Hi, everybody,
I am now trying to use MuSiC to calculate BMR of BRCA from TCGA. One input requirement of MuSic is the coverage of each sample from wig file, which can be downloaded using firehose_get. But when I download OV or GBM data, there are coverage wig file for each sample, but for BRCA, there is only one wig file. So I am wondering if I want the coverage information of other samples, what should I do? Thanks.
When Broad's Firehose lists only 1 WIG file for a tumor type, then it is a place-holder WIG file with faked coverages from a representative sample. This WIG is used to model the coverage for all samples in that tumor type. They do this when they don't have access to the original BAM files (if the sequencing was done elsewhere, and they haven't downloaded it for Firehose). You will find that this is the case for AML, BRCA, LUAD, and UCEC.
The Genome Institute at WashU (my place of work) did the sequencing on several TCGA tumor types, and we have uploaded WIG files for AML, UCEC, and BRCA on the TCGA Jamboree server. We are waiting on Broad to import these into Firehose, but if you have credentials to download from the TCGA Jamboree, then you can find them there (/tcgajamboree/awg_mafs/*/tcga_*_wu_wigs.tar.gz)
If not, then I would recommend faking coverages. In newer projects like BRCA, lack of coverage for variant calling, was not as much of a problem as in older projects like GBM and OV.
