search

geneontology / geneontology.github.io

Repository for storing GO documentation, directly available through the general GO site
http://geneontology.org
MIT License
6 stars 10 forks source link

Tighter integration and access of GO-CAMs in view of the article release #180

Open lpalbou opened 5 years ago

lpalbou commented 5 years ago

@thomaspd created a mockup to tighten the integration and ease the access of GO-CAMs from the main go site.

Proposed action items to be discussed and prioritized with @pgaudet and @suzialeksander :

I estimate this to be roughly a 2 days work.

The trickiest part is 4) as the data is not in the pipeline. If it can wait a month, I could modify our go-stats and more particularly the go-meta.json which is already fueling the other displayed stats. An alternate solution would be to do live queries over the rdf store but I would favor the update of the go-meta.json file over multiplying HTTP calls.

cmungall commented 5 years ago

Also:

cmungall commented 5 years ago

add a GO-CAM radio button below the main search bar.

I can see arguments for and against. I think if we do this, we do this more universally, not just from search bar.

Also it looks like this link isn't using the closure yet https://geneontology.cloud/browse/GO:0045444

I think it may be best to "silo" the sites a bit until they work in the same way, TBD

pgaudet commented 5 years ago

@lpalbou

Which GH site ? Why not keep the documentation on the GO website ?

3 - create a 4th card box for GO-CAM (next to annotations and before tools&guide). Link to the above documentation and to the browse models page

Maybe this is necessary, but it's too bad that GO-CAMs are not considered annotations (ie that they would be a separate box)

4 - add to the upper-right stats, data about GO-CAMs (# models, # genes, # species)

Given that we have not done exhaustive QC, perhaps those will change a lot, when we implement the ShEXs ? I would prefer to wait before showing stats, until the data is more stable.

5 - possibly add a top-level menu "GO-CAM" next to "Annotations" and before "Downloads"

Again, I am not clear why GO-CAMs are not annotations. I didn't this think was the message we wanted to send.

This is more Annotations2.0 but still annotations.

Thanks, Pascale

pgaudet commented 5 years ago

Also - did we review the text in the documentation ? https://www.geneontology.cloud/docs

It would be nice if it followed more the language of the GO-CAM specs https://docs.google.com/document/d/1OsE19zh8KE_2wT3-8oJysqEGqy-fahnhcjthO34IWP4/edit#

For example I thought we wanted to move away fro 'standard' or 'conventional' annotations. In the document we have 'Isolated (unconnected) GO-CAM statements'; perhaps we can call them 'unconnected annotations?

Thanks, Pascale

lpalbou commented 5 years ago

I've put Paul's mockup on the GO google drive: read only access link

use proxypass

@cmungall Yes, there is a ticket for that (also we are not clear on the why it currently doesn't work yet): https://github.com/geneontology/go-site/issues/965

I can see arguments for and against. I think if we do this, we do this more universally, not just from search bar.

Me too (e.g. we can search for GP or GO terms in either GO annotations or GO-CAMs... so it's a bit strange that these radio buttons are at the same level). Having said that, the GO-CAM site actually allows for a search over anything, so it would work.

Also linked to our discussion of having a sharable search widget with autocomplete I believe. Not achievable by October meeting, but can certainly be a target for May/June.

link isn't using the closure yet

True, there was a requirement for this (https://github.com/geneontology/web-gocam/issues/6), but it wasn't a priority.

Which GH site ? Why not keep the documentation on the GO website ?

@pgaudet GitHub site (pages to be more precise) = technical way to refer to the GO website. Yes the goal would be to move the GO-CAM docs from the GO-CAM site to the GO website

Maybe this is necessary, but it's too bad that GO-CAMs are not considered annotations (ie that they would be a separate box)

For the moment, GO annotations and GO-CAMs look quite different, including in the way users can interact with them. Maybe in the future ? If you have a proposal, we can certainly discuss it at a manager call.

