Design cyclic peptides with disulfide cyclization

[lelaihoangson]

I was wondering if we could generate the peptide's backbone and specifically put 2 Cys residues on 2 ends to cyclize it with a disulfide bond. Based on what I read, I assume we can only cyclize the peptide using the backbone right?

Thank you

[rclune]

Since RFdiffusion only generates backbone structures, not sequences, there is no way to tell it to include 2 Cys residues in the generated portions of the structure.

There may be other work arounds. One thought is to use sequence generation tools such as LigandMPNN to take the generated backbones and bias them towards having cysteine residues in specific locations (see option --bias_AA_per_residue). Another is to play tricks with the input structure for RFdiffusion, though I'm not sure how feasible this idea is.