Category: T2 - General Drug Discovery, Including COVID PS ID : DDT2-09

Automated Analogue library generation tool using bioisosteres

Bioisosteric libraries contain reaction fragment pairs that can be used to generate many analogues from a starting molecule. One can thus potentially generate analogues of current antivirals and those showing promise in clinical trials or literature. This helps drug discovery researchers work with known chemical space with slightly higher probability of bioactivity . Such analogues can then also be used as test data for Machine learning models or docking. The challenge is to write a user friendly software that will take bioisostere libraries in SMILES/SMARTS format as input, candidate or known antiviral molecules like Hydroxychloroquine, remdesivir, human ACE inhibitors as the second input and generate a library of several thousand molecules by automatically replacing fragments on molecules with their bioisosteric replacements. The generated molecules then need to be filtered by Lipinski filters and then clustered in families of molecules based on similarity. This final set of analogues can be output to a SMILES file. Analogues can additionally be filtered by using computational tox models although this is desirable but not mandatory for the scope of this challenge. Minimum acceptance criteria will be a tool that automates analogues generation in python + rdkit with two given inputs and stores results. Usage of open-source software like knime, rdkit, Indigo is desirable .

Login to Download Input Form