fuse fails when attaching C-terminal NMe residue

[mtessmer]

Describe the bug

PyMol fails to attach amino acids to preexisting structures using builder or editor cmd interfaces.

To Reproduce

Steps to reproduce the behavior:

1. fetch 1ubq
2. editor.attach_amino_acid('last name C', 'nme')

returns:
Error: no coordinate for source anchor atom

Expected behavior

The NMe or other residue should be attached to the structure

Environment

• **PyMOL Version: 3.1.0a
• **PyMOL Source: Gholke pymol-open-source-wheels (https://github.com/cgohlke/pymol-open-source-wheels)
• **Operating System: Windows 11

Additional context

Attaching amino acids seems to work fine when using a peptide built entirely from the builder interface but breaks when using PDB or CIF files from the PDB.

[JarrettSJohnson]

Thanks for the report; there's unfortunately a couple of issues underlying here from what I see. One of them involves keeping around invalid fragments when a function has failed. I'll open an issue for awareness about this one. The specific reason though why this fails seems to be that builder fragments have their terminal oxygens removed to provide an open valence for amino acid attachments.

Thus,

fetch 1ubq
remove last name OXT
editor.attach_amino_acid('last name C', 'nme')

might be able to get you what you need here (assuming the final oxygen name is OXT).

Since we provide the spherical indicators that valid attachments should occur at that location, PyMOL should allow connections there. I'll see if I can just automatically replace these oxygens with the new fragment.