GNN approach to face with the problem of siRNA-mRNA efficacy prediction.
This repository provides the source code for "insert the paper's title".
-
The "data" Folder contains both "raw" and "processed" data. We reported three siRNA-mRNA interaction network datasets from 702 to 3518 interactions.
-
The "preprocessing" folder contains five scripts we released for transforming raw data into the input of our model. It also includes a "params.py" file with paths of input files and other parameters.
-
The "model" folder contains the main script "GNN4siRNA.py" for predicting siRNA-mRNA interactions. Once again, the "params.py" file contains all the parameters reported in the paper. Results of this script are given in terms of Pearson correlation coefficient and mean squared error.