piRNN Program(Python) 
Kai Wang 
01/13/2017

GENERAL INFORMATION
-----------------------------------------------------------------

- piRNN is a program that used for piRNA detection. The deep learning 
  library that we used is Keras that running on top of TensorFlow.

- The program can be downloaded at htps://github.com/wangk4/piRNN
  After downloading and unziping, the following items should be 
  found in the folder:
  	- piRNN.py
  	- functions.py
  	- Ele_piRNN.h5
  	- Dro_piRNN.h5
  	- Rat_piRNN.h5
  	- Hum_piRNN.h5
  	- README
  	- test_input_data.fa

- Four species' (C. elegans, D. melanogaster, Rat, and Human) piRNAs
  data were used for training four differnet models, respectively. 

USAGE NOTES
-----------------------------------------------------------------

- The program can be easily executed in a terminal like:

python piRNN.py [-h] -s <species code> -i <input file in fasta format> -o <output file>

- The species code should be a number that from 1, 2, 3, or 4.
  - 1 means C. elegans model.
  - 2 means D. melanogaster model.
  - 3 means Rat model.
  - 4 means Human model.

- The input file should be small RNA data that in fasta format.
- The output file is also in fasta format.
- Use "-h" to show help information.

*please unzip the model files first.

RUN EXAMPLE DATA
-----------------------------------------------------------------

- The test_input_data.fa is used for testing the program. Users can 
  run the following command to test the program.

python piRNN.py -s 4 -i test_input_data.fa -o out.fa

- The result should be in out.fa

CONTACT INFORMATION
-----------------------------------------------------------------
Kai Wang,
Department of Biology, Miami University, Oxford, Ohio, USA   
Email: [email protected]

Prof. Chun Liang, 
Department of Biology, Miami University, Oxford, Ohio, USA
Email: [email protected]