CNVkit GRCh38 tumor-normal pair

Question

[plink_9857]

I recently tried to use CNVKit in DNAnexus. CNVKit requires a tumor and normal bam, reference fasta (I used GRCh38 from Gencode) and a target or bait file.

I tried to use the hg38 bed file from ucsc (hg38_GENCODE_v23_basic.bed), but the CNVKit program did not find common chromosome names between the TCGA WES bams and the hg38 bed file...

Does anyone know where I can find the right bed file for GRCh38?

[WouterDeCoster]

You need the target file specific for your kit which was used to enrich samples for the exome sequencing, which you can probably find on the website of the company.

Answer 1

[Eric T.]

If your reference genome is missing the chr prefixes but the other files have them, then you can either:

• Add the chr prefixes to the GRCh38FASTA file yourself (e.g. sed 's/^>/>chr/g, which you can run directly on DNAnexus with the Swiss Army App) -- slightly risky, e.g. the mitochondrial chromosome might be different, if you care about it, and any haplotype/alternative contigs captured by TCGA could be mismatched.
• Download UCSC's version hg38 from the Genome Browser website -- or look for it in the DNAnexus public resources.