GRU4ACE: A novel GRU-based approach for improving the prediction of ACE inhibitory peptides using multi-source deep feature representations
###GRU4ACE: uses the following dependencies:
-
Python 3.10.12
-
esm
-
biopython
-
numpy
-
scipy
-
scikit-learn
-
pandas
-
TensorFlow keras
-
pandas==2.2.2
-
lightgbm==4.5.0
###Guiding principles:
**The data file contains a dataset
**Feature extraction：
ESM2.ipynb
bert.ipynb
BPF.ipynb
ProtT5.ipynb
Fasttext.ipynb
FEGS.ipynb
**Feature selection
Features_selection is the implementation of PCA, mRMR, MRMRD, LASSO and Elastic-Net.
**Classifier:
BigGRU.ipynb CNN_BiLSTM.ipynb CNN_GRU.ipynb DNN.ipynb GAN.ipynb GRU.ipynb LSTM.ipynb
and ML.py implement LR, KNN,DT,NB,ExtraTree,RF, Xgboost, SVM and LightGBM.