If I understand correctly, the starter notebook "Bootstrap notebook" trains models to predict -log(Standard Value) where "Standard Value" is potency. Is there a consensus on what values of potency indicate that a molecule is a good candidate or is that part of the problem?
You are looking at a wrong outdated notebook. Apologies is there. the actual starting notebool is https://gitlab.com/instadeep-public/grandchallenge/-/blob/master/IndabaChallenge.ipynb
Thanks for the reply. The notebook you linked to looks like it produces a .pdb file when given a protein and a ligand. Is this correct?
I thought that you would use the Bootstrap notebook to find good candidates for a protein-ligand pair (via machine learning), and then use pyrosetta to evaluate the pair and produce a .pdb file. Is this correct?