search

geneontology / pathways2GO

Code for converting between BioPAX pathways and Gene Ontology Causal Activity Models (GO-CAM)
8 stars 0 forks source link

Document transformation rules for Cambridge GO meeting #57

Closed goodb closed 5 years ago

goodb commented 5 years ago

Get something - e.g. a slide deck ready such that interested parties can see the current rules for transforming reactome into GO-CAMs.

goodb commented 5 years ago

@ukemi unless you have something else in mind, I'm going to use the existing Google Slidedeck to capture the rules and relevant screenshots. Let me know if this is okay with you or if you have another preferred approach. (Can be exported as .ppt if you prefer)

ukemi commented 5 years ago

Hi @goodb. I wasn't planning to bring my laptop with me, an experiment. Instead I'd like to have all of the slides in the meeting folder on the Google drive. Basically @deustp01 will give an introduction to Reactome curation and represenation and then I will go through the rules with examples. Ideally we will have slides that represent: rules for extracting molecular functions, rules for extracting biological processes, rules for extracting cellular components and rules for making causal connections between functions. The ideal slide would have a brief description of the rule, a screen shot of how the information looks in Reactome and a screen shot of how the information looks in GO-CAM.

goodb commented 5 years ago

@ukemi okay then it sounds like Google Slideshow is a good format for this. Will continue on there today.

goodb commented 5 years ago

Note from today's meeting. Beyond capturing rules of transformation, this project is generating general purpose templates for how to model different aspects of biology within the GO-CAM framework. For example, how to model small molecular regulation of a process. We should attempt to capture these patterns or 'templates' in a way that makes it easy for curators to see them, use them, and contribute to them. As a first step, we can put them on the wiki. Ideally we would see deeper integration with the go-cam editing interface. Curators could choose templates to use to build skeletons for their models as well as comment on or even directly extend the templates as needed.

ukemi commented 5 years ago

Thanks @goodb. I am going to download it and do a bit of culling so the talk won't be too long, but these look good. I will add another version to the meeting folder.

ukemi commented 5 years ago

OK @goodb. I have removed a bunch of the specific rule slides because I think it will take too much time to go through every one of them. I have uploaded a new deck to the meeting Google folder. I had one question about slide 24. The figure at the right indicates that the reaction is happening in two places. Don't we want to show an example where this resolves nicely to one location?

deustp01 commented 5 years ago

@ukemi @goodb Looks good to me. Small things - I replaced the screenshot of a bit of the BMP pathway on slide 1 with one that shows the events as corrected by our discussions last month, and added comments on slides 10 and 15.

ukemi commented 5 years ago

Awesome. @deustp01 you can update me on last week's discussion when we are in Cambridge.

goodb commented 5 years ago

@ukemi my apologies. Those slides aren't quite finished yet. It looks like you have merged together the new ones I was working on in the beginning of the doc with the old ones that were being replaced at the end of the doc and may not be up to date. And now we have three versions..

I'm going to go through the one that Peter was most recently working on and make sure what is there is correct with what is currently working. https://docs.google.com/presentation/d/1m0Xxyoalrtt2XFgkZLUPKOlrUstG5Lax7fvpLvOHJvM/edit#slide=id.p1

Sorry, I should have been more on top of this. The review for the new slides ended up identifying a couple bugs that distracted me.. (noctua-dev has the latest translation with those bugs cured now.)

Let me know if you want to hop on skype and discuss before the meeting.

goodb commented 5 years ago

OK @goodb. I have removed a bunch of the specific rule slides because I think it will take too much time to go through every one of them. I have uploaded a new deck to the meeting Google folder. I had one question about slide 24. The figure at the right indicates that the reaction is happening in two places. Don't we want to show an example where this resolves nicely to one location?

If we are talking about the same slide 24, I believe that the slide is out of date and has already been replaced by Slide 15 - 'Inferring Occurs_in'

goodb commented 5 years ago

@ukemi and @deustp01 hold on edits/comments on slides please. Let me get through updating the Google Slides version in the meeting drive that you generated so its in sync.

deustp01 commented 5 years ago

Meanwhile, I have posted a draft of the "Reactome" slide set to the "presentations" folder on the Google drive https://drive.google.com/drive/folders/1uTDBjMGYJUIQ9OOWp5P28rFem6WwPJEQ Comments?

goodb commented 5 years ago

Okay @ukemi and @deustp01 my panic is over now. I deleted old/redundant slides and added a couple of the ones you deleted (not all) back into the slides. (Those currently have yellow backgrounds so you can easily find and re-delete them if you wish. Now that I am confident what you have matches what I have actually done, please feel free to delete/edit what you like to make your presentation. (I'd suggest keeping the slides you deleted as back ups in case they come up - e.g. if people browse the models and ask you what the 'union' thing is or why the complexes are so long. But of course up to you.) Sorry again for not having this finished last week.

goodb commented 5 years ago

@deustp01 Depending on your view of the audience's interests, I think it might be worthwhile to mention the BioPAX export when you bring up interoperability. The fact that you do make use of the BioPAX standard (however problematic it can be) has made it possible for me to write code that not only works on Reactome, but also on most other pathway databases. Just like the work on identifier alignment, it really does increase interoperability systemically, resulting in a lot of saved work.

It may be a bit down the road, but I also think it might be worth talking about the potential for using the GO-CAM conversion as the main way that you generate the GPAD files shared with and then by the GO consortium. That would touch on other conversations the software team is having about the pipeline and various places where rules for validating annotations/models should be put into place.

deustp01 commented 5 years ago

David will be talking immediately after me about the GO-CAM project, so I expect the GO-CAM to GPAD, and validation material will make more sense to the audience as part of his talk. I will add a slide to mine to cover exports of data from Reactome and highlight BioPax there as a link to David and GO-CAM.

ukemi commented 5 years ago

Hi @goodb . So just to be sure, the slides I am supposed to use are the ones on the meeting drive? I pared them down because I don't think it is necessary to give curators all of the rules we have generated, just give them examples. I was definitely going to mention the GPAD files at the end as future work. Most of the people at the meeting are gene annotators, not ontology developers or software engineers.

goodb commented 5 years ago

@ukemi Please use the current Google Slides (not the ppt file) that are in the Google Drive for the meeting as your starting point for today. I'd suggest just editing them directly and presenting from the web unless that is too painful for you.

ukemi commented 5 years ago

I'm actually not bringing my laptop to the UK (an experiment), so presenting from the web is the best way for me to go.

goodb commented 5 years ago

@ukemi great. That way any other changes we want to make can be sure to be in sync.

can we delete the .ppt file from the drive to avoid potential confusion ?

ukemi commented 5 years ago

Sure.

goodb commented 5 years ago

And yes.. unless you have 90 minutes it will probably take some more pruning. My objective was to give a complete coverage of the rules as they exist now to give you a starting point and a reference.

The only slides I added back currently have yellow backgrounds so you can find them easily.

ukemi commented 5 years ago

Do you have time for a quick skype?

goodb commented 5 years ago

yes. hopping on now

deustp01 commented 5 years ago

Added BioPax slide to Reactome_overview_draft; replaced earlier pptx and GoogleSlides version of file with this new one.

goodb commented 5 years ago

Closing issue with outputs:

  1. Slides capturing all of the current operational rules for transforming reactome to go-cam. This one is meant to live on and capture future changes as they come in.
  2. Slides that @ukemi will present to the group (shorter, subset of rules). This one will just be used for this meeting.
  3. Slides that @deustp01 will present regarding the state of Reactome and its integration with GO and RHEA.