DeepCRBP: Improved predicting function of circRNA-RBP binding sites with deep feature learning
- numpy (==1.19.2)
- torch (==1.9.1)
- pandas (==1.3.2)
- scikit-learn (==1.0.0)
python main.py
The circular RNA binding proteins include:
AGO1, AGO2, AGO3, ALKBH5, AUF1, C17ORF85, C22ORF28, CAPRIN1, DGCR8, EIF4A3, EWSR1, FMRP, FOX2, FUS, FXR1, FXR2, HNRNPC, HUR, IGF2BP1, IGF2BP2, IGF2BP3, LIN28A, LIN28B, METTL3, MOV10, PTB, PUM2, QKI, SFRS1, TAF15, TDP43, TIA1, TIAL1, TNRC6, U2AF65, WTAP, ZC3H7B.
You can choose from the list of proteins in the main.py
batchsize 50
learningrate 5 ร 10โ4
model.py defines our model structure
gp2lib.py extracts the secondary structure of circRNA
pytorchtools.py implements an early stop mechanism
Deal_Kmer.py and file_encoding.py preprocess the data