These are chat archives for openworm/ChannelWorm

1st
May 2015
Stephen Larson
@slarson
May 01 2015 12:44
@VahidGh let's just chat here
Stephen Larson
@slarson
May 01 2015 12:53
@VahidGh @travs you guys reading me?
Travis Jacobs
@travs
May 01 2015 12:53
I am yes
Vahid Ghayoomie
@VahidGh
May 01 2015 12:54
yes
Stephen Larson
@slarson
May 01 2015 12:54
OK great
Sorry for the frustration on hangout
So; as I was saying
I'm hoping to build energy around ChannelWorm and help to integrate it with other pieces over the summer
Vahid Ghayoomie
@VahidGh
May 01 2015 12:55
Actually I should appologize
Stephen Larson
@slarson
May 01 2015 12:55
No worries
I'm sure you aren't slowing your own bandwidth down @VahidGh :)
Not your fault, we understand :)
Travis Jacobs
@travs
May 01 2015 12:55
Yes totally!
Vahid Ghayoomie
@VahidGh
May 01 2015 12:55
Yes :D
Stephen Larson
@slarson
May 01 2015 12:55
As part of his project for the summer, @travs will be doing some coordinating of the community
So its helpful if we sync up and paint the larger roadmap that includes ChannelWorm / muscle_model & PyOpenWorm
Probably also the c302 effort
ChannelWorm seems to be coming along very nicely
you guys are doing great stuff
The first thing I was just trying to understand is how it is going
@travs was showing the Channel and Experiment data elements that have been added to the branch of PyOpenWorm
That's great!
I'm curious what else is going on recently with ChannelWorm
Vahid Ghayoomie
@VahidGh
May 01 2015 12:57
As I said I'm working on the estimation part
Stephen Larson
@slarson
May 01 2015 12:58
OK that's which issue?
Vahid Ghayoomie
@VahidGh
May 01 2015 12:58
Kinetic estimation of ion channels with no patch clamp data available
Stephen Larson
@slarson
May 01 2015 12:59
OK cool -- is there a github issue for that one yet?
Vahid Ghayoomie
@VahidGh
May 01 2015 12:59
Not yet
Travis Jacobs
@travs
May 01 2015 13:00
Maybe we can make that into one of our "milestones" Vahid, what do you think?
Similar to the one we have for modelling with patch-clamp data available
Vahid Ghayoomie
@VahidGh
May 01 2015 13:01
I'm still reading some papers for the feature selection step of the machine algorithm
Yes, I'll open an issue for this
Stephen Larson
@slarson
May 01 2015 13:01
What is your strategy for kinetic estimation of ion channels with no patch clamp data available ? Are we assuming at least an I/V curve is available?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:01
*machine algorithm
Yes, of course
Stephen Larson
@slarson
May 01 2015 13:02
OK good
Vahid Ghayoomie
@VahidGh
May 01 2015 13:03
My plan is to estimate conductance, V_half, and activation/inactivation parameters
Stephen Larson
@slarson
May 01 2015 13:04
Tell me more about the database that milad is working on?
