DeepT4 is a deep learning tool for Gram-negative bacteria Type IV secretion effectors prediction. The backbone of the framework is a convolutional neural network (CNN), which automatically extracts T4SEs-related features from 50 N-terminal and 100 C-terminal residues of the protein.
To run: $ python model.py The model file implements the convolutional neural network to train the model, evaluate the model, plot loss-acc and ROC cruves, and save the model. If you want to use different training and test data, please change the file name inside the file.
To run: $ python test_saved_model.py The test_saved_model file predicts the T4SEs using saved CNN model. If you want to predict different data, please change the file name inside the file
The latest (devel) version of DeepT4 requires£º
python 3.5 Keras 2.2.4 numpy 1.16.2 matplotlib 3.0.2 scikit-learn 0.20.3
Reference: Li Xue, Bin Tang, Wei Chen,* and Jiesi Luo. A Deep Learning Framework for Sequence-Based bacteria type IV secreted effectors Prediction