We propose a series of scripts, which are still in development, to label and train a multilayer Bidirectional long short-term memory recurrent neural network to base-call ONT-nanopore reads with modified bases.
Install BWA and make sure executable is in the PATH:
git clone https://github.com/lh3/bwa.gitcd bwamakeexport PATH=$PATH:$PWD
The easiest way to deal with the dependencies is to download Anaconda
- Download and Install Anaconda https://docs.continuum.io/anaconda/install
git clone --recursive https://github.com/BD2KGenomics/nanopore-RNNcd nanopore-RNNconda env create --file requirements.ymlsource activate nanotensormake- C executables created by signalAlign need to be in the PATH
export PATH=$PATH:$PWD/signalAlign/bin- Test installation
make test
If you want to use Nanonet to train a network, you can use the script located here to embed aligned kmers into the fast5 file which can then be used with Nanonet.
If you decide to help contribute please use pylint so our code is consistent. https://www.pylint.org/