mutect2 analysis with one tumor vs two normals from the same individual?

May 02, 2024 02:27

gatk Mutect2 \
     -R reference.fa \
     -I tumor1.bam \
     -I normal1.bam \
     -I normal2.bam \
     -normal normal1_sample_name \
     -normal normal2_sample_name \
     --germline-resource af-only-gnomad.vcf.gz \
     --panel-of-normals pon.vcf.gz \
     -O somatic.vcf.gz

I am using gatk/4.2.3.0

Is it correct if I run this? for some reason, when i compared the tumour with two normals some variants disappeared, but when I did it with either of the two normals, about 4 new variants appeared.

Thanks for your help.

3 comments

Gökalp Çelik

May 02, 2024 10:47
Hi María Fernanda Talavera Cruz

Using multiple normals is no problem as soon as they are indicated as normal to Mutect2. When you perform FilterMutectCalls what is the state of those 4 variants? Do they get filtered somehow? Can you share IGV views of those 4 variant sites from tumor and normal samples if possible?
0

Comment actions Permalink
María Fernanda Talavera Cruz

May 10, 2024 19:14
Hi Gökalp Çelik,

Apparently, they get filtered even before using FilterMutectCalls. The four variants are deletions and I can't see them on IGV (two are just in one read) so I am really confused of why they are appearing. The IGV views of the tumours are attached. Thank you for your help!

The variants that appear when just comparing it with either one of the normals are:

4 137936757 GTTT G

4 137936762 TGTCAGCGCGGACTCCAAGAATTCCAACAGCAATGTGATCCAGGTGGGCTGAG T

1 88407393 GTCCCTGAGTCCTTACCTGTTTCGGGCGACCCGGAGCTCCTTGAAGAGAGTAAGTGAATGGAGATACC G

10 22148560 CAA C

4 88419812 TAGATGGTGAATGCTGTTTAAAGAAAAAGCATCATGA T
0

Comment actions Permalink
Gökalp Çelik

May 11, 2024 07:38

Edited
Hi María Fernanda Talavera Cruz

The fact that they get filtered out even before FilterMutectCalls tells us that they may also get filtered out in the filtering step even when they appear in the original file. Our general practice is to keep suspicious of those sites that are not supported by more than one spanning read and variants only happen to occur within reads with multiple alignments such as secondary or supplementary.

Mutect2 and HaplotypeCaller uses a local reassembly approach to detect variants at sites with high entrophy. During reassembly step certain impurities such as improperly removed adapters/PCR primers, chimeric PCR products, cross-species contaminants etc.. may interfere with the local reassembly especially for Mutect2 where allele fractions count. We strongly advise using FilterMutectCalls step to go over all variants to see what filters they have and then reassess the validity of calls by visually or other orthogonal approaches where necessary.

Here are 2 writeups from our team that may be useful for understanding the variant calling steps.

https://gatk.broadinstitute.org/hc/en-us/articles/360043491652-When-HaplotypeCaller-and-Mutect2-do-not-call-an-expected-variant

https://gatk.broadinstitute.org/hc/en-us/articles/360035891111-Expected-variant-at-a-specific-site-was-not-called

I hope these help.

Regards.
0

Comment actions Permalink