And how it is going?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:05
Milad is not working on the DB right now
He just submitted his proposal this week
And I hope start to work on plan
The priority is with the pipeline for modeling ion channels with patch clamp data available
But the problem was with the fitting process
Stephen Larson
@slarson
May 01 2015 13:07
OK is he going to also have time this summer?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:07
It supposed to be done using Neurotune/Pyelectron
Yes, me and Milad will work on this for the next 4-5 monthes
Stephen Larson
@slarson
May 01 2015 13:08
OK great
syncing with github issues is probably going to be one of the most effective ways we can keep all this into perspective
Vahid Ghayoomie
@VahidGh
May 01 2015 13:09
Yes
Travis Jacobs
@travs
May 01 2015 13:09
I agree
Vahid Ghayoomie
@VahidGh
May 01 2015 13:10
till this week I was not sure about timings of contributers
But now that's really better
Stephen Larson
@slarson
May 01 2015 13:10
Yes
And we should bring back and merge others
Vahid Ghayoomie
@VahidGh
May 01 2015 13:10
And of course we will try to work on a better plan
Travis Jacobs
@travs
May 01 2015 13:10
Totally
Stephen Larson
@slarson
May 01 2015 13:10
so the question is how can we open things up for other contributions
this is often tricky :)
but it is doable
Vahid Ghayoomie
@VahidGh
May 01 2015 13:11
Yep :D
Travis Jacobs
@travs
May 01 2015 13:11
Joe started making a django project for this repo if I'm not mistaken
Stephen Larson
@slarson
May 01 2015 13:11
Because the next meeting we'd like to have is to get @travs coordinating some of the other contributors and merge the efforts with muscle_model and pyopenworm
Travis Jacobs
@travs
May 01 2015 13:13
Yes, I think we'll need to have a basic outline of where we want things to go by then
@VahidGh There was an idea for a sort of web app for this project, is this correct?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:14
Yes
Travis Jacobs
@travs
May 01 2015 13:14
Can you describe what the web app would do, just to remind me?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:15
At the first step we need to structure and represent data related to ion channels in C. elegans
Also developing the pipeline #16
Stephen Larson
@slarson
May 01 2015 13:17
OK
Travis Jacobs
@travs
May 01 2015 13:17
Right, this makes sense
I've started a joint diagram that you guys should be able to edit
Vahid Ghayoomie
@VahidGh
May 01 2015 13:17
For letting users to build models from patch clamp studies
Travis Jacobs
@travs
May 01 2015 13:18
Excellent
Stephen Larson
@slarson
May 01 2015 13:18
@VahidGh you've done a better job than I have of making milestones!
Vahid Ghayoomie
@VahidGh
May 01 2015 13:18
#17 would be the first step
Stephen Larson
@slarson
May 01 2015 13:18
So now we have to step up our game to meet you :)
So one thing is that the data modelling piece in ChannelWorm is really all PyOpenWorm, right?
Travis Jacobs
@travs
May 01 2015 13:20
That's correct
Vahid Ghayoomie
@VahidGh
May 01 2015 13:20
Yes, we tried to work based on the PyOW approach
Stephen Larson
@slarson
May 01 2015 13:20
Have we run into roadblocks ?
I would imagine you did...
Vahid Ghayoomie
@VahidGh
May 01 2015 13:21
But also we will have another DB for the web app purposes
Stephen Larson
@slarson
May 01 2015 13:22
ok
Travis Jacobs
@travs
May 01 2015 13:22
One issue that Vahid and I have discussed is using PyOpenWorm, but wanting to have code that we can generalize back into PyOW from ChannelWorm further downstream
Basically, how can we model our data in PyOpenWorm without making it too specific, and not really using PyOpenWorm anymore, but some heavily-modified fork
Stephen Larson
@slarson
May 01 2015 13:23
OK -- ideally PyOpenWorm is usable as a data backend for the web app
If we can figure out the specific reasons it is not, that is useful feedback
but we can also merge the efforts back in later as well if it is quicker to just get started with something simple
Vahid Ghayoomie
@VahidGh
May 01 2015 13:24
I was speaking about user accounts, profiles, etc.
Stephen Larson
@slarson
May 01 2015 13:24
Right right
I see
so the plan is to have that other DB not capture the scientific information?
Travis Jacobs
@travs
May 01 2015 13:24
That's right
Stephen Larson
@slarson
May 01 2015 13:24
Maybe do we need a web app milestone as well?
