Skip to content

Latest commit

 

History

History
20 lines (11 loc) · 766 Bytes

README.md

File metadata and controls

20 lines (11 loc) · 766 Bytes
Raw

Predicting multiple types of Modification sites in RNA using GAN (MR-GAN)

MR-GAN model training takes the transcriptomic sequences as input in fasta format (*.fa).

Requirement:

  1. Python3
  2. Tensorflow

To run code:

  1. At first, run onehot_encode_geneSeq.py for one-hot encoding of transcriptomic sequences and save the embedded matrix as HDF5 file (*.h5)

python3 onehot_encode_geneSeq.py --genome_file_path your/saved/fasta/file/directory

*** Note that, this code generates ~30 GB data for whole gene sequence and save to your local disk.

  1. Feed the saved HDF5 files into the mrgan_main_wgan.py to train the MR-GAN model.

python3 mrgan_main_wgan.py --genome_file_path your/saved/HDF5/file/directory --log_dir directory/where/trained/model/will/be/saved