PHArmacoKINetic PROperty calculator: Used to predict varies pharmacokinetic properties using QSAR models. If you use please cite our paper. There is a webserver that runs these models, but for large numbers of compounds, running locally using this code is much more effective

• Conda

git clone https://github.com/jonswain/PhaKinPro.git
cd PhaKinPro
conda env create -f environment.yml
conda activate PhaKinPro

The packages necessary to run the project are now installed inside the conda environment.

After downloading, PhaKinPro/phakinpro.py can be called from the command line with python phakinpro --help

--infile is required and is the fileloc for a csv of SMILES to predict properties for. Requires that csv has header and is comma seperated --smiles_col is the name of the column containing the SMILES strings of interest. Defaults to "SMILES" --id_col is the name of the column containing the compound identifiers. Defaults to "identifer" --outfile is the fileloc of where the output csv file should go. Defaults to current-working-directory/phakin_output.csv --ad flag to turn on applicability domain calculation for the models

python PhaKinPro/phakinpro.py --infile input.csv --smiles_col smiles --id_col indentifiers --outfile out.csv

This repository also contains the code to run a local webserver (or host your own). You can start the server by running qunicorn wsqi:app (or using the devolpment flask server by setting the FLASK_APP variable: $env:FLASK_APP = "main" on windows or export FLASK_ENV=main on unix). From that access 127.0.0.1:5000 to view the local server

Thanks to JSME for a free and easy to use molecule editor for webpages Bienfait, B., Ertl, P. JSME: a free molecule editor in JavaScript. J Cheminform 5, 24 (2013).