Travis Jacobs
@travs
May 01 2015 13:25
I think so. I can make that now
Vahid Ghayoomie
@VahidGh
May 01 2015 13:26
Yes, That's what @miladjafary and if @joebowen still is interested in, would work on
Stephen Larson
@slarson
May 01 2015 13:26
Great -- let's get that into issues & make a milestone so we can put that on the roadmap too
OK -- then I think the muscle_model stuff becomes one specific example of what ChannelWorm is intended to help create
Vahid Ghayoomie
@VahidGh
May 01 2015 13:27
Another critical work to be done, is the fitting process with data from patch clamp studies
We had a discussion about this with @pgleeson
We need some improvements to thee Neurotue/Pyelectron
Stephen Larson
@slarson
May 01 2015 13:29
OK -- So let's add this to our roadmap
what is the repo for that? I always get confused which one it is
@travs how much of the muscle_model stuff have you been able to run / play with yet?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:30
But I don't know what do we need exactly for this customization
This is the updated version
Stephen Larson
@slarson
May 01 2015 13:31
OK -- I see no issues
hehe
Travis Jacobs
@travs
May 01 2015 13:31
@slarson I was able to generate the tuning curves if I'm not mistaken, but this was a few months ago. As I rewatch the hangouts I'll be going through the muscle_model repo in full as well
It looks like making some milestones in the muscle_model repo would also be a good idea, so I'll take a look at that
There are already some issues, so I can organize these into milestones
Stephen Larson
@slarson
May 01 2015 13:34
Ok so @VahidGh I see that for #16 the next step we are stuck on is this Neurotune step, right?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:36
Yes
If @pgleeson could help on specifying required changes, we can start working on this too
Stephen Larson
@slarson
May 01 2015 13:38
Yep exactly
OK I'm asking him to structure some of that
Vahid Ghayoomie
@VahidGh
May 01 2015 13:40
Great
Stephen Larson
@slarson
May 01 2015 13:40
Hmm -- it may be good for us to structure a unified waffle board for this stuff
Vahid Ghayoomie
@VahidGh
May 01 2015 13:42
I guess I did it some time ago!!
Stephen Larson
@slarson
May 01 2015 13:42
Yes?
link?
Otherwise I could aggregate it into the muscle model one
Stephen Larson
@slarson
May 01 2015 13:45
OK so we can use yours
Can you integrate the other repos tho?
Probably you don't have access to neurotune b/c its Padraig's personal one
Vahid Ghayoomie
@VahidGh
May 01 2015 13:46
yes
We need access for other repos
Stephen Larson
@slarson
May 01 2015 13:46
how about openworm/muscle_model
you should have access there
and pyopenworm actually
you know I mean the settings for that one, right?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:49
You mean merging all the Waffles in one?
Stephen Larson
@slarson
May 01 2015 13:49
Screenshot 2015-05-01 06.48.36.png
Yeah like here's how we have it for muscle_model
Screenshot 2015-05-01 06.49.35.png
the gear here
and then additional sources
or if you want to keep that one separate for now
then I can merge ChannelWorm into muscle_model
up to you
But yes, it is nice then because you can see the collection of issues in relation to others across repos
Vahid Ghayoomie
@VahidGh
May 01 2015 13:51
I could connect PyOW
Stephen Larson
@slarson
May 01 2015 13:52
ok -- muscle_model ?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:52
yes connected
Stephen Larson
@slarson
May 01 2015 13:52
can you make me a collaborator on ChannelWorm repo too?
Vahid Ghayoomie
@VahidGh
May 01 2015 13:52
Oh yes :D
Stephen Larson
@slarson
May 01 2015 13:52
thanks :)
Did we get a milestone for the web app @travs ?
Travis Jacobs
@travs
May 01 2015 13:54
Yeah!
Stephen Larson
@slarson
May 01 2015 13:54
I mean I see it on the diagram
but also in github?
Yes i see it
ok great
Travis Jacobs
@travs
May 01 2015 13:54
Yea, it only has one issue yet
Stephen Larson
@slarson
May 01 2015 13:54
gotcha
Travis Jacobs
@travs
May 01 2015 13:54
ok
Stephen Larson
@slarson
May 01 2015 13:55
Is there any issue with a roadmap for next steps with the web app?
Or a pointer to the code?
Travis Jacobs
@travs
May 01 2015 13:55
I think so, one sec
It was mentioned here
but the code is not on github yet.
I've reached out to Joe so getting some input from him will be informative
Stephen Larson
@slarson
May 01 2015 13:57
OK yeah I just commented on the issue
OK so Padraig will illuminate us on the neurotune thing which is a bottleneck for the pipeline
Travis Jacobs
@travs
May 01 2015 13:58
Right
Stephen Larson
@slarson
May 01 2015 14:00
in the mean time...
