Code to produce the main figures of the paper.
- Figures in .AI, .PDF and .PNG format are accessible at
results
. - Executable IPython Notebooks to produce most panels related to this figures can be found in
notebooks
. - Panels are stored in the
assets
folder.
All dependencies related to these notebooks are pip
-installable and you should be able to install them on-the-go as you encounter them. Some modules you may want to install are fragmentembedding, mini-automl, mini-xai and stylia:
git clone https://github.com/ligand-discovery/fragment-embedding
cd fragment-embedding
python -m pip install -e .
git clone https://github.com/ligand-discovery/mini-automl
cd mini-xai
python -m pip install -e .
git clone https://github.com/ligand-discovery/mini-xai
cd mini-xai
python -m pip install -e .
python -m pip install git+https://github.com/ersilia-os/stylia.git
- This repository is likely to evolve as figures are revised over manuscript iterations. Therefore, it is possible that, before publication, panel names in
assets
do not correspond to panels in the final figures. - Likewise, some panels produced in the executable notebooks may not be part of any final figure.
- Also, please note that some figure panels were not produced with executable notebooks (e.g. in-fluorescence gels, or networks laid out with Gephi).
- There are some custom paths associated to Supplemental Figures, since the data related to them is large and computations were conducted elsewhere.
- Similarly, for Figure 6 data is read from the results of the
interactome-signatures
repository.