TDC
TDC copied to clipboard
mims-harvard
(new feature) large molecule multi-objective fluorescent protein oracle
Describe the problem It would be great to add the large molecule fluorescent protein task introduced here as a molecular generation oracle. Code is available here.
Describe the solution you'd like
In brief the oracle would return the folding stability (-dG) and solvent accessible surface area (SASA), given a primary amino acid residue sequence and a reference PDB. The oracle would first translate the input sequence to a list of substitution mutations, then use FoldX to generate a predicted structure and the predicted folding stability. SASA can be computed from the predicted structure using BioPython.
Additional context I would potentially be willing to make a PR if this is something the maintainers would approve.
Hi Samuel, this sounds like a great oracle to include in TDC! We usually implement individual oracle function at here while the tdc.Oracle interface is the wrapper around all oracles. Let us know if you need further pointers and happy to discuss more!
Hi Kevin, haven't had time to draft a PR bc of some impending deadlines. I'm hoping to defend my dissertation in July so realistically it'll probably be a while before I'll be able to give this proper attention. If you like we can close the issue for now, or leave it open until I have more time, up to you.
