Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Aug 23 16:32
    pgaudet closed #2605
  • Aug 23 16:32
    pgaudet commented #2605
  • Aug 23 16:14
    pgaudet closed #2612
  • Aug 23 15:58
    pgaudet edited #2588
  • Aug 23 15:45
    pgaudet labeled #2614
  • Aug 23 15:45
    pgaudet assigned #2614
  • Aug 23 15:45
    pgaudet opened #2614
  • Aug 23 15:43
    pgaudet edited #2613
  • Aug 23 15:18
    pgaudet edited #2588
  • Aug 23 14:25
    ukemi unassigned #2612
  • Aug 23 14:05
    ukemi commented #2607
  • Aug 23 14:03
    pgaudet commented #2607
  • Aug 23 13:59
    ukemi commented #2607
  • Aug 23 13:46
    ukemi commented #2607
  • Aug 23 13:45
    pgaudet commented #2607
  • Aug 23 13:38
    ukemi commented #2607
  • Aug 23 13:14
    ukemi commented #2607
  • Aug 23 08:50
    pgaudet commented #2598
  • Aug 23 08:45
    ValWood closed #2260
  • Aug 23 08:45
    ValWood commented #2260
sabrinatoro
@sabrinatoro
@goodb That works for me.
@vanaukenk @ukemi : is this is the official convention?
vanaukenk
@vanaukenk
@sabrinatoro - afaik, we are still modeling this using the enabled_by relation for the MF. In that case, I would make two separate protein binding annotations with each protein as the enabler and the interacting protein as the has_input.
goodb
@goodb
@vanaukenk is there documentation on this on the wiki somewhere ?
vanaukenk
@vanaukenk
@goodb - specifically for the protein binding MF annotations, no.
goodb
@goodb
@vanaukenk seems a little wonky to use enables on this to me. Out of curiosity is that choice made because it results in gpad output? the has_input structure above won’t produce any gpad as things stand.
vanaukenk
@vanaukenk
@goodb - yes, that is currently the only way we'll get the expected GPAD output, although that's not to say we can't re-visit the rules for this.
sabrinatoro
@sabrinatoro
@vanaukenk Thank you!
vanaukenk
@vanaukenk
@goodb - I do think a broader discussion of how 'protein binding' annotations fit into GO-CAM models is warranted. One issue I'd like to sort out is: when is protein binding really the main MF of a gene product vs when is protein binding just supportive evidence for a direct regulatory relationship between two activities? If only the latter, could the binding be used as evidence for the relation, but then could we still get the conventional binding annotations back out in the GPAD?
goodb
@goodb
Hi @vanaukenk agree that this ought to be nailed down. Perhaps a subject of your annotation call to resolve the question? My take based on how we are approaching the reactome information and my understanding of the go-cam model would be (1) to model the case @sabrinatoro brought up as I proposed because there is no evidence of signaling or that the “function” of either protein is really to bind. Its just a starting point for building up a more complete model (and/or its a pattern for the GO to record PPIs which aren’t really “functional” but are useful information). If we need to make tabular output to link these proteins to “protein binding” then I think that is a rule we can add to the exporter. As I understand things, we should only be using ‘enables’ when linking a gene product to what is understood as its purposeful function. (2) After spending a few hours in the room with Paul, I don’t think the function of many proteins is “to bind”. That is just something they do along the way to achieving their actual purpose. So i’d suggest focusing on the structuring of the regulatory relationships as the main priority. In that case, IMO I don’t like mixing up the biological model with the evidence model. Would rather not use binding as evidence unless that was explicity pulled out of the go-cam and into ECO. Probably easier if we had a white board or shared some screentime.
Petra Fey
@pfey03
Interesting discussion and thought about that a while back. Sometimes we do not have any other function (yet) but the binding. And when you say binding as evidence, can we still add the protein binding partner? Good to discuss in annotation call(s).
sabrinatoro
@sabrinatoro
I am having some difficulties creating a model for the following statement.
"caspy2 binds directly to lipopolysaccharide, resulting in caspy2 oligomerization, which is critical for pyroptosis" (PMID:30076291)
I started a GOCAM model, but I am not sure it is correct: gomodel:5c4605cc00004193
Could someone review this model, please?
