search

sokrypton / ColabFold

Making Protein folding accessible to all!
MIT License
1.96k stars 493 forks source link

Partial binder hallucination? #283

Closed blakely3dnm closed 2 years ago

blakely3dnm commented 2 years ago

Hi,

Would it be possible to combine the partial hallucination and binder hallucination protocols to hallucinate loops or other new structural elements within a given binder scaffold?

A trial case I have in mind would involve a binding interface formed by a multi-chain protein binder and multiple receptor chains. I'd like to add a new loop between the multi-chain binder but leave the target receptor sequence and connectivity consistent with the input. I almost got this to work using the partial hallucination examples, however I'm not sure how to avoid designing unnecessary loops at the chain breaks. loops of zero length force single chain outputs.

Should I look into specifying a sequence with "/" instead of a PDB?

Thanks

sokrypton commented 2 years ago

please reopen at ColabDesign github