Straightforwarding Scoring Suite (3S) is a collection of several tools to ease the procedure of desiging a machine learning scoring function by providing a GUI.
Milad Rayka, milad.rayka@yahoo.com
Below packages should be installed for using 3S. Dependecies:
- python (3.9), numpy, pandas, seaborn, streamlit, matplotlib, biopandas scipy, scikit-learn ,progressbar2, xgboost, jupyterlab, pdb2pqr
First install python then make a virtual environment and activate it.
On windows:
python -m venv env
.\env\Scripts\activate
Which env is the location to create the virtual environment. Now you can install packages with one of the following methods:
1- pip install package_name
2 - python setup.py install
3 - python install -r requirements.txt
4 - Double click on install.bat (Recommended)
For running 3S:
streamlit run webapp.py
or
Double click on webapp.bat (Recommended)
So far, this suite contains five tools:
1-Feature Generation:
In this mode, features for different structure of complexes based on aforementioned method are genereted[1].
2-Model Training:
In this mode, a machine learning scoring function (Gradient Boosting Trees) is designed for a dataset of provided complex structures.
3-Prediction:
Binding affinity of complexes are predicted using a ML-Score.
4-Normality Test:
In this mode, if the test data has binding label, normality property of errors is analysed.
5-Add Hydrogen:
Add hydrogens to ligand and protein at pH=7.4 using PDB2PQR and Openbabel. For installing Openbabel check on this link.
Check the provided Tutorial.pdf file for more information and example.
3S is tested on the following system:
OS | RAM | CPU | Browser |
---|---|---|---|
Windows 10 | 4.00 GB | Intel Core (TM) i5 - 7200U (2.50 GHz) | Firefox - 97.0.1 (64-bit) |
We don't assume using macOS, Linux, or other browsers make problems.
To ensure code quality and consistency the following tools are used during development:
Copyright (c) 2021-2023, Milad Rayka