Deep-Ion

Please cite our paper. In this site you can download our benchmark data in fasta format, feature sets and source code.

Benchmark Data

Download-Data in Fasta Format

1. Ion Channels: Download

2. Ion Transporters: Download

3. Membrane Proteins: Download

Feature Sets

Download-Feature Sets

1. Ion Channels:

2. Ion Transporters:

3. Membrane Proteins:

Note: these feature sets already included feature normalization and balanced dataset techniques.

Source Code

Download Our Code

1. PSSM Feature Extraction Method and Fature Normalization Techniques:

2. Imbalanced dataset technique (SMOTE & ADASYN):

3. Deep Learning code:

Note: When build our deep learning model, we utilize the Keras 2.1.5 (https://github.com/fchollet/keras) as an open source neural network library written in Python with the Tensorflow 1.8.0 backend and NVIDIA GeForce GTX 980 Ti GPU was used for training the model.