search

geneontology / go-ontology

Source ontology files for the Gene Ontology
http://geneontology.org/page/download-ontology
Creative Commons Attribution 4.0 International
216 stars 40 forks source link

NTR: Negative Regulation of CRISPR-cas system #16423

Closed jdc-cpt closed 2 years ago

jdc-cpt commented 5 years ago

New Term: Negative regulation of CRISPR-cas system Definition: Negative regulation of an adaptive immune response of bacteria that serves to clear host cells of foreign DNA and RNA. It has three distinct stage: acquisition of foreign DNA by integration into CRISPR loci in the host chromosome, CRISPR RNA (crRNA) biogenesis, and target interference. CISPR stands for Clustered Regularly Interspaced Short Palindromic Repeat, which describes the nature of the loci. Notes: This reference describes the function of anti-CRISPR proteins within a bacterial genome to protect itself from its CRISPR system. PMID: 30190308, 30190307 @sandyl27 @jimhu-tamu This is for CACAO

pgaudet commented 5 years ago

Shouldn't it be 'defense response to against CRISPR-cas system', rather than just negative regulation?

jimhu-tamu commented 5 years ago

We didn’t think “response to” was appropriate here. These are from species where the guide RNA would target somewhere in their own genome. The anti-CRISPRs that are presumably constitutively expressed to keep the CRISPR cas from causing suicide. No phage or mobile element is being responded to. I think of it as preventing an autoimmune response.

Other anti-CRISPRs are carried by viruses for overcoming host defense, but not these.

jrr-cpt commented 5 years ago

@vanaukenk Any progress on this NTR?

vanaukenk commented 5 years ago

@jrr-cpt

I've added regulation of CRISPR-cas system and negative regulation of CRISRP-cas system to the ontology.

In looking over what we have more generally for the CRISPR-cas system in GO, I came across the term 'evasion by virus of CRISPR-cas system':

http://amigo.geneontology.org/amigo/term/GO:0098672

There are no annotations to the 'evasion' term, but I am wondering if and how this term and the negative regulation of CRISPR-cas system should be related in the ontology.

The definition for the 'evasion' term seems quite broad, but the associated references suggest some commonalities between it and the negative regulation term.

I am concerned about having two conceptually related terms that have no relationship in the ontology, but it is hard to know exactly how the evasion term was meant to be used with no associated annotations.

Do you have any thoughts about this?

Also tagging @SIBvirus for input here.

Possibly we will need to think about having more specific child terms of 'negative regulation of CRISPR-cas system' to account for the different mechanisms, and purposes, of the regulation but I'd like to know more about the processes before doing that.

Thx.

vanaukenk commented 5 years ago

Also, @jimhu-tamu

If bacteria encode anti-CRISPRs that serve to essentially prevent bacterial suicide, do you think our placement of CRISPR-cas system in the ontology as a defense response is still correct?

image

jrr-cpt commented 5 years ago

@vanaukenk In short, I think that GO:0098672 'evasion by virus of CRISPR-cas system' is not a good term, rather using GO-CAM (?) use with 'negative regulation of CRISRP-cas system' would better capture what we know now and what we suspect we might find later.

See PMID: 20598393 for an early (2011) look at self-targeting spacers. While many are actually targeting mobile genetic elements or proviruses, others are targeting cellular genes. This phenomenon was used as the strategy in PMIDs: 30190308, 30190307 referenced above for finding proteins that inhibit CRISPR function. What we do not know is whether all these self-targeting spacers were acquired from the chromosome, or came in with a transducing phage/mobile genetic element. I am not sure how much we know about whether those self-targeting spacers could actually target the chromosome genes (dependence on PAM), or whether all genomes with self-targeting spacers have mechanisms to repress cleavage by the CRISPR-Cas complexes. Some postulate that this could be a gene regulation mechanism, though I haven't seen evidence for it.

About classifying it under defense response, not sure.

cmungall commented 4 years ago

@jrr-cpt

I think that GO:0098672 'evasion by virus of CRISPR-cas system' is not a good term

I agree! To me 'evasion' is a bit vague, it encompasses things like mutations in the target sequence, which isn't even a biological process

(((Ontologists side note: this term has an invalid has-participant relationship to a process

relationship: has_participant GO:0099048 ! CRISPR-cas system

we will catch these when we switch to Elk5)))

rather using GO-CAM (?) use with 'negative regulation of CRISRP-cas system' would better capture what we know now and what we suspect we might find later.

