somatic SNV/InDel number is too many

[DayTimeMouse]

Hi,

I used clairs(v0.4) to call somatic SNV/InDel(tumor/normal paired), but the PASS variants number is ~6w. Is it normal?
Compare to deepsomatic, the somatic SNV/InDel number is ~1.7w.

Do you have recommend the cutoff of QUAL value for SNV and InDel, seperately?

ClairS-0.4.0/run_clairs --tumor_bam_fn hifi_tumor_pbmm2.bam --normal_bam_fn hifi_normal_pbmm2.bam --ref_fn chm13v2.0.fa --threads 30 --platform hifi_revio --output_dir clairs_out --enable_indel_calling --enable_clair3_germline_output

Should I set a higher number for --min_coverage ? I'm confused about it.

Best regards.

[aquaskyline]

It depends on the cancer type and other clinicalpathological details of your sample to make a rough guess of the range of the somatic variants. Use HCC1395/BL from SEQC2 for example, it has over 40k somatic classified as truth. Nevertheless, you can always do downstream filtering based on the qual and tags given in ClairS' result to reach the number of somatic variants you except.