The metacaspase–Peps–PEPR immune module confers resistance to Fusarium head blight in wheat (PMID: 40674698)

[Hsinyugithub]

New extension needed to curate this kind of data (pre-treatment with host elicitor peptides). In PHI-Canto, we can already add "PTI inducer flg22" in the "Conditions" (see below).

To curate the elicitor peptides, we could have "host elicitor peptide: XXXX" in the conditions extension.

Or use an annotation extension, "in the presence of" XXXX.

However, we will need to modify PHI-Canto to be able to replace XXXX with any given peptide. We will need a way to record these peptides as well. We could use, for instance, A0A3B6IL84 (100-121) for now.

TacPep-a EVRSRGVPQRPRAPREGRGGTHH

The other thing, we have discussed before, about creating an entry or making metagenotype without a pathogen gene. This will be the case here.

[jseager7]

@Hsinyugithub If we need to be able to use an arbitrary peptide, then an annotation extension will be the best choice.

I'm not sure about the name; maybe something like 'with_host_peptide' to match the wording of our existing extensions 'with_host_species'. For the full name in the annotation extension popup we could be more specific and use 'with host elicitor peptide'.

We'd need to decide on a controlled way to write the peptide. It seems like the ProteinName(start-end) notation that you suggested is fairly standard. I'd recommend using the UniProtKB accession number as the protein name.

Ideally we would hint at this notation in the name of the extension. Maybe something like this, but I'm not keen on the double brackets.

with host elicitor peptide (UniProtKB_ID(start-end))

[Hsinyugithub]

@jseager7 I agree that the annotation extension "with host elicitor peptide (UniProtKB_ID(start-end))" might be the best way at the moment. Is it easy to implement this in PHI_Canto? We can discuss this during our meeting on Monday.

[jseager7]

Is it easy to implement this in PHI_Canto?

@Hsinyugithub Yes, adding new annotation extensions is easy, especially when they will be entered as text (as this one will be). If the rest of the team agrees that an annotation extension is the best way to capture peptide information then I can add it straight away.

[jseager7]

@Hsinyugithub Here's the configuration for the new peptide annotation extension. Please can you give feedback on the following points:

The domain ID is the PHIPO term that the annotation extension will be available for. I've set this to PHIPO:0000001 so the extension will be available for 'pathogen host interaction phenotype' and any of its child terms. If the extension should be on single species phenotypes instead, please let me know.

If the extension is meant to be for pathogen host interaction phenotypes, please also let me know if you want the extension enabled for gene-for-gene phenotypes as well. If not, it can be restricted to just pathogen host interaction phenotypes or just gene-for-gene phenotypes.

The cardinality is the number of extensions of this type that can be added for one annotation. Currently it allows zero or one peptide extension per annotation. If you think you'll need more than one peptide per annotation then I can remove this limit.

domain ID	subset relation	extension relation	range ID	Canto display text	Help text	cardinality	role	annotation_type_name
PHIPO:0000001	is_a	with_host_peptide	Text	with host elicitor peptide	the UniProtKB accession number for the protein containing the peptide, then the start and end of the range of amino acids, in parentheses (100-200)	0,1	user

[Hsinyugithub]

@jseager7 All look sensible to me. I could try to curate one entry to see if there are any problems with this approach.

[jseager7]

@Hsinyugithub Thanks. I'll test the annotation extension in the test version of PHI-Canto today and if all goes well the extension should be available for testing in a curation session tomorrow morning.

[jseager7]

@Hsinyugithub The with_host_peptide annotation extension should be available on any pathogen-host interaction phenotype term now, in case you want to test it with an entry.

[Hsinyugithub]

[Hsinyugithub]

Without the placement gene for the pathogen:

We still need a placement gene for host in order to make the metagenotype. In these case, I used the WT host genes as the placement genes. Not perfect, but not incorrect.

[jseager7]

@Hsinyugithub If you've found cases where both sides of the metagenotype need to be wild-type, it's possible to remove the restriction on needing a gene in either the pathogen genotype or the host genotype.

But I think we should discuss this first to make sure it's really needed, because allowing fully wild-type interactions could lead to community curators leaving out important information about mutations.