luntergroup / octopus

Bayesian haplotype-based mutation calling
MIT License
305 stars 38 forks source link

Reducing Indel False Positives in Somatic Variant Detection #285

Open jasmin774 opened 2 months ago

jasmin774 commented 2 months ago

Hello,

I am using Octopus to detect somatic variants, but I am encountering a high number of false positives, particularly with indels. In many cases, ALT alleles are also found in the Normal sample, contributing to these false positives.

I have two questions:

Are there any Octopus options or parameter adjustments that could help reduce indel false positives? Is there a recommended approach for re-filtering the VCF based on allele frequency (or read depth) to further reduce false positives? Any guidance would be greatly appreciated.

Thank you!

Octopus: version=0.7.4 Genome: mouse mm10

octopus \ -R ${ref} --threads ${threads} \ -I ${dir}/${normal_bam} ${dir}/${tumorbam} \ -o ${tumor}${normal}.octopus.vcf.gz \ -N ${normal}