Likely somatic variant not called by DeepSomatic

[QichenFU]

Hi, I found some likely somatic variants that were not called by DeepSomatic (heterozygous in normal and homozygous in tumor). May I ask what might be the reason?

Top: tumor WGS data. middle: normal WGS data. bottom: somatic varian calls

[danielecook]

@QichenFU If you observe a variant that is heterozygous in the normal and homozygous in the tumor, then what you are observing is a Loss-of-Heterozygosity (LOH) event. LOH is generally caused by large structural deletions or even full chromosome loss.

DeepSomatic is designed to call somatic single-nucleotide variants (SNVs) or small Indels.

You can use alternate tools (such as ASCAT) to map out where LOH regions are in tumor genomes. DeepSomatic can still call SNVs and indels in these regions.

[danielecook]

@QichenFU I will close this issue for now, but feel free to reopen if you have any further questions. Thank you.