Yes! even without GO-CAM or extensions I think it is good to be more precise than 'evade'. If it's negatively regulating or interfering we should say it!

@vanaukenk

There are no annotations to the 'evasion' term

actually some were added after you commented 2 years ago! And unfortunately they chose the 'evasion' term rather than your more appropriate negative regulation term

note there is a keyword mapping to the evasion term, but not keyword mapping to the evasion term

The majority of our experimental annotations to 'evasion'/GO:0098672 come from this paper:

http://amigo.geneontology.org/amigo/reference/PMID:30046034

This paper resulted in annotations only to GO:0098672, no other terms

From my reading of the paper it's clear the mechanism is known, I would not say it is "evasion", it is by targeting the nuclease activity of Cas9. This could be captured by go-cam or extensions, as jrr-cpt suggests.

But even just with a simple annotation, these annotations should have been to a more mechanistically precise term.

I prefer the newer neg-reg-of-CRISPR term, but what I don't like is that it's still a bit vague about the mechanism. We use neg-reg a lot of the time when we don't know what's happening. I think here it's clear, it's interfering with the nuclease activity?

A bold solution would be to merge the the existing 'evasion' term into the new unused 'negative regulation' term -- and maybe even call the term "anti-CRISPR", this is what is used in the literature.

This might make some uneasy, as we'd be making a stronger statement (I haven't fully absorbed the thoughtful comments and about self targeting spaces). But there is something to be said for having fewer terms to choose from!

cmungall commented 4 years ago

@vanaukenk the placement of CRISPR-cas system (see above in https://github.com/geneontology/go-ontology/issues/16423#issuecomment-438764909) feels a little odd to me. The 'clearance' part seems odd, even when it's targeted against phages.

it seems like it should itself negatively regulate some process, such as viral lifecycle (when not employed to avoid self-suicide)

the def has a typo (CISPR):

def: "An adaptive immune response of bacteria that serves to clear host cells of foreign DNA and RNA. It has three distinct stage: acquisition of foreign DNA by integration into CRISPR loci in the host chromosome, CRISPR RNA (crRNA) biogenesis, and target interference. CISPR stands for Clustered Regularly Interspaced Short Palindromic Repeat, which describes the natur

also the def states foreign, which sounds like not always true...

pgaudet commented 4 years ago

@ValWood we should look at this.

@vanaukenk I am stealing this as part of the multiorg work. I hope that's OK.

ValWood commented 4 years ago

Yes I saw that go by this morning. Chris has a good point about "evasion"

vanaukenk commented 4 years ago

@vanaukenk I am stealing this as part of the multiorg work. I hope that's OK.

@pgaudet that is fine since I am no longer on the ontology rota

pgaudet commented 4 years ago

These 2 related terms have not been used:

Look into merging - and look for what would be the best way to formulate this.

@pmasson55 There is a link to https://viralzone.expasy.org/3962?outline=all_by_protein

Maybe 'defense against (host) defense system/response'?

@dsiegele @mgiglio99 Thoughts on this ?

pgaudet commented 3 years ago

Discussion on multiorganism call:

pgaudet commented 3 years ago

@addiehl Does the proposal above work for you ? This puts those terms under 'immune system development', which is incorrect since this is under multiorganism process. Do you have suggestions as to how to deal with this ?

addiehl commented 2 years ago

Not sure if it's too late to comment on this, but I can see there are outstanding issues.

A fundamental issue here is the definition of what a "system" is in the GO. Per the gloss on the definition of GO:0048731 'system development' the working definition of system for this term is "A system is a regularly interacting or interdependent group of organs or tissues that work together to carry out a given biological process."

This is fundamentally at odds with systems like CRISPR-CAS that work on the molecular or cellular levels, sometimes in unicellular organisms. Examples of GO terms that make reference to systems which are not systems per the above definition (this not necessarily an exhaustive list): GO:0099048 CRISPR-cas system GO:0012505 endomembrane system GO:0009521 photosystem GO:0009522 photosystem I GO:0009523 photosystem II GO:0030253 protein secretion by the type I secretion system GO:0015628 protein secretion by the type II secretion system GO:0030254 protein secretion by the type III secretion system GO:0044097 secretion by the type IV secretion system GO:0030255 protein secretion by the type IV secretion system GO:0044098 DNA secretion by the type IV secretion system GO:0046819 protein secretion by the type V secretion system GO:0062051 lipopolysaccharide transport system GO:0000160 phosphorelay signal transduction system GO:0071629 cytoplasm protein quality control by the ubiquitin-proteasome system GO:0071630 nuclear protein quality control by the ubiquitin-proteasome system

While we typically think of an immune system as a system as a system that occurs in multicellular animals, at least some or many plant biologists like to talk about plant immune systems (which makes sense given that many innate immune recognition strategies are shared by plants and animals), and at least some microbiologists talk about CRISPR-Cas and other molecular level defense systems as immune systems of bacteria (nicely reviewed in PMID:31695182, among other places). CRISPR-Cas even exhibits memory of past viral infections.

