Hi, this is a question that I have asked before and I know it has been discussed in some other threads, however, I don't feel there is a sufficient solution yet.
Given a receptor, I would like to de-novo design a binder in a particular spot. I have tried two approaches.
De novo generation with no hotspots: This works well, simply using a receptor PDB with one chain and creating a new chain in the Chroma API. However, there is no way to specify where to place the chain, so only 1 in 20 binders are in the right region.
I have tried supplying a PDB with a complex of the receptor and existing binder, and used the substructure conditioner to only keep one residue from the existing binder, to sort of ground the Chroma diffusion in the right place. While this works in keeping the Chroma design in the binding pocket, the binders that Chroma generates are not good, usually being long, improbable single alpha helixes despite using the CATH conditioner for alpha helix bundles. (in contrast, in the first approach the CATH conditioner works really well to design good looking binders, albeit in the wrong spot).
Please let me know if there is any recommended solution for this. Thank you! Chroma is awesome!
Hi, this is a question that I have asked before and I know it has been discussed in some other threads, however, I don't feel there is a sufficient solution yet.
Given a receptor, I would like to de-novo design a binder in a particular spot. I have tried two approaches.
Please let me know if there is any recommended solution for this. Thank you! Chroma is awesome!