Questions on design constraints: penalizing multi-residue motifs and suppressing interface beta-sheets #408

Hi,
thank you for this great work and making it publicly available

I am working on nanobody design against a target with an exposed edge strand using ESMFold2, and I have two questions regarding steering and constraints:

1. Sequence liability motifs
   Excluding single residues via logit masking (e.g., Cys) works smoothly. Is there currently a way—or any plans—to handle multi-residue motifs (such as N-X-S/T glycosylation sites) during design, e.g., via soft sequence penalties?

2. Intermolecular beta-sheet pairing with the target
   A significant portion of my top designs show CDR3 pairing with the target as an antiparallel beta-strand. This seems largely driven by the target's exposed backbone. Is this a known behavior with distogram-based loss, and is there a recommended way to suppress/discourage unwanted beta-sheet formation at the interface?

Thanks!