search

marcpaga / basecalling_architectures

The Unlicense
10 stars 5 forks source link

what is the step of using models to caseball? #2

Closed zwshan closed 1 year ago

zwshan commented 1 year ago

Hello,

I know if I want to train a model, I should first use tombo to resquiggle fast5 files. And then convert them to numpy array.

But if I want to basecall a fast5 file, should I use tombo to resquiggle fast5 files? Because I think I should not use the ground truth reference gnome to act as the tombo reference. But the combo need a reference patameter.

zwshan commented 1 year ago

oh I know!

just follow your readme, use python ./scripts/basecall_original.py is ok