Orange3 Chemoinformatics is an add-on for the Orange3 data mining suite. It provides extensions for embedding molecule dataset annotated with SMILES through a pre-trained deep neural networks.
This is fork of the original Chemoinformatics widget with custom embedder. Embedder widget implements four different methods: substructure, circular, topological fingerpritns and and encoder of the autoencoder.
Orange3-Chemoinformatics requires RDKit, that is conda-installable running
conda install -c rdkit rdkit
and NumPy based deep learning package (https://github.com/PrimozGodec/ndf)
To install the add-on from source run
python setup.py install
To register this add-on with Orange but keep the code in the development directory (do not copy it to Python's site-packages directory) run
python setup.py develop
You can also run
pip install -e .
which is sometimes preferable as you can pip uninstall packages later.
After the installation the widgets from this add-on are registered with Orange. To run Orange from the terminal use
orange-canvas
or
python3 -m Orange.canvas
New widgets are in the toolbox bar under the Chemoinformatics section.
Example of an annotated database: https://s3-us-west-1.amazonaws.com/deepchem.io/datasets/molnet_publish/clintox.zip