How to compare a tumour sample to multiple normal samples in Freebayes

Question

[nanana]

I'm currently running Freebayes on tumour normal pairs to detect somatic SNPs in the tumour sample:

Raw germline calls:

freebayes -f genome.fa --pooled-discrete --genotype-qualities --read-mismatch-limit 70 \
tumour.bam normal.bam -v tn_raw.vcf

Decompose complex events into single SNPs and, tag the tumour and normal samples and quality-filter:

vcfallelicprimitives tn_raw.vcf | \
vt decompose_blocksub - | \
vt normalize -q -r genome.fa - | \
vcfsamplediff -s VT normal tumour - | \
vcffilter -f "DP > 20" \
-f "QUAL > 1 & QUAL / AO > 10" \
-f "SAF > 0 & SAR > 0" \
-f "RPR > 0 & RPL > 0" \
-f "TYPE = snp" > tn_snps_filt.vcf

Filter for somatic SNPs

vcffilter -f "VT = somatic" tn_snps_filt.vcf > tn_snps_somatic_filt.vcf

I would like to be able to include a panel of normals, and exclude any SNPs called as somatic that are present in any of the normal samples. I know that you can supply Freebayes with multiple samples, but how can I feed this information into vcfsamplediff?

Answer 1

[AcademicDialysis]

I believe you mean to filter in somatic SNPs and filter out germline SNPs. Using multiple normal samples is often known as a pool or panel of normals. More about that here. I looked around a bit at Freebayes and couldn't find anything directly from them. Looks like bcbio supports pooled normals via Freebayes though. You should check out MuTect - they are pretty commonly used and have pretty good documentation of pooled normals creation.

[nanana]

I appreciate the comment, but this has got nothing to do with the question - of how to include a PON in Freebayes