So I see the new PyOpenWorm bits for Channel and Experiment
Let's see if the status is up to date on the cards
also @VahidGh were you able to merge the muscle_model into the waffle board? not sure I see those issues in there
Vahid Ghayoomie
@VahidGh
May 01 2015 14:03
Yes I connected them
Should I do anything else?
Stephen Larson
@slarson
May 01 2015 14:03
Oh I see they are mostly all in backlog
Travis Jacobs
@travs
May 01 2015 14:03
In the backlog
yes
Stephen Larson
@slarson
May 01 2015 14:04
ok so one thing we should do is to review the state of these cards across the projects
so one thing Vahid is you should move cards into the "ready" column that are the next steps that need to get done, and move cards into "in progress" that are getting worked on now
and we should collectively brainstorm on appropriate milestones for muscle_model that compliments what is going on in channelworm
Vahid Ghayoomie
@VahidGh
May 01 2015 14:05
OK
Travis Jacobs
@travs
May 01 2015 14:05
Ok that makes sense
Stephen Larson
@slarson
May 01 2015 14:06
Ok one obvious thing about muscle_model that is different from ChannelWorm is aggregating channels into the cell level
unless @VahidGh was this covered in your plan?
It isn't an explicit milestone to do this work but maybe you have put it in your roadmap somewhere I don't see?
Vahid Ghayoomie
@VahidGh
May 01 2015 14:07
Not yet
Stephen Larson
@slarson
May 01 2015 14:08
OK
So I'm thinking about what is the meaningful milestone for that aggregation
Vahid Ghayoomie
@VahidGh
May 01 2015 14:09
Yes, also merging with C302 project
We can do this in the validation phase
Stephen Larson
@slarson
May 01 2015 14:11
Ah yes
So that's the other one we should probably add to our roadmap
Vahid Ghayoomie
@VahidGh
May 01 2015 14:11
Every ion channel model created, should pass the overall validation phase in the OW
Stephen Larson
@slarson
May 01 2015 14:11
phew this is getting intense haha
Travis Jacobs
@travs
May 01 2015 14:11
Ok, so validation is going to be in CW or Muscle_model?
Haha yes it is, awesome :smile:
Stephen Larson
@slarson
May 01 2015 14:12
My thought would be that there would be CW validation and muscle_model validation both
Vahid Ghayoomie
@VahidGh
May 01 2015 14:13
Yes
Stephen Larson
@slarson
May 01 2015 14:13
ok Vahid can you also add the https://github.com/openworm/CElegansNeuroML repo to the waffle board?
Travis Jacobs
@travs
May 01 2015 14:13
Ah yes of course, at channel and cell levels respectively
ok
Stephen Larson
@slarson
May 01 2015 14:13
right
openworm/muscle_model#31 is the closest to capturing the idea of an "integration test" for validation
Vahid Ghayoomie
@VahidGh
May 01 2015 14:14
Done
Stephen Larson
@slarson
May 01 2015 14:14
thanks
so i'm trying to envision how to really structure this
Some of this gets down to the vision for the testing and documentation framework we want to have here
I'm sketching at the bottom of the roadmap to see if there is a diagram that hooks this together
Travis Jacobs
@travs
May 01 2015 14:24
Looks good!
Stephen Larson
@slarson
May 01 2015 14:31
OK -- that's getting closer. @VahidGh and @travs what do you think? Edits?
Travis Jacobs
@travs
May 01 2015 14:32
That looks good to me!
I'm not sure if the web app would fit into the diagram, since it's more of a means to an end
Stephen Larson
@slarson
May 01 2015 14:33
Ah right
that bit
Vahid Ghayoomie
@VahidGh
May 01 2015 14:33
That's cool
Travis Jacobs
@travs
May 01 2015 14:34
Yeah I think that's awesome
Stephen Larson
@slarson
May 01 2015 14:35
So the question is what is the functionality of the web app we want to expose from this system diagram?
