Varscan2 warning : resetting normal

Question

[bharata1803]

Hello,

So, I have 8 normal sample and 8 tumor sample. Each normal and tumor group had been processed to get mpileup file by samtools. After that, I did the varscan2 soomatic function to call variation. After that, I got warning resetting normal and from what I found in the document, it's better to not call mpileup for each group. Instead, calling mpileup for 2 bam file, 1 normal and 1 tumor. So, I had finished doing this. I take 1 normal sample and 1 tumor sample for mpileup input to generate 1 mpileup file. After that, I use that mpileup file for varscan input. The result is no warning and seems no problem. My question is, how do I integrate the result for 8 normal/tumor pair varscan results? I figure if I proocess the data with 1 normal and 1 tumor, I will need to do varscan call 8 times and will get 8 results. Is there any method to join the results?

Answer 1

[russhh]

I always make a combined vcf for all patients studied.

After you have your patient-specific vcfs, Merge the positions present in those files to generate a single bed.

Then, for each patient, extract the variants (and non-variants) lying in the positions in that bed using varscan somatic --validation.

The varscan-validation vcfs can then be merged.

If you'v a good number of samples, you can use the merged vcf to find biases etc

HTH

[bharata1803]

Thank you for your reply.

So, basically, you generate VCF file for each sample, not normal-tumor pair? Is the workflow like this:

1. Generate mplieup for each sample

2. Generate vcf for each mpileup

3. Merge vcf into bed

4. Use varscan validation to the bed

5. Merge the result

[russhh]

For each PATIENT I run varscan to generate a vcf; each run of varscan requires two samples: the tumour and the normal