These are chat archives for openworm/ChannelWorm

21st
Jun 2015
Wilson Zhao
@wilzh40
Jun 21 2015 03:07
Okay, I setup a virtualenv for python and installed Channelworm...wow. That took about 20 minutes.
Wilson Zhao
@wilzh40
Jun 21 2015 03:31
im getting a no module named wsgi when i try running python manage.py runserver however
Chee Wai Lee
@cheelee
Jun 21 2015 04:42
I'm getting the same issue with wsgi. Tried PYTHONPATH=<path>/Channelworm/channelworm/wsgi but that did not fix the problem. I'll mess around a little more.
Currently reading through http://webpython.codepoint.net/wsgi_tutorial. Turns out I was wrong thinking about wsgi as a typical python module.
Milad Jafary
@miladjafary
Jun 21 2015 04:55
@cheelee @wilzh40 Hi. I think we have issue about wsgi after updating project in order to upload in openshift.
I look at it for finding problem. :d
Chee Wai Lee
@cheelee
Jun 21 2015 04:59
@miladjafary cool thanks! I'll keep an eye out on waffle!
Vahid Ghayoomie
@VahidGh
Jun 21 2015 09:36
Hey @cheelee, Hey @wilzh40. As Milad said this is due to some new changes for deploying on the OpenShift.
For now, you can just comment out this line to run it locally.
One in progress issue to tackle these kinds of conflicts during updates, is #89.
Chee Wai Lee
@cheelee
Jun 21 2015 12:59
ok sounds good.
Chee Wai Lee
@cheelee
Jun 21 2015 13:05
@VahidGh That did the trick. Navigated to 127.0.0.1:8000 and was "logged in" as a well-drawn cartoon portrait of @slarson (at least I think that's him!) :)
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:15
:D Yeah, they look the same
@joebowen, Working on the migration issue...
Here I made some changes to the cronjob script
Joe Bowen
@joebowen
Jun 21 2015 13:16
Hey. I saw that you added those makemigrations and migrate. Did that work>
?
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:18
It works unless there need a confirmation from manage.py when renaming an object in models
Asks for confirmation
I read somewhere --noinput resolves the problem
Joe Bowen
@joebowen
Jun 21 2015 13:19
yea, I was just reading that
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:19
But did not work, neither locally nor on the server
Joe Bowen
@joebowen
Jun 21 2015 13:20
do you have access to see the cron logs?
hmmm
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:20
So I decided to use the hacky way echo "y" |
Joe Bowen
@joebowen
Jun 21 2015 13:21
I'm still seeing the EOF in the logs
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:21
No, I just wanted to open an issue for writing some logger for the cronjob :)
How can I have access to logs?
Joe Bowen
@joebowen
Jun 21 2015 13:23
we could do a minutely cron copy of the serverl logs to a remote somewhere... or I could just give you ssh permissions
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:25
Would you please post here the log?
Joe Bowen
@joebowen
Jun 21 2015 13:26

Sun Jun 21 09:25:12 EDT 2015: START minutely cron run


/var/lib/openshift/55454af95973ca347e00011b/app-root/runtime/repo//.openshift/cron/minutely/updated_from_git.sh:
Cloning into 'ChannelWorm'...
You are trying to add a non-nullable field 'deltat' to patchclamp without a default; we can't do that (the database needs something to populate existing rows).
Please select a fix:
1) Provide a one-off default now (will be set on all existing rows)
2) Quit, and let me add a default in models.py
Select an option: Please select a valid option: Traceback (most recent call last):
File "manage.py", line 10, in <module>
execute_from_command_line(sys.argv)
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/core/management/init.py", line 338, in execute_from_command_line
utility.execute()
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/core/management/init.py", line 330, in execute
self.fetch_command(subcommand).run_from_argv(self.argv)
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/core/management/base.py";, line 390, in run_from_argv
self.execute(args, **cmd_options)
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/core/management/base.py";, line 441, in execute
output = self.handle(
args, **options)
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/core/management/commands/makemigrations.py";, line 125, in handle
migration_name=self.migration_name,
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/db/migrations/autodetector.py";, line 43, in changes
changes = self._detect_changes(convert_apps, graph)
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/db/migrations/autodetector.py";, line 185, in _detect_changes
self.generate_added_fields()
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/db/migrations/autodetector.py";, line 763, in generate_added_fields
self._generate_added_field(app_label, model_name, field_name)
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/db/migrations/autodetector.py";, line 793, in _generate_added_field
field.default = self.questioner.ask_not_null_addition(field_name, model_name)
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/db/migrations/questioner.py";, line 131, in ask_not_null_addition
"Quit, and let me add a default in models.py",
File "/var/lib/openshift/55454af95973ca347e00011b/python/virtenv/lib/python2.7/site-packages/Django-1.8-py2.7.egg/django/db/migrations/questioner.py";, line 99, in _choice_input
result = input("Please select a valid option: ")
EOFError: EOF when reading a line
Operations to perform:
Synchronize unmigrated apps: staticfiles, messages
Apply all migrations: admin, contenttypes, ion_channel, auth, sessions
Synchronizing apps without migrations:
Creating tables...
Running deferred SQL...
Installing custom SQL...
Running migrations:
Rendering model states... DONE
Applying ion_channel.0005_auto_20150621_1014... OK


