Virtual screening for SARS-CoV-2 inhibitors is an useful and easy to implement approach based on RNN and deep models in general.
This is an interesting study conducted by 'The European Laboratory for Learning and Intelligent Systems' (ELLIS), utilizing ChemAi a deep architecture with a SMILES string representation of molecules.
Preprint Large-Scale Ligand-Based Virtual Screening for SARS-CoV-2 In...
Sometimes drugs having a strong and negative impact on the function of the heart must be applied. Sometimes drugs like hydroxychloroquine are used in some COVID-19 patients, which can induce arrhythmias.
For such situations, the following approach combining complexity and AI/ML methods was developed.
It predicts ventricular tachycardia, ventricular fibrillation, and most importantly drug-induced Torsades de Pointes arrhythmias.
A preprint is available now (full, extended version later)