How to determine statistically significant recurrence or enrichment of LOH

Question

[Obi Griffith]

Is anyone aware of an existing tool for identifying statistically significant recurring regions of loss-of-heterozygosity (LOH) in (for example) a cohort of tumor vs normal sequence or array data. I'm thinking of something like GISTIC but specifically for LOH.

[Chris Miller]

I don't think there's any reason why you couldn't apply GISTIC with some tweaking of the inputs and params. If I'm wrong and that turns out to be impractical, the underlying algorithm isn't that crazy, IIRC, and wouldn't be awful to implement something similar. It's essentially peak finding, followed by random permutation of the same size events to establish significance.

[Obi Griffith]

Yeah. I figured as much but was curious if anyone out there had seen something that was LOH-specific.