PHI high level term (PHI-base5)

[Hsinyugithub]

In the PHI high level term section, it seems to record all the "infective_ability" terms. Therefore, the terms from WT, deletion strains, complement strains are all recorded in this section. For instance in this case (https://test-phi5.phi-base.org/#/search-detail-page/PHIG:1655), we have both "Increased virulence" and "Reduced virulence" terms at the same time. I can delete the "infective_ability" from all entries that are not linked to WT pathogen genes, if there is not other solution to this. @jseager7 please let me know if I should go in and delete these terms.

https://canto.phi-base.org/curs/ee6b5e4f83bc0a4b/ro

[Hsinyugithub]

Just had another look. It's useful to know if the deletion strain increase or reduce virulence. However, the "infective_ability" term for complement strain might just confuse our users. Perhaps, I could just delete those ones?

[jseager7]

I think this is the expected behaviour of the 'PHI high level term' section. It's meant to aggregate all the high level terms that are used in any annotation on the page. It doesn't imply that both 'Increased virulence' and 'Reduced virulence' are present simultaneously. It only states that there are known mutations (or conditions, or experimental tests) that produce the given phenotypes, and that those phenotypes are found elsewhere on the page.

So in one case, the virulence is increased; in another case, it's reduced; and in another case, it's unaffected – but never in the same case.

We might need to clarify this on the Help page since the current wording doesn't make this clear:

PHI high level term: high-level terms from PHI-base 4 that describe the overall outcome of the interaction (e.g. Loss of pathogenicity) or that tag the gene as an effector.

---

However, the "infective_ability" term for complement strain might just confuse our users. Perhaps, I could just delete those ones?

I think it's fine to keep the infective ability annotations here, though perhaps complement strains should use the term 'restored pathogenicity' instead of something like 'increased virulence', since the reference point for the infective ability is more ambiguous for complement strains: presumably the reference point is normally a wild-type pathogen, but for complement strains the reference point is the mutant pathogen.

[Hsinyugithub]

I will try to replace "unaffected pathogenicity" with "restored pathogenicity". @jseager7 Is it possible to find a list of entries that have this high level term ("unaffected pathogenicity") from the entries that I curated? This could save me some time to find them.

[martin2urban]

I believe this ties into our earlier discussion on how to curate complemented deletion mutants. From my perspective, the gene-centric page is primarily concerned with the wild-type gene and the consequences of its mutation. In this context, complementation is a means of experimentally confirming the gene's function, rather than a phenotype attributable to the gene itself.

As such, I’m not convinced we need to explicitly annotate the result of complementation as “increased or restored virulence,” as this may be implicitly understood from the experimental design. That said, I’m open to other perspectives. If there is concern about potential confusion for users, perhaps this could be handled via annotation extensions to clarify the role of complementation without conflating it with the gene’s primary phenotypic annotation.

[jseager7]

In this context, complementation is a means of experimentally confirming the gene's function, rather than a phenotype attributable to the gene itself. As such, I’m not convinced we need to explicitly annotate the result of complementation as “increased or restored virulence,” as this may be implicitly understood from the experimental design.

This is a good point, but note that 'restored pathogenicity' will not (or should not) be displayed in the Entry Summary section since the website and data upload tool has no logic to handle it. So the risk of confusion for users will be minimised in that respect. While the restoration of pathogenicity can be implicitly understood by humans, it may not be as easy for machines or AI to understand (whereas it should be easier if the change in pathogenicity is made explicit through an annotation extension).

perhaps this could be handled via annotation extensions to clarify the role of complementation without conflating it with the gene’s primary phenotypic annotation.

The problem is that arguably, any annotation extension is somewhat conflated with the primary phenotype annotation (the PHIPO ontology term) due to the annotation extension being an extension of the primary phenotype annotation. Perhaps what we need to properly capture complementation is a new annotation type?

I was just looking at the PomBase issue tracker and a complementation annotation type seems to have been implemented for FlyBase, though that was almost six years ago, so I don't know if it's still supported. See pombase/canto#1985

[jseager7]

There is another option, which is to add an evidence code for 'Functional complementation'. The existing evidence codes in Canto seem to be sourced from the Evidence & Conclusion Ontology (ECO), just with the text 'evidence' removed from the term name.

ECO has a term for functional complementation evidence.

However, this will only work if we don't already need an evidence code for the phenotype that is linked to the complementation.

[martin2urban]

@jseager7 These are interesting options we can explore further.
Initially, my main concern was to introduce a new high-level gene annotation: 'restored virulence'.

I believe this concept sits at a different level than the existing high-level categories such as loss of pathogenicity, reduced virulence, increased virulence, and unaffected pathogenicity. Those describe the outcome of disrupting a single gene, whereas restored virulence typically applies to a second-site mutation or complementation event that reverses a previously observed phenotype which can happen at different levels: partially, 60%, or restored (100%).

I recall that there is a term that can be used in curation but is suppressed from display in PHI5. Perhaps something like that could be applied here as a temporary workaround until we decide how best to represent this concept?