Getting oxygen from glycosydic bond `link` object

[NessimRaou]

Hi,

When using the glypy library to traverse a glycan's links with sugar.links.items(), it seems I can only retrieve the carbons from the link's parent (link.parent_position) and child (link.child_position) but not the oxygen from the acceptor sugar's hydroxyl group that is part of the glycosidic bond.

Is there something that I overlook to get that atom object with its numbered name (like "O6") ?

Thank you for reading me.
Nessim Raouraoua

[mobiusklein]

glypy doesn't use an atomic graph, only a rough molecular graph for its primary representation. This means individual atoms aren't addressable. glypy.Monosaccharide assumes that each backbone carbon has a hydroxyl group and a hydrogen atom unless there is a bond at that position, or it is the ring's anomeric carbon, which closes ring and doesn't have a hydroxyl group. When dealing open chain monosaccharides, the ReducingEnd modifies the total composition to correct for the regained hydroxyl group, but it is indirect.

What is your particular use-case?

[NessimRaou]

We develop a tool that wraps AlphaFold3 to parallelize its computations and make it easier to use for novice command line users.

One of its features is providing IUPAC formatted glycan to model glycoproteins structures. We construct the glycan with AlphaFold3's bondedAtomPairs entry, in which we specify the atoms (element + numbering in the structure, e.g. "O6") that bond together between sugars. To that end, we use glypy to parse the glycan, traverse it and obtain these information.

Currently, parent and child carbon from the link between sugars are the atoms we bond together while we should connect the oxygen of the base sugar to the carbon of the newly added sugar.
I don't know if we can use the carbon numbering to reliably get the oxygen name. That would be perfect but I don't find a clear statement that insures it, hence me trying to program a consistently right solution.