DeepOlf is a deep neural network-based prediction model, developed on a large dataset of odorants, non-odorants and olfactory receptors (ORs), using a potential set of physiochemical properties and molecular fingerprints. It allows accurate inference of an odorant over a set of olfactory receptors.
Install Anaconda3-5.2 or above.
Install or upgrade following libraries (python, numpy, tensorflow, keras, scikit-learn).
Download and extract zipped file.
Prepare input file (user_input.csv). Calculate molecular properties and fingerprints from SMILES or sdf format of a chemical compound using software’s like PaDel, alvaDesc (recommended), DRAGON, etc. and save as user_input.csv. Keep the header information (first row) as such in input file.
Change value of path variable in DeepOlf_script.py to the extracted folder and execute the script.