peace195 / mirna-identification-conv2d

Precursor microRNA Identification Using Deep Convolutional Neural Networks

Precursor microRNA Identification

Code for project "Overview of Biomedicine" under lab course "Deep Learning for Computer Vision and Biomedicine" - Technical University of Munich, Germany. Depends on Python 3, numpy, pytorch, scikit-learn, Bio, RNA, matplotlib.

Our paper: Binh Thanh Do, Vladimir Golkov, Göktuğ Erce Gürel, and Daniel Cremers, "Precursor microRNA Identification Using Deep Convolutional Neural Networks"

• biorxiv: https://www.biorxiv.org/content/early/2018/09/16/414656

Setting up the Environmental Variable

Ubuntu/Mac OSX

To set up the environmental variable in Mac OSX, follow the steps:

• Open ~/.bash_profile with a text editor.
• Add the line export VIENNA_PATH="the-path-of-your-viennarna-package" and save.
• Restart the terminal, check if it is working by typing echo $VIENNA_PATH

An example: export VIENNA_PATH="/Users/peace195/ViennaRNA/lib/python3.6/site-packages"

Usage

python filename.py

where:

File / Folder	Description
dataset/	human, cross-species and new RNA sequences
results/	running log of each epoch in testset
weights/	saved model parameters
cv.py	5-fold cross validation for human and cross-species dataset using fixed-sized inputs ConvNet architecture
test.py	Train and test in human and cross-species dataset with fixed-sized inputs ConvNet architecture
test_new.py	Train and test in new dataset with fixed-sized inputs ConvNet architecture
cv_variable_size.py	5-fold cross validation for human and cross-species dataset using variable-sized inputs ConvNet architecture
test_variable_size.py	Train and test in human and cross-species dataset with variable-sized inputs ConvNet architecture
test_new_variable_size.py	Train and test in new dataset with variable-sized inputs ConvNet architecture
utils.py	Read sequences, encode sequences and measurements
ConvNet.py	Some ConvNet architectures such as Alexnet, Resnet, etc.
statistics.py	Statistics of dataset