Varscan somatic error

Question

[tejaswikoganti]

Hello,

I am trying to run Varscan somatic command for the first time. I have single tumor-normal mpileup file with reads only for certain regions on chr 17 and with a tumor purity of 50%. My command looks like this -

java -jar Varscan.jar somatic file.mpileup OUTPUT --mpileup 1 --tumor-purity 0.5

I encountered an error saying

Error: Invalid format or not enough samples in mpileup: 
ï¿ï¿½BC�

I am not sure if this is because my file is really small. Has anybody else faced this problem?

Thank you!

Tejaswi

Answer 1

[Charles Warden]

I typically use VarScan somatic calls with two separate pileup files (one for tumor and one for normal). It seems like a combined mpileup file could work, but would recommend trying that if you are still having difficulties.

[Chirag Nepal]

Hey Charles,

Can varscan somatic call be stored in vcf format. Because i could get output only in .snp and .indel.

Since some other downstream tools use vcf format. What tools did you use after u called somatic mutations.

thanks !

[Charles Warden]

I just manually convert the file format for annotation in ANNOVAR.

I am only aware of the mpileup commands (mpileup2snp, mpileup2indel, mpileup2cns) having a parameter to produce .vcf files. However, I didn't develop the program, so it is possible that I missed something.

[Chirag Nepal]

Thanks Charles !

I wanted to extract the read count that support the reference/variant call for each Normal/Tumor pair, to see if the variant is occurring in all or few libraries. There is readCount script from Varscan

java -jar VarScan.jar readcounts [pileup file] OPTIONS

But the problem is that I have assembled all Normal/Tumor for 5 patients in the same file. Do you know some easy way on how to get read count supporting reference/variant frequency for each patient. If not, I might just have to redo mpileup for each normal/tumor pair.

Thanks!

[Charles Warden]

I would recommend asking your question on a VarScan discussion group.

I think the separate .pileup files will work, but I'm guessing that this function won't really save you much time (since the pileup files already contain a coverage column). However, I honestly haven't tried this particular function.