Compatibility of DeepSomatic Tumor-Only Mode for Short-Read RNA-Seq without Control Sample

google / deepsomatic

DeepSomatic is an analysis pipeline that uses a deep neural network to call somatic variants from tumor-normal and tumor-only sequencing data.

BSD 3-Clause "New" or "Revised" License

127 stars 16 forks source link

Compatibility of DeepSomatic Tumor-Only Mode for Short-Read RNA-Seq without Control Sample #22

Open Yousuk-Song opened 1 week ago

Yousuk-Song commented 1 week ago

Hello,

Thank you for developing those fancy variant callers.

We conducted short-read RNA sequencing, and since we do not have a control sample, we are planning to call variants using the DeepSomatic tumor-only mode. However, the tumor-only mode is only supported for Illumina WGS, PacBio, and ONT, and not for WES. Would it be acceptable to apply the WGS mode to our sample? Additionally, are there any significant differences in the algorithms or methods between WGS and WES for variant calling, and would this lead to meaningful differences in the results?

Many thanks, Song

kishwarshafin commented 6 days ago

Hi, we are internally discussing on how best to support this. Please give us some time to figure out the best way to support your use-case.

Yousuk-Song commented 5 days ago

Additional question: Does setting region (chr2:????-???) makes change in mutation call result?

kishwarshafin commented 4 days ago

Hi @Yousuk-Song,

Currently we are testing a WES tumor-only model to see if we can make it available to you. Please allow us a few days to run some sanity checks before making it public.

If you are running on a smaller chunk like chr1:11550000-11590000 the results may not be the same as running chr1 entirely as local realignment can change if you specify too small of a window. Hope that answers your question.