README.html
<!DOCTYPE html>
###### Decription of the files on GitHub ######
Usage: python reptb_nbi.py inputfile.csv outputfile.csv
Prerequisites:
- This algorithm has been tested on Python 2.7
- The Python program needs following modules to be imported:
- numpy
- csv
- numpy.matlib
Input File
/RepTB/inputfile.csv : This is the input file that we have used used in our study. This file contains the drug-target interactions from drugbank and GO mapping.
NBI Algorithm
/RepTB/reptb_nbi.py : The NBI code snippet contains comments that would help the researchers with their own data. All the researchers have to do is upload their GO mapped DTI file (inputfile.csv)and run the python code. The code will first generate an adjacent matrix using the input file and then create a prediction matrix called NBIscore which will be a m x n matrix. The output file (outputfile.csv) will be created which will contain predicted DTIs having scores within 20% of the max score for each Target.
Output File
/RepTB/outputfile.csv : This is the final output file generated using our input file and contains the prioritized list of predicted DTIs with the predicted scores. If you are uploading your own input file containing the DTIs, you might need to change the name of the output file in the python code provided.
Supporting Files:
- Generating_random_netoworks.m: This is a matlab file that was used to generate random networks with the same number of nodes but varying drug interactions and running through NBI. This file was only used to establish the robustness of the NBI.
- uniprot_go_annotations_mf_strong_evidence_new.csv: This file contains the mapping of the molecular function GO terms and Uniprot Ids.
***Remarks***
Customised GO mapped input files can be fed into the NBI algorithm for different model systems.