CATCaller: An end2end ONT Basecaller using Convolution-augmented Transformer.
bash run_caller.sh <model file> <fast5 folder> <signal window length> <basecalled_dir>
command parameters
model file: we provide the model/model.2048.chkpt trained on Klebsiella pneumoniae genome.
fast5 folder: directory of original sequencing fast5 files.
signal window length: the length of the signal segment, default: 2048.
basecalled_dir: the output directory of basecalled files.
- Python version >= 3.6
- CUDA/10.0
- pytorch_gpu version >= 1.2.0
- ctcdecode and fast-ctc-decode
- dynamicConv/lightweightConv. To install dynamicConv/lightweightConv module:
cd dynamicconv_layer
python cuda_function_gen.py
python setup.py build
python setup.py install --user
If you want to process new datasets and parallelly train a new model, the following tools are also needed: