Skip to content

covalent bond violation in predicted RNA structures #348

@sahakyanhk

Description

@sahakyanhk

I noticed a problem with phosphodiester bonds in artificial protein-RNA structures when the RNA forms a hairpin.

Here is the structure predicted with esmfold2-2026-05 for:

protein
YNQHRRLQDQLN
rna
GAGACGACCGGUCGUUGC

Image Image

But then, I noticed that this is a general problem for structures with strong complementarity in the hairpin loops; if you predict any sequence with a GCGC or GGCC in the loop, the phosphodiester bond will be broken.

e.g., here is GCGCGCGCGCGCGCGCGCGC

Image

Interestingly, that AUAUAUAUAUAUAUAUAUAU is fine, but if there are GC pairs in the loop (i.e. AUAUAUAUGCGCAUAUAUAU ), it will cause the same problem

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions