These are chat archives for openworm/ChannelWorm

21st
Aug 2015
Milad Jafary
@miladjafary
Aug 21 2015 18:29
Hi @travs
Travis Jacobs
@travs
Aug 21 2015 18:33
@miladjafary @VahidGh @slarson Hey guys
Vahid Ghayoomie
@VahidGh
Aug 21 2015 18:34
Hi everyone
Milad Jafary
@miladjafary
Aug 21 2015 18:35
Hi @VahidGh @slarson
what's up?
Stephen Larson
@slarson
Aug 21 2015 18:36
Hi all
Cheers!
Well it has come up to the end of @travs internship :sob:
I was hoping we could discuss the next steps
Travis Jacobs
@travs
Aug 21 2015 18:40
Yes, it is a sad day :weary:
Milad Jafary
@miladjafary
Aug 21 2015 18:40
@travs why?
Travis Jacobs
@travs
Aug 21 2015 18:40
I've got a small agenda here.
Agenda:
@miladjafary This is my last official day for GSoC!
Milad Jafary
@miladjafary
Aug 21 2015 18:41
@travs really?
Travis Jacobs
@travs
Aug 21 2015 18:43
@miladjafary Yeah :(
Stephen Larson
@slarson
Aug 21 2015 18:43
Great agenda @travs -- so how do we start?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 18:44
@travs, hoping to have a good collaboration (unofficially) :D
Travis Jacobs
@travs
Aug 21 2015 18:44
@VahidGh Absolutely! @slarson Thanks :)
Milad Jafary
@miladjafary
Aug 21 2015 18:44
@travs YES. I agree with @VahidGh
Travis Jacobs
@travs
Aug 21 2015 18:45
Ok so @VahidGh can you give us all a quick update on the status of #67, and optimization in general?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 18:46
As for #67, it's somehow related to #63
@miladjafary has started to work on these issues, since the last week
We decided to use the cwFitter as a util within the django app
Travis Jacobs
@travs
Aug 21 2015 18:50
@VahidGh Ok gotcha. I saw some outputs in some issues that looked really nice! So what do we need to be able to close #67 and #68?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 18:51
As it's a bit computationally expensive the optimization process (for a server like the OpenShift), we also decided to augment the process to be able to run as a background job
But generally, we can run the process locally for now
And postpone the considerations for running on the web server for later
Travis Jacobs
@travs
Aug 21 2015 18:56
@VahidGh Ok gotcha. So are the two issues I mentioned able to be closed now do you think? Since we can run the fitter fine locally?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 18:56
For #67, we need to complete TODOs I left in the Initiators class, and get parameters from DB/interface
But #68 is a bit different, I think it could be closed by closing #96
Travis Jacobs
@travs
Aug 21 2015 18:59
OK, I will update #67 to reflect these criteria
And what you say about #68 makes sense to me as well. As long as we can keep #96 moving forward, it should close naturally
@miladjafary @VahidGh So there is going to be another panel on the web app UI that says "Fitter" or something like that?
Milad Jafary
@miladjafary
Aug 21 2015 19:04
It could be
but
we are working on expriment dashboard for fitter
may be move this functionality to another page
Travis Jacobs
@travs
Aug 21 2015 19:06
@miladjafary Ok gotcha. So that is #63 I suppose. How far along is that issue?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:06
Yeah, as @miladjafary said, we will have this in both the unified experiment curation process, and as a stand alone process
Milad Jafary
@miladjafary
Aug 21 2015 19:07
@travs yes it's relate to UI integration
Travis Jacobs
@travs
Aug 21 2015 19:09
@VahidGh Ohh ok. And in terms of code, is there anything written for this one yet?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:11
I'm deeply involved in the modeling with no patch clamp data milestone these days
Milad Jafary
@miladjafary
Aug 21 2015 19:11
@travs yes I develop some part of it locally
but it's not finished yet.
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:12
And helping @miladjafary to be familiarized with the optimization and related issues
@miladjafary, I guess it's going to be done by the end of August!?
Travis Jacobs
@travs
Aug 21 2015 19:13
Ok! Well if there's anything you think could use a helping hand, just @-mention me (or whoever) in the issue of course :)
Stephen Larson
@slarson
Aug 21 2015 19:13
Excellent guys -- one of the reasons I know @travs is asking is because we want to make it clear to GSoC what he accomplished this summer and helping out with all these tasks has been a big part of it
Milad Jafary
@miladjafary
Aug 21 2015 19:14
@VahidGh I hope :ي
Stephen Larson
@slarson
Aug 21 2015 19:14
Some of these pieces he has been working on more than others
Milad Jafary
@miladjafary
Aug 21 2015 19:14
:D
Travis Jacobs
@travs
Aug 21 2015 19:14
Stephen Larson
@slarson
Aug 21 2015 19:14
nice transition :)
Milad Jafary
@miladjafary
Aug 21 2015 19:16
@travs thank you for collaboration :D
Travis Jacobs
@travs
Aug 21 2015 19:16
Ok, since @VahidGh mentioned the no-patch-clamp-data milestone, let's jump to point 3 in the agenda
@miladjafary And thank you as well! :)
It seems like this one has a lot of background information that needs to be assembled before much progress can be made on it. What do you think @VahidGh ?
Have you spoken with @a-palyanov recently about these? I think he was interested in this milestone
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:18
@travs, yeah, this one is a really different process
I've done some good works in the background
And trying to put all pieces together these days
Needed some background on proteins, structures, ion channel gating, and kinetics etc.
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:24
I hope to finish the general structure of the predictor within the next 2 weeks
Then I'll continue the process of the modeling of digitized channel plots
And will apply my predictor algorithm for the remaining ones (also with the help of the new models)
Travis Jacobs
@travs
Aug 21 2015 19:25
@VahidGh You're always doing good work Vahid! :)
Ok, this sounds really really cool
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:26
@travs, thanks :D
Travis Jacobs
@travs
Aug 21 2015 19:26
Is there any room for someone to jump in and help with this milestone? With data collection or with the predictor?
Stephen Larson
@slarson
Aug 21 2015 19:28
Is there any part of that process that collaborating with folks here would make better
?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:29
@travs, the doors are always open for this kind of works :D
@slarson, I think we can improve available data about ion channels we have
many of them are still not completed
The thing everyone needs as the starting point is the basic knowledge about ion channels, their types, structure, etc
After that there are lots of rooms open, and many interesting subjects to work on
Stephen Larson
@slarson
Aug 21 2015 19:33
Ok!
Travis Jacobs
@travs
Aug 21 2015 19:33
@VahidGh Ok, good point about the prerequisite knowledge.
Not sure what @a-palyanov's schedule is like right now, but he would be a good candidate and has expressed interest in this before.
Stephen Larson
@slarson
Aug 21 2015 19:35
Yeah if we want to make that more formal as a group we could
Travis Jacobs
@travs
Aug 21 2015 19:35
I will shoot him an email and see what he's up to these days, and get him in contact with you Vahid, cool?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:35
@travs, sure, I would really appreciate his help
Travis Jacobs
@travs
Aug 21 2015 19:35
Ok sweet :)
Stephen Larson
@slarson
Aug 21 2015 19:35
I know that you are also writing up a thesis right @VahidGh
Travis Jacobs
@travs
Aug 21 2015 19:36
Ohhh, I did not know that. Thank you so much for taking the time out for this meeting Vahid!
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:36
@slarson, yeah, I'm also busy with writing the thesis (I've done lots of it before, but going to be prepared for the defense)
Stephen Larson
@slarson
Aug 21 2015 19:38
Understood, yeah-- so we want to of course balance out with your available time
So just let us know
on those other pieces what priority you will put on them and how soon you want to address these
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:39
@slarson, yeah, thanks for opening the issue
Stephen Larson
@slarson
Aug 21 2015 19:39
Especially as we will shift away from @travs full time effort to a probably more moderate pace of activity on organising
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:41
Ok, as said before, in my side, I'm going to finish this work within the next 2 weeks
Then I'm ready for the modeling process of digitized ion channels
And then the remaining ones
And finally going to see how to re-move the worm with generated models :D
Stephen Larson
@slarson
Aug 21 2015 19:46
Ok great!