A better working definition of system for the GO would be "A system is a regularly interacting or interdependent group of lipid membranes, nucleic acids, proteins, cells, organs, or tissues that work together to carry out a given biological process."

We could then move 'system development' directly under 'developmental process' and broaden the definition to be "The process whose specific outcome is the progression of a system over time, from its formation to the mature structure. A system is a regularly interacting or interdependent group of lipid membranes, nucleic acids, proteins, cells, organs, or tissues that work together to carry out a given biological process."

We could even add a child (thinking fondly of @ukemi), 'multicellular organismal system development' with the definition "The process whose specific outcome is the progression of an organismal system over time, from its formation to the mature structure. An organismal system is a regularly interacting or interdependent group of organs or tissues that work together to carry out a given biological process." This term would also be under 'multicellular organism development'

Then we would keep 'immune system development' as a child of 'system development' and move all the rest of the current children of 'system development' under 'organismal system development' ('animal organ development', 'limbic system development', 'vasculature system development', etc.)

An additional change would be to definitions of 'immune system development' to include the broader definition of system: "The process whose specific outcome is the progression of an organismal system whose objective is to provide calibrated responses by an organism to a potential internal or invasive threat, over time, from its formation to the mature structure. A system is a regularly interacting or interdependent group of lipid membranes, nucleic acids, proteins, cells, organs, or tissues that work together to carry out a given biological process."

With these changes we can add 'tolerance to self CRISPR-cas system' as a child of 'tolerance induction'.

New hierarchy: biological_process --developmental process ----system development ------immune system development --------tolerance induction ----------tolerance to self CRISPR-cas system ------multicellular organismal system development [new term] ----multicellular organismal process ------multicellular organism development --------multicellular organismal system development [new term] --immune system process ----immune system development ------tolerance induction --------tolerance to self CRISPR-cas system

Perhaps someday there will be additional children of 'system development' that aren't children of 'multicellular organismal system development'

Thanks for your tolerance of my late reply and long winded response and daring to suggest changes to the development process hierarchy! Alex

cmungall commented 2 years ago

Perhaps we can obsolete "system development"? I think your broadened definition makes sense but it's so broad it sure covers everything in GO, it's all systems biology.

pgaudet commented 2 years ago

Discussion on the multiorganism call today:

pgaudet commented 2 years ago

Dear all,

The proposal has been made to obsolete GO:0110132 regulation of CRISPR-cas system and GO:0110133 negative regulation of CRISPR-cas system

The reason for obsoletion is that these terms do not clearly capture the processes described in papers cited. The existing term GO:0098672 "evasion by virus of CRISPR-cas system" is more appropriate.

There are no annotations, no mappings to these terms. These terms are not present in any subsets.

You can comment on the ticket: https://github.com/geneontology/go-ontology/issues/16423

Thanks, Pascale

pgaudet commented 2 years ago

+[Term] +id: GO:0140721 +name: nuclease inhibitor activity +namespace: molecular_function +def: "Binds to and modulates the activity of a nuclease." [PMID:28785032, PMID:29554913, PMID:30046034] +intersection_of: GO:0098772 ! molecular function regulator +intersection_of: negatively_regulates GO:0004518 ! nuclease activity +property_value: term_tracker_item https://github.com/geneontology/go-ontology/issues/16423 xsd:anyURI +created_by: pg +creation_date: 2021-11-16T09:09:45Z

cmungall commented 2 years ago

Hmm, I am not so sure about this - I thought we all agreed "evasion" was very vague. There is a very precise mechanism with anti-CRISPR. I am also not sure about distributing the concept of anti-CRISPR across two branches

pgaudet commented 2 years ago

Not sure what you refer to. I have changed of label GO:0098672 "evasion by virus of CRISPR-cas system" to "inhibition of host CRISPR-cas system by virus" - seems like we're saying the same thing?