parentage of ADP ribosylation

[ValWood]

This is a child of "protein glycosylation" which seems wrong, so it seems that "protein glycosylation" should be resticted further in some way?

[ValWood]

results in GPs like Tryptophan--tRNA ligase, cytoplasmic P23381 DNA repair protein XRCC1 P18887 Poly [ADP-ribose] polymerase 1 P09874 being annotated to protein glycosylation. Seems wrong? but maybe its OK? It might look odd to me because maybe fungi don't have this particular protein modification?

[pgaudet]

Indeed, "protein ADP ribosylation" is an asserted subclass of "protein glycosylation". This does seem wrong (dates from 2012); it should be directly under "protein modification process", right ?

Thanks, Pascale

[deustp01]

From Wikipedia: Many different amino acid side chains have been described as ADP-ribose acceptors. From a chemical perspective, this modification represents protein glycosylation: the transfer of ADP-ribose occurs onto amino acid side chains with a nucleophilic oxygen, nitrogen, or sulfur, resulting in N-, O-, or S-glycosidic linkage to the ribose of the ADP-ribose[14]. Originally, acidic amino acids (glutamate and aspartate) were described as the main sites of ADP-ribosylation. However, many other ADP-ribose acceptor sites such as serine[15][16], arginine[17], cysteine[18], lysine[19], diphthamide[20], phosphoserine[21], and asparagine[22] have been identified in subsequent works. (https://en.wikipedia.org/wiki/ADP-ribosylation)

Maybe a workaround is to distinguish glycosylation processes that incorporate an adenine as part of the glyco moiety from ones that incorporate only monosaccharides and derivatives of them.

[pgaudet]

Right - another point where biologists and chemists diverge..... :(

[ValWood]

I guess my issue is that the ADP-ribosylation and the protein glycosylation appear to occur in completely different contexts.

I don't mind the "nucleotide" part so much

GDP-Man:Man1GlcNAc2-PP-Dol alpha-1,3-mannosyltransferase activity for example, sits comfortably in canonical "protein glycosylation"

but the "ribose" instead of the "hexose" addition appears to be a strong biological partitioning (in ER as part of glycosylation of cell surface proteins vs ribose as part of nuclear event, ( as far I can see...... but I would need to look into this further)

[deustp01]

That's the beauty of the nucleotide part - without meaning to, it gives you exactly the differentium (?) that you need without further looking, because the ER events never incorporate one into a glyco-whatever-modified protein.

[ValWood]

Of course, got you!

[pgaudet]

Fixed by https://github.com/geneontology/go-ontology/pull/15396