage you will probably have address the normal being contaminated with the tumor cells and recover additional filtered SVs.

[Jane M]

Thank you for your answer m_two.
Could you (and others) please tell me what you do concretely when you look for SVs?
Do you use several tools (2, 3?)? Do you combine or intersect the results?