search

dkoboldt / varscan

Variant calling and somatic mutation/CNV detection for next-generation sequencing data
152 stars 34 forks source link

MNV or two SNV #42

Open aledj2 opened 5 years ago

aledj2 commented 5 years ago

How does Varscan (v2.4.3) determine that two consecutive SNVs are part of the same event and call the variants as a MNV?

I have multiple examples where BAM files (which have gone through differing preprocessing/alignment steps) have resulted in different descriptions of the same variants.

There are two adjacent SNVs on the same strand. In one BAM file varscan calls this as a MNV, but in another it calls it as two SNVs. Variants are called as SNVs even with near identical MAFs (40.11% and 40.17%) and similar read counts.

I am using varscan v2.4.3 and the mpileup2cns command: samtools mpileup -f genome/grch37.fa -B -d 500000 -q 1 input.bam | java -Xmx12030m -jar /usr/bin/VarScan.v2.4.3.jar mpileup2cns --min-coverage 10 --min-reads2 5 --min_avg_qual 15 --min-var-freq 0.01 --min-freq-for-hom 0.75 --p-value 0.05 --strand-filter 0 --output-vcf 1 --variants

zgtman commented 2 years ago

Hi did you please solve this issue? I am facing the same problem. Or is there any tool to convert two phased SNV to MNV?

Best,

Paul