adapt for ultra-deep sequencing data

[WeiweiBian]

Hi,

I try to adapt your model for my deep-targeted sequencing data(> 100,000X depth) for somatic mutation calling. I noticed that the default read depth is identified as 100 in the proprocess.py.
https://github.com/bioinform/neusomatic/blob/master/neusomatic/python/preprocess.py#L243

I wonder whether this depth could be modified to a larger number for our deep sequencing data and whether our data can be encoded into layers for the architecture of the CNN model you built.

Thanks a lot!

Best,
Weiwei