Mutect2 Somatic Variant Calling for WES Data from Different Library Versions
Hi all,
I am working with matched whole-exome sequencing (WES) data generated using different library versions (V5 and V8) and want to call somatic variants using Mutect2 in GATK 4.2.0.0.
To account for the differences in target regions, I created an intersecting BED file between the two versions using bedtools and filtered the BAM files accordingly with samtools:
samtools view -b -h -L intersecting.bed originalbam.bam > newbam.bamThen, I ran Mutect2 using the following command:
subprocess.check_call("{java} -Xmx{memory} -jar {gatk} Mutect2 -R {bwa_ref} \
-I {folder}/Alignments/{tumour}.sorted.dedup.recal.bam \
-I {folder}/Alignments/{normal}.sorted.dedup.recal.bam \
--germline-resource {gnomAD} --panel-of-normals {pon} -normal {normal} \
--f1r2-tar-gz {folder}/Mutect2/{tumour}.f1r2.tar.gz \
-L {bed} -ip 150 \
-G StandardMutectAnnotation --tmp-dir ./ \
-O {folder}/Mutect2/{tumour}.mutect2.vcf.gz". format(**dict), shell = True)
My Questions:
- Is this an appropriate approach for calling somatic variants when WES data comes from different library versions?
- Are there any additional considerations (e.g., bias correction, filtering strategies) that I should implement to improve variant calling accuracy in this scenario?
I’d appreciate any insights or suggestions. Thank you in advance!
Best,
Dakyung
-
In order to call variants in Mutect2 you certainly don't need to find any intersection between your tumor and matched normal since Mutect2 can use any information that can be derived from matched normal sample. You certainly don't need to perform any additional bias correction tasks. You may need to enable genotyping germline and pon sites just to make sure that you include all the germline data in the file to perform comparisons later on.
I hope this helps.
Regards.
-
Thank you so much for your reply!
I’m a bit confused, though—if the matched normal and tumor samples were generated using different library versions (V5 and V8, respectively), which BED file version should I use for the
-Loption in Mutect2? If you suggest that I don't need to use an intersecting BED file, does it mean that the BED file version doesn’t matter?Kind regards,
Dakyung Lee -
Hi again.
We would recommend using the bed file for the capture kit that cancer samples were enriched. It would be sufficient to call the entire coverage of the cancer sample.
I hope this helps.
Regards.
Please sign in to leave a comment.
3 comments