Incorrect UniProt KW mapping: DNA repair Source:UniProtKB-KW Inferred from electronic annotation

[ValWood]

• UniProt KeyWord (ID and label):

DNA repairSource:UniProtKB-KW
Inferred from electronic annotation

• Sequences with problematic annotation (ID + gene/protein name):

https://www.uniprot.org/uniprotkb/O60066/entry

• Description of issue

not "DNA repair"

Also, has
GO:0006307 | NOT DNA alkylation repair | IGI with mag1 | Korvald H et al. (2012)
for many years but this doesn't show up in uniprot

[cmungall]

Is this a mapping issue, or a uniprot annotation issue?

Can we work with uniprot to reverse propagate NOTs back to the KW mappings to remove erroneous KW annotations?

[ValWood]

I'm not sure. in most yeast , classical DNA methylation (such as 5-methylcytosine, 5mC) is either absent, and the lyase activity appears to be absent

Interestingly the human UniProt entry says

Dioxygenase that acts as on nucleic acids, such as DNA and tRNA (PubMed:18603530, PubMed:27497299, PubMed:27745969).
Requires molecular oxygen, alpha-ketoglutarate and iron (PubMed:18603530, PubMed:27497299).
A number of activities have been described for this dioxygenase, but recent results suggest that it mainly acts as on tRNAs and mediates their demethylation or oxidation depending on the context and subcellular compartment (PubMed:27497299, PubMed:27745969).
Mainly acts as a tRNA demethylase by removing N1-methyladenine from various tRNAs, with a preference for N1-methyladenine at position 58 (m1A58) present on a stem loop structure of tRNAs (PubMed:[27745969](https://www.uniprot.org/citations/27745969)).
Acts as a regulator of translation initiation and elongation in response to glucose deprivation: regulates both translation initiation, by mediating demethylation of tRNA(Met), and translation elongation, N1-methyladenine-containing tRNAs being preferentially recruited to polysomes to promote translation elongation (PubMed:27745969).
In mitochondrion, specifically interacts with mt-tRNA(Met) and mediates oxidation of mt-tRNA(Met) methylated at cytosine34 to form 5-formylcytosine (f5c) at this position (PubMed:27497299).
mt-tRNA(Met) containing the f5c modification at the wobble position enables recognition of the AUA codon in addition to the AUG codon, expanding codon recognition in mitochondrial translation (PubMed:27497299).
Specifically demethylates DNA methylated on the 6th position of adenine (N6-methyladenosine) DNA (PubMed:30017583, PubMed:30392959).
N6-methyladenosine (m6A) DNA is present at some L1 elements in embryonic stem cells and probably promotes their silencing (By similarity).
Demethylates mRNAs containing N3-methylcytidine modification (PubMed:31188562).
Also able to repair alkylated single-stranded DNA by oxidative demethylation, but with low activity (PubMed:18603530).
Also has DNA lyase activity and introduces double-stranded breaks at abasic sites: cleaves both single-stranded DNA and double-stranded DNA at abasic sites, with the greatest activity towards double-stranded DNA with two abasic sites (PubMed:19959401).
DNA lyase activity does not require alpha-ketboglutarate and iron and leads to the formation of an irreversible covalent protein-DNA adduct with the 5' DNA product (PubMed:19959401, PubMed:23577621).
DNA lyase activity is not required during base excision repair and class switch recombination of the immunoglobulin heavy chain during B lymphocyte activation. May play a role in placental trophoblast lineage differentiation (By similarity).

I wonder how much of the DNA repair is physiologically relevant. It seems to be mainly about removing 3-methylcytidine (m³C) modification from tRNAs and mRNAs

Probably the evidence isn't strong enough for a "DNA repair " keyword, because it's clearly not the main role, and may be an artefact.