Please install:
python >= 3.6 version
Tensorflow >= 1.12.0
scikit-learn >= 0.20.2
python cmfw.py data/F1_F2_SL_binary data/F1_F2_coexpr_for_train data/F1_F2_me_for_train data/F1_F2_pathway_for_train data/F1_F2_ppi_for_train data/F1_F2_proteincomplex_for_train data/F1_F2_index_for_train data/F3_index_for_test 332
1st column: training set with {1,0} binary value, 1=SL, 0=non-SL (gene(row) * gene(col) matrix)
2nd column: co-expression training set with co-expression score (gene(row) * gene(col) matrix)
3rd column: mutual exclusivity(ME) training set with ME score (gene(row) * gene(col) matrix)
4th column: pathway training set with {1,0} binary value, 1=co-membership, 0=non co-membership (gene(row) * gene(col) matrix)
5th column: ppi training set with {ppi interaction score}
6th column: protein-complex training set with {1,0} binary value, 1=complex co-membership, 0=non co-membership (gene(row) * gene(col) matrix)
7th column: gene pairs selected for training (row, column, SL/non-SL {1,0}). The row and column index is started with 0
8th column: gene pairs selected for testing (row, column, SL/non-SL {1,0}). The row and column index is started with 0
9th column: the row counts of the input matrix
If you find this useful for your research, we would be appreciated if you cite the following papers:
@article{liany2020predicting,
title={Predicting synthetic lethal interactions using heterogeneous data sources},
author={Liany, Herty and Jeyasekharan, Anand and Rajan, Vaibhav},
journal={Bioinformatics},
volume={36},
number={7},
pages={2209--2216},
year={2020},
publisher={Oxford University Press}
}