lima1
commented
5 years ago Hi,
PureCN can calculate cancer cell fractions (CCFs) of somatic mutations. If you are not interested in sub-clonality of somatic mutations, you can only provide germline calls. For callers other than Mutect1, we don't do a lot of artifact filtering (except for low coverage, poor read count support and poor base quality), so you want to do that yourself. Let me know if you run into problems, possible that Varscan2 VCFs don't work out-of-the-box.
If you have a pool of normals, merge the normals in a single multi-sample VCF and follow the Quick vignette. This will remove variants not in dbSNP, but present in the PoN. It will also automatically clean up variants in poor quality regions.
If you want to use CNVkit, follow the example in the Quick vignette, no need to recalculate coverages again with PureCN.
Best, Markus
qindan2008
vedellpt
Hello,
The input vcf for PureCN should contain both germline and somatic variants. Does this mean both the germline and somatic vcf from variant caller should be combined together into a single one ? Or, just flag the variants recored in germline databases in the somatic vcf ? Or, something else?
I just want to evaluate tumor purity and ploidy. I got vcf files with varscan2. Then, the high confidence germline and somatic vcf was combined together into a single vcf file, and annotated with dbSNP, marked the INFO with "DB" flag to indicate the variant was recored in the dbSNP database.
Run PureCN with the following command:
Rscript PureCN.R --out Tumor1/Tumor1 \ --normal PureCN_Coverage/Normal1.sort.dedup_coverage_loess.txt \ --tumor PureCN_Coverage/Tumor1.sort.dedup_coverage_loess.txt \ --vcf Tumor1_Normal1.varscan2.combinded.germline.somatic.vcf \ --parallel \ --cores 6 \ --sampleid Patient1 \ --intervals baits_hg19_intervals.txt \ --segfile Tumor1.segments \ --genome hg19
Both the coverage files and intervalFile was generated from PureCN. The segment file was generated from CNVkit. The above command completed without errors.
Is this OK ?