Notebook exploring the use of several ML toolkits (DeepChem, DeepPurpose, Alphafold...) to find potential drug candidates against highly pathogenic avian influenza virus H5N1 estimated to have a mortality rate of approximately 60%.
Divided in 5 sections: Drug screening, Toxicity, ADME, Docking and (Retro)synthesis, it contains a general example of some capabilities and ML aspects (data formatting, hyperparameter optimization, etc.). To use it just open the notebook in Google Colab and replace the dummy paths with the path to the corresponding file (files available in the folder H5N1_files).
Material adapted from different DeepChem, DeepPurpose and Therapeutics Data Commons tutorials, among others. Make sure to check the corresponding repositories for more in depth information.