Also, I can imagine this might require more discussions,... maybe at one of the GOCAM calls?
Thank you!
vanaukenk
@vanaukenk
@sabrinatoro - we could discuss this on tomorrow's GO-CAM call if you like!
sabrinatoro
@sabrinatoro
@vanaukenk : if there is time, it would be great! (we can at least start the discussion) Thanks
vanaukenk
@vanaukenk
Great; let's do it!
cerivs
@cerivs
Another question about binding and modeling protein interactions. Gene is asah1b Drer (ZFIN:ZDB-GENE-040426-1512)
"The difference between masses on size-exclusion chromatography, SDS-PAGE suggested that zebrafish AC exists as a homodimer of heterodimer (α+β) having two α-subunit and two β-subunits. Our findings resemble previous findings of the human AC as homotrimers of heterodimer in urine"
The homodimers form from the same polypeptide " Majority of the purifiedenzyme was obtained in the unprocessed form; however, small amounts of α-and β-subunits were also obtained".
Right now I just say protein herodimerization in my model since they showed that 2 ways. I'm not sure how much detail is helpful for users.
sabrinatoro
@sabrinatoro
Hello everyone!
Is there an existing GO-CAM model for ligand-receptor pathway that can be used/looked at as a reference? @vanaukenk
Jim Balhoff
@balhoff
In a GAF file, can an annotation extension contain both commas and pipes, e.g. rel(term),rel(term)|rel(term),rel(term)?
I looked here (http://geneontology.org/docs/go-annotation-file-gaf-format-2.1/#annotation-extension-column-16) but it doesn’t really have a precise grammar
Petra Fey
@pfey03
In P2GO, we distinguish between AND (comma) and OR (pipe)
Jim Balhoff
@balhoff
Is it legal to have an OR of two ANDs? (a&b) or (x&y)
Petra Fey
@pfey03
Yes I think so. At least it’s possible to do in P2GO
Jim Balhoff
@balhoff
thanks
vanaukenk
@vanaukenk
@balhoff Yes, it is possible to have both commas and pipes in the annotation extensions. The pipes are meant to signify independent extensions, while the commas indicate extensions that, together, provide context for the GO term. I'll double-check our GPX2.0 specs to make sure we're clear on this, but let me know if you want any clarification.
Jim Balhoff
@balhoff
@vanaukenk I need to parse these in a tool so I’m trying to make sure I know exactly what can be there. If both are used in one extension, do commas take precedence over pipes? Is it possible to group using parentheses? I thought I had seen a more explicit grammar doc at one point but I can’t find it since the website overhaul. This page (http://geneontology.org/docs/go-annotation-file-gaf-format-2.1/) isn’t explicit enough.
vanaukenk
@vanaukenk
@balhoff I'm not sure what you mean by precedence. Maybe we could Skype?
At the bottom of the gpad table is the current explanation of annotation extensions format, but we can make this clearer, if needed: https://github.com/geneontology/go-annotation/blob/master/specs/gpad-gpi-2-0.md
Jim Balhoff
@balhoff
@vanaukenk thanks, that’s what I was looking for
Annotation_Extensions ::= [Extension_Conj] ('|' Extension_Conj)*
vanaukenk
@vanaukenk
Okay; just let me know if you have any other questions.
Jim Balhoff
@balhoff
For precedence I mean you first group by commas, then those groups are OR'd
I think that’s what the spec is saying
vanaukenk
@vanaukenk
Yes, for precedence that's correct: extensions are first grouped by commas, and then those groups are separated by OR.
Petra Fey
@pfey03
Hi All, working on the oligomerization review and we saw and added a few GO:0042802 identical protein binding. The question we have: use IDA orcIPI? I think IDA should be fine as the same protein in the WITH seems just redundant. But there are more IPI out there. Thanks!
Birgit Meldal
@bmeldal
@pfey03 With my interaction head on, if the expt shows physical binding it's IPI... and you need the partner in the WITH field for those who parse the data out. Identical protein binding is very common, hence why I keep saying we need to capture homomeric complexes :) It would be different if it was self-interaction of the same molecule, then you can't have a GP in WITH but I don't know if GO captures that. It would require an exception rule. We do :) (and we may export this data to GO...)
