Closed chanur closed 4 years ago
I would suggest to try and run from their segmented files if possible e..g using CNVKit to transform log values into discrete. If that doesn't work we would need to run from scratch but as you said we will likely get different results. But we cant control that.
Ben said they will upload segmentation files and said that he doesn't think we should convert these to discrete values due to tumor purity issues. However, for the sake of import into cBIO, he suggested an approach where he can tell us which samples are highly amplified and we can remove these during the cBIO import step. I am not quite sure how this will work, will have to see..
Ben says segmentation file provided. Anish will review all the files Do we need to rerun from scratch Will a different algorithm produce different results.