Misplaced MI:0789 (knock down), MI:0788 (knock out) is also a mutation

[lukasz99]

Knock downed (in contrast to knocked out) gene is not mutated but its level of expression is manipulated, typically, by siRNA or morpholinos. Thus the term should be made a child term of MI:0223) under expressed level.

Knocked out gene is a mutation and thus could be a child of MI:0113 (mutation). The problem is mutation terms combine mutation (as gene change) with its effect (increasing/decreasing/etc) on one specific process. In principle, mutation effects should be handled the same way as the effects of PTMs.

[pporrasebi]

These terms are both under the 'genetic experimental form' branch, which is meant to capture a terms related to genetic interactions exclusively. As such, these terms should not be applied to experimental preparations in physical interactions.

If you are fine with that, I will keep them as they are and close the issue.

[lukasz99]

On 7/30/19 9:05 AM, Pablo Porras wrote: These terms are both under the 'genetic experimental form' branch, which is meant to capture a terms related to genetic interactions exclusively. As such, these terms should not be applied to experimental preparations in physical interactions. If you are fine with that, I will keep them as they are and close the issue.

Pablo, the move of 'knock down' will be, effectively, also *within* 'genetic experimental form' (223 is a child of, both, 221 *and* 803). knock down is not a mutated gene as the cv seems to imply at the moment. this is exactly how the cv is supposed to work. lukasz

…

-- ------------------------------------------------------------------------- Lukasz Salwinski PHONE: 310-825-1402 UCLA-DOE Institute FAX: 310-206-3914 UCLA, Los Angeles EMAIL: [email protected] -------------------------------------------------------------------------

[pporrasebi]

I am not sure we are fully understanding the design choices behind this branch of the ontology. Perhaps @chris-grove can help here. Chris, did you also work on this when you did your work for genetic interactions in PSI-MI?

[chris-grove]

@pporrasebi
I didn't build this branch of the ontology. I would agree that "knock down" does not necessarily mean a mutated gene, but it could. I agree that "knock down" should not be a subclass of "mutated gene" (as it isn't necessarily the case). Perhaps "knock down" ought to be a synonym of "under expressed level" (MI:0223). I'm not sure there would be a meaningful distinction between "under expressed level" and "knock down".

[lukasz99]

On 08/01/2019 08:52 AM, Chris Grove wrote: @pporrasebi I didn't build this branch of the ontology. I would agree that "knock down" does not necessarily mean a mutated gene, but it could. I agree that "knock down" should not be a subclass of "mutated gene" (as it isn't necessarily the case). Perhaps "knock down" ought to be a synonym of "under expressed level" (MI:0223). I'm not sure there would be a meaningful distinction between "under expressed level" and "knock down".

sounds ok although I'm not sure if utilizing native regulatory mechanisms to decrease expression of a gene would be called a knock down. the most common use of the term seems to be in the context of siRNA-based depletion. apart from that, folding a term into a synonym makes it, effectively, disappear. it might make these who actually use it somewhat unhappy. though it would be still better than the current version. lukasz

[pporrasebi]

I would say that the distinction could be in the mechanism used to generate the under expression. I agree with Lukasz that I wouldn't call a knock down something you do to alter expression of a gene that is not happening at a genetic level (with a drug, for example). Knock down seem to have a very specific meaning in that it needs to be achieved via alteration at the genetic level, right?

So what I think is that we need to get rid of 'mutated gene' term (MI:0804), seems fairly meaningless, but I do not know if it has been used in any context, so I am a bit reluctant to make it obsolete.

I would propose changing this term to 'alteration at genetic level' or something similar (I am terrible with names, so please do not hesitate to lean in for this one). The definition would be "A modification that alters a gene function or expression at genetic level directly.". The terms a-/anti-/hyper-/hypomorph could join KD/KO under this definition. KO/KD can both be under 'under expressed level' because they are specific cases of under-expression. We would keep the 'expression alteration' branch separated and under both 'genetic experimental form' (for use in genetic interactions).

Does that sound reasonable to both of you?

[chris-grove]

Sorry for the delayed response. Given that I/we don't know how this branch has been used, I agree that deprecating the term without such an investigation would be premature. On the same lines, I would be hesitant to change the term label. I do agree, though, that "knock down" needs to move out from under "mutated gene", particularly since "knock down" is defined as reduction in gene expression by "introduction of an external substance, e.g. by RNA interference".

The branch "genetic experimental form" appears to be a place to collect types of genetic perturbations. I feel that "alteration at genetic level" would be redundant with the existing purpose of this branch and it wouldn't be clear what "at genetic level" means from the label alone and would have to be carefully defined, as it could be liberally interpreted (some would likely interpret "at genetic level" to mean alteration to the DNA sequence itself).

I would say "knock down" should either be made a child term of "genetic experimental form" or it should be made a child term of "under expressed level", as opposed to being made a synonym of that term. I think the latter is the most expedient and most correct, for the time being.