Travis Jacobs
@travs
Aug 21 2015 19:48
Ok excellent :)
I'll paste the agenda again since we seem to have a consensus about the status of this one
Stephen Larson
@slarson
Aug 21 2015 19:49
Cool!
Travis Jacobs
@travs
Aug 21 2015 19:49
Ok, for #44 there, and this ties in with what @miladjafary is doing in the UI regarding export
I think the last comment you made @VahidGh, you agree that we can store these series in pyopenworm? Wanted to hear your thoughts on this one
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:51
@travs, yeah, this is possible if considered in the adapter too
But something would be also very useful, for the purpose of both PyOpenWorm and ChannelWorm would be to store reported parameters in literature
This would be also very helpful for the validation process
Parameters like the reported area/capacitance for each cell
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:56
Reported values like V_half activation, E-rev, and tau_max for each channel
Travis Jacobs
@travs
Aug 21 2015 19:57
@VahidGh Ok, do you think we can store those in CW and export the same way we will for other types of data?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:57
These reported parameters are the most helpful data, which is not very hard to capture from literature and store somewhere
Travis Jacobs
@travs
Aug 21 2015 19:57
It seems like this type of literature and metadata can pass through CW as a curation interface
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:57
@travs, sure
Travis Jacobs
@travs
Aug 21 2015 19:58
Ok, when this is in place it should not be too hard to update the export to capture this information as well
As for the plots, I can fix this up in a short amount of time
Ok, so that brings us to the last point on the agenda
Vahid Ghayoomie
@VahidGh
Aug 21 2015 19:59
@travs, yeah, but we can also add them to a sheet for now, and add to the PyOW later
Travis Jacobs
@travs
Aug 21 2015 20:00
@VahidGh Ok, the parameters?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 20:00
This would help resolving the main issue with the validation process too
Yes
Travis Jacobs
@travs
Aug 21 2015 20:00
I'm not sure exactly what kind of structure this sheet would have
What would it look like, approximately?
Vahid Ghayoomie
@VahidGh
Aug 21 2015 20:01
As the main objective of a scientific validation test is to validate a model against observed and reported data from experiments
Something like the electrophysiology section here
In addition to those mentioned above for cells
Travis Jacobs
@travs
Aug 21 2015 20:05
Ok, that makes sense to me. Can you raise an issue for it maybe? Since I can't do it right now, and I don't want this idea to get lost in the shuffle of GSoC ending for me
Vahid Ghayoomie
@VahidGh
Aug 21 2015 20:07
I've opened some issues for this before
Should look for it
Travis Jacobs
@travs
Aug 21 2015 20:08
Ok sure, well as long as it doesn't get lost!
Alright guys, well that wraps it up for the agenda that I had planned out.
Vahid Ghayoomie
@VahidGh
Aug 21 2015 20:09
#138
Travis Jacobs
@travs
Aug 21 2015 20:09
@miladjafary @VahidGh @slarson
Did you guys have anything else you wanted to discuss before we break for now?
@VahidGh Ok excellent, I'm glad this is rendered in an issue. So this is a prerequisite to being able to store/export these params
Vahid Ghayoomie
@VahidGh
Aug 21 2015 20:12
I've written a model for storing key-value pairs in the django models, but could be also replaced with some static parameter names
But these values are some important values that would be very useful if maintained via PyOW
Travis Jacobs
@travs
Aug 21 2015 20:14
Yeah, it would make the data models in PyOW more rich as well. I agree we should add them there
Once this data is available and the parameters we want are defined, we should raise issues in PyOW to add these to the data models
OK guys, I've gotta run right now. It's been great catching up on some of these issues. Thank you for taking the time out for this meeting :)
Vahid Ghayoomie
@VahidGh
Aug 21 2015 20:18
OK, thanks everyone
Later
Milad Jafary
@miladjafary
Aug 21 2015 20:24
OK, thanks everyone