Petra Fey
@pfey03
@bmeldal ok, since nobody else commented I changed mine that I did recently to IPI. All have been shown to self interact
vanaukenk
@vanaukenk
@pfey03 We are recommending that 'identical protein binding' be annotated with IPI and the ID of the same annotated entity be included in the With field for exactly the reasons Birgit states. We need to be consistent in how we capture physical interactions as much as possible, even if the binding is to the same protein. I know in some cases we have said that when we can't assign a specific ID to an interactor, e.g. actin, it is okay to annotate to a term like 'actin binding' using IDA, but I think it'd be good to enumerate these specific cases so there is no confusion about why we think these IDA annotations are okay.
Petra Fey
@pfey03
@vanaukenk Yes, I saw her point and therefore updated those today. But we might have legacy of IDA and I saw others too when I first quickly checked all EXP annotations to GO:0042802. And I agree with the actin example, Dicty has 29 actin genes, 17 make identical proteins.
Birgit Meldal
@bmeldal
We have a discussion here in the office : in 21844191, Fig 7 shows that reelin binding its receptor (binding shown in F6) indirectly activates Dab1 phosphorylation (which is much much weaker if reelin is mutated). I annotated it to GO:0038026 reelin-mediated signaling pathway as IDA (maybe it should be IMP - different discussion!) but Livia questions if we can say this is a direct evidence for reelin's involvement in the pathway. Discuss!
http://www.jbc.org/content/286/40/35247.long
sabrinatoro
@sabrinatoro
I am creating a model which requires a lot of adding individuals and making relationship. Every time I add an individual or creating a new relationship, the view of the whole model changes, and everything is sent WAY far right to the right, making it extremely difficult (and frustrating).
Is there any way around this? Is there a way to stop the "boxes" to change location everytime I do something? Thanks!
kltm
@kltm
@sabrinatoro This is an outstanding issue that we still need to get to: geneontology/noctua#173
In the meantime, there are workbenches that can help rally and control the node layout ("Cytoscape layout tool"). Of things should more-or-less fit on the screen, you can also use "Compact layout" under "Tools".
sabrinatoro
@sabrinatoro
Thank you @kltm . This is not great (as the "compact layout" mix up everything), but it is better that what I have been dealing with.
Does the "Cytoscape layout tool" let you make changes (or is it only a view option)?
kltm
@kltm
You can rubberband, move things as groups, sort however you want, and then save the new layout (as long as you are logged in).
sabrinatoro
@sabrinatoro
OK. Thank you for your help!
kltm
@kltm
NP!
Eric Douglass
@dougli1sqrd

Hello! In GORULE:0000015, we check that when a GAF has two taxon IDs, that they are formed correctly. I had a couple questions to clarify what the intention is. Here's the link to the rule: https://github.com/geneontology/go-site/blob/master/metadata/rules/gorule-0000015.md

Is this rule saying that 1) the first taxon should be attached to the gp, and 2) the second should be the interacting species (is that just what it means, or is there a way to verify this?) 3) the 2 allowed terms that can be used with dual taxons ('GO:0044419 : interspecies interaction between organisms', and 'GO:0044215 : other organism'), is that to mean only those terms are allowed? Or are children terms also allowed?

vanaukenk
@vanaukenk
@dougli1sqrd - As I understand the rule, the first of the pipe-separated taxon IDs should correspond to the taxon of the annotated gp, and the second taxon to that of the other species involved in the process. Children of the two GO terms listed in the description are also allowed for annotation. The statement about ancestor is the relevant part for this: "These annotations should be used only in conjunction with terms that have the biological process term 'GO:0044419 : interspecies interaction between organisms' or the cellular component term 'GO:0044215 : other organism' as an ancestor." Let me know if you want to touch base to discuss further. Thx.
pgaudet
@pgaudet
@dougli1sqrd Which ticket are you working on ?
I put a lot of tickets in the GO-rules projects; I dont recall this was one of them
We really need to start being better at following projects