A cheminformatics package to perform Applicability Domain of molecular fingerprints based in similarity calculation.
Project description
pyADA
pyADA (Python Applicability Domain Analyzer) is a cheminformatics package to perform Applicability Domain of molecular fingerprints based in similarity calculation. In this case, the calculation of the Applicability Domain consists of a scan of similarities of the structures present in the test set in relation to the training set, the best similarity threshold is the one with the lowest error and also the lowest number of molecules with similarity below the threshold. A notebook file with an example of using this package is present in the directory 'example/example_of_use.ipynb'
Dependencies
- numpy
- pandas
- tqdm
- scikit-learn
- Tested in python3.6 and python3.8
Install
Via pip
pip3 install pyADAqsar
Via github
git clone https://github.com/jeffrichardchemistry/pyADA
cd pyADA
python3 setup.py install
How to use
This package has three classes: Smetrics (perform some statistical parameters like Q2ext R2ext etc), Similarity (realize similarity calculations based in differents metrics ) and ApplicabilityDomain (run a scan of AD with differents thresholds). The line code bellow import all classes.
from pyADA import Smetrics, Similarity, ApplicabilityDomain
A file containing a jupyter-notebook with a few examples of use is in 'example' folder. For more information about documentation run the help function of classes.
help(Smetrics)
help(Similarity)
help(ApplicabilityDomain)
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for pyADAqsar-1.1.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | c4cfcdfb5119a6ded11c22428015f1c1be82e0e51801b561b9c570436dead0da |
|
MD5 | 68fc6509fa15c1a5bcb9873b674667c6 |
|
BLAKE2b-256 | fc30da06c1f668f878c1734d4056749c65f330fab50a4dad7993787de9652445 |