Predicting lncRNA–disease associations based on singular value decomposition and deep learning techniques
tensorflow==1.3.0
numpy==1.11.2
scikit-learn==0.18
scipy==0.18.1
In this GitHub project, we give a demo to show how SDLDA works. In data_processing folder, we give following datasets we used in our study.
-
lda_interMatrix.mat is the raw lncRNA-disease interaction matrix with matlab format. Its shape is 577 lncRNAs x 272 diseases.
-
matrix.npy is the lncRNA-disease interaction matrix with numpy format.
-
data.pkl is used to store the sampled positive and negative samples.
-
u_feature.npy is the U matrix of SVD technique used in our study, its shape is 577x64.
-
v_feature.npy is the V matrix of SVD technique used in our study, its shape is 272x64.
-
lnc_index.txt is the names of lncRNAs with corresponding serial number.
In our demo, we provide a leave-one-out cross validation to evaluate our model. You can use cross_validation.py to see experimental results and predict lncRNA related diseases. If you want to tune some hyper-parameters, you can change some values of hyper-parameters in hyperparams.py.
The other details can see the paper and the codes.
This project is licensed under the MIT License - see the LICENSE.txt file for details