Given that we have not done exhaustive QC, perhaps those will change a lot, when we implement the ShEXs ? I would prefer to wait before showing stats, until the data is more stable.

OK

Also - did we review the text in the documentation ?

Yes, there are some emails and google doc ~5 months ago

It would be nice if it followed more the language of the GO-CAM specs

Now that we have the GO-CAM specs, it's certainly a good idea ! I could look into it.

move away fro 'standard' or 'conventional' annotations

The GO-CAM article uses "standard annotations" to speak of GO annotations. There should be no mention of "conventional" anywhere, if there is, can you provide the links so we can correct ? As I understand, "standard annotations" is the accepted way to refer to GO annotations

Isolated (unconnected) GO-CAM statements

I think this was more referring to the CC only annotations ? I don't remember any mention of "isolated" or "unconnected" GO-CAM statements to refer to GO annotations.

lpalbou commented 5 years ago

I added this ticket to the next manager call, since there are some points to be discussed.

pgaudet commented 5 years ago

add a GO-CAM radio button below the main search bar. This will redirect for instance to https://geneontology.cloud/browse/GO:0005634 or https://geneontology.cloud/browse/sox2

Discussion at managers' call: We won't do that for now because the logic is different (does not implement closure)

pgaudet commented 5 years ago

We'll do # 2 and 3 for now (this is what I understood from today's managers call).

Pascale

lpalbou commented 5 years ago

I deployed the proposal on a temporary URL: http://www.geneontology.xyz/

I also ported the GO-CAM documentation as a GH page.

@pgaudet @cmungall @thomaspd if ok, this PR can be merged: https://github.com/geneontology/geneontology.github.io/pull/181

pgaudet commented 5 years ago

Hi @lpalbou

Overall it looks good - a few comments:

On http://www.geneontology.xyz/docs/gocam-overview/

  1. the link to 'evidence code' on the first line of the first paragraph should link to the GO website, not the wiki (/docs/guide-go-evidence-codes/). 1.1 You can also add a link to ECO http://www.evidenceontology.org/ (since you mention it later in the page).
  2. Under 'Linking different molecular activities' it states "The most common relations are directly (positively / negatively) regulates and directly provides input for, but there are other relations of greater and lesser specificity," ; is this true that these are the most common relations ? (do we have stats on these anywhere?)
  3. Change the formulation so that it doesn't sound like annotation guidlines: "Regulates should be used to denote biological control of a downstream activity. Provides input for should be used when there is no control, but an upstream function creates a molecular entity that is the target of the downstream function, such as in a metabolic pathway" -> Regulates denotes the control of a downstream activity. Provides input for indicates there is no control, but an upstream function creates a molecular entity that is the target of the downstream function, such as in a metabolic pathway.
  4. Under "Incomplete GO-CAMs" "We recognize that the knowledge of biology is incomplete; in cases where some or most of these aspects are unknown, a model may still be constructed with details added as more information becomes available. Users should attempt to specify functions as fully as possible, but partial models are expected and still contribute to the GO knowledgebase." -> Change to "The knowledge of biology is constantly evolving, therefore many models are incomplete, and details may be added as more information becomes available. ~Users should attempt to specify functions as fully as possible, but partial models are expected and still contribute to the GO knowledgebase.~"
  5. Last line: Do we want to give the link directly to the models ? https://github.com/geneontology/noctua-models/tree/master/models Is that how we expect people to download them ? Should we not link to the 'GO-CAM browse' page instead ?

Looks great otherwise !

Thanks, Pascale

lpalbou commented 5 years ago

@pgaudet I added your suggestions and after discussing with @thomaspd about the GO-CAM download link, we agreed to create a small Download GO-CAM page based on the Download annotation.

With the article now available, I just merged the PR: https://github.com/geneontology/geneontology.github.io/pull/181 . We can discuss later about other requirements / refactoring.

Notes