Is it channel worm only?
is it the bigger picture?
Travis Jacobs
@travs
May 01 2015 14:35
And that is a very good question
What was your intention for the web app @VahidGh ?
Stephen Larson
@slarson
May 01 2015 14:37
ok back to the waffle board for a minute -- @VahidGh and @travs -- which are the issues in channel worm that should be in the ready and in progress columns? we should just drag them over from backlog
What's the difference between #21 and #22 ?
Travis Jacobs
@travs
May 01 2015 14:42
My mistake :flushed:
I made it twice
Stephen Larson
@slarson
May 01 2015 14:42
no worries
#11 is in progress, right?
Travis Jacobs
@travs
May 01 2015 14:44
Yess, just moved it
@VahidGh is #4 complete?
Vahid Ghayoomie
@VahidGh
May 01 2015 14:47
Yes
Stephen Larson
@slarson
May 01 2015 14:48
OK guys -- we've been doing this for a few hours now; I have to break
We're not done :) but we've made good progress
So I propose we get together again very soon to get the next bit. Priorities are to flesh out the milestones for the other projects that are needed to actually build the roadmap picture
we'll want to add tracking issues
Vahid Ghayoomie
@VahidGh
May 01 2015 14:50
Yes, that was a good progress
Stephen Larson
@slarson
May 01 2015 14:50
and following the beautiful effort Vahid has put into the channelworm docs, I am similarly inspired for us to write this plan up in docs form so that it integrates nicely with his
Travis Jacobs
@travs
May 01 2015 14:50
Yeah totally! :smile: I can make another meeting time (this time with more options, sorry again Stephen)
Stephen Larson
@slarson
May 01 2015 14:50
this way when we start to organize others we will have a comprehensive picture to present to everyone in writing
Great -- yes please travis -- send out another poll -- just check PT time zone when you pick times :D
let's have options that are over the weekend and early next week
Travis Jacobs
@travs
May 01 2015 14:51
Will do!
Vahid Ghayoomie
@VahidGh
May 01 2015 14:52
:D Actually I was really surprised you waked up at that time
you rock Stephen
Stephen Larson
@slarson
May 01 2015 14:52
Hehe yeah -- I had been waking up even 2 hours earlier than that all this week for some other meetings with the GMT time zone, so it wasn't that bad
Also it gave lots of time to dig into this today since now things are just starting over here
so all good :)
You guys rock
We're going to be really pleased with what we can accomplish over the next few months, I feel it :)
Bye for now
Travis Jacobs
@travs
May 01 2015 14:53
Me too.. This is really going to take off :rocket:
Seeya later Stephen!
Vahid Ghayoomie
@VahidGh
May 01 2015 15:01
@slarson, One thing I forgot to say, is the problem with limitation in the number of group members in Mendeley.
I remember Travis had some troubles with accessing papers in Mendeley.
Travis Jacobs
@travs
May 01 2015 15:02
@VahidGh That's all resolved now :smile: I'm in the group
Vahid Ghayoomie
@VahidGh
May 01 2015 15:02
Really? how?
Travis Jacobs
@travs
May 01 2015 15:02
Stephen invited me
Or maybe this is a different group?
Just checked; I am in the "OpenWorm" mendeley group
@joebowen
I will have a look at setting up a mirror now, since I have access to the rhcloud git repo. That could be just what we need
Vahid Ghayoomie
@VahidGh
May 01 2015 15:05
OK, I remember there was a limitation before, I hope it is solved now!
Stephen Larson
@slarson
May 01 2015 15:10
It is still limited unfortunately which is why i can only add folks on an invite basis-- but i freed up a slot for Travis and gave it to him
Vahid Ghayoomie
@VahidGh
May 01 2015 15:10
I see
Stephen Larson
@slarson
May 01 2015 15:11
If others need access, let me know and i'll see if i can free up more slots
Vahid Ghayoomie
@VahidGh
May 01 2015 15:12
OK, thanks
Joe Bowen
@joebowen
May 01 2015 16:21
@travs ok, let me know if you need anything more from me and we'll chat more tomorrow.