Sun Jun 21 09:25:20 EDT 2015: END minutely cron run - status=0

Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:27
The `echo "y" |' trick worked for me locally without this error!
Joe Bowen
@joebowen
Jun 21 2015 13:28
what about that "temporary workaround" shown in that django ticket?
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:32
Didn't try, but another odd thing is that, now I've updated the db in repo with the new scheme and commented out the line removing db in the cronjob but still the db file is the old one!
Joe Bowen
@joebowen
Jun 21 2015 13:35
how long ago did you comment out that line?
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:36
Recently, but the date om my VM was not update, and that's why the git showing the wrong time :)
Synced the time now..
OK, the online app is back now
Commented the migrate command for now
Joe Bowen
@joebowen
Jun 21 2015 13:38
ok, I see that in the logs now
Joe Bowen
@joebowen
Jun 21 2015 13:44
@VahidGh do you have an openshift account?
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:45
Nope
Joe Bowen
@joebowen
Jun 21 2015 13:46
If you create one, I can give you access to the admin of the servers.
I just need your username to add to the servers
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:51
Joe Bowen
@joebowen
Jun 21 2015 13:52
ok, you've been added to the domain.
Vahid Ghayoomie
@VahidGh
Jun 21 2015 13:53
Right, thanks
Joe Bowen
@joebowen
Jun 21 2015 13:54
then if you add your ssh public key, you can then access the server through ssh. The logs are in app-root/logs/*
Vahid Ghayoomie
@VahidGh
Jun 21 2015 14:46
Have access to logs now, thanks
Travis Jacobs
@travs
Jun 21 2015 16:24
Hey everyone @/all :)
Joe Bowen
@joebowen
Jun 21 2015 16:24
Hey @travs
Chee Wai Lee
@cheelee
Jun 21 2015 16:25
Hey all!
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:25
Hey, I'm Shreejoy, interested in sucking up channel data from literature
Milad Jafary
@miladjafary
Jun 21 2015 16:26
Hi ever one
every one*
Travis Jacobs
@travs
Jun 21 2015 16:26
:)
Richard C Gerkin
@rgerkin
Jun 21 2015 16:26
Rick here, SciUnit guy
Chee Wai Lee
@cheelee
Jun 21 2015 16:26
I'm trying to figure out what I can do to help out with the code and integration.
and the workflow
Richard C Gerkin
@rgerkin
Jun 21 2015 16:27
Hey all, I am going to step out for a few minutes for breakfast with family. What I'd like to do when I get back is dive into the Channel Analysis test code.
I submitted this pull request a few days ago: NeuroML/pyNeuroML#8, to get pyNeuroML/NML2ChannelAnalysis in shape to integrate with SciUnit
Vahid Ghayoomie
@VahidGh
Jun 21 2015 16:27
Hi everyone again :)
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:28
the channelworm webapp looked really cool - Id love to see/hear more about it
Stephen Larson
@slarson
Jun 21 2015 16:28
Did you get to the live link @stripathy ?
Richard C Gerkin
@rgerkin
Jun 21 2015 16:28
Then I will be able to work it in to this: https://github.com/VahidGh/ChannelWorm/blob/master/tests/IVCurve_Model.ipynb, so that the first half (example trivial SciUnit tests) can be updated to actually do the tests being shown in the second half.
So brb in about 30 minutes
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:28
yes - i'm playing with the app now
Stephen Larson
@slarson
Jun 21 2015 16:29
cool
Joe Bowen
@joebowen
Jun 21 2015 16:29
Hey @stripathy, I was gonna be doing some work on the web app, if you want to help.
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:29
definitely - is there any prioritization on the papers / figures that need extracting?
Joe Bowen
@joebowen
Jun 21 2015 16:30
Are we gonna go forward with the modeled grid dashboard or do we want to put that on hold and work on other things first?
Stephen Larson
@slarson
Jun 21 2015 16:30
So one thing guys as I’ve been mentioning the effort on milestones, it may be worth talking at the level of milestones and what issues are under what milestones
this may be less familiar to us, but its an experiment to try this
so for example
Vahid Ghayoomie
@VahidGh
Jun 21 2015 16:31
@stripathy you can take a look at #8
this card is wrapping up the web interface tasks
and you can click into the filtered waffle board from there
for @stripathy this is the milestone on gathering data: https://trello.com/c/rkV2k51F/12-channelworm-collect-and-digitize-channel-data
Chee Wai Lee
@cheelee
Jun 21 2015 16:32
@slarson am looking through the high level milestones to get an idea now.
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:32
#8 yeah that sounds great, let's work on that
Joe Bowen
@joebowen
Jun 21 2015 16:33
I'll grab some low hanging fruit and see if I can see what's causing this, VahidGh/ChannelWorm#94
Travis Jacobs
@travs
Jun 21 2015 16:34
@stripathy I'm not sure if the issues link it, but we have a folder of ion channel papers thatare relevant to C elegans
grabbing the link for you now
@VahidGh Stephen and I have been going through the papers and getting parameters/plots that will inform our ion channel modeling
@stripathy here is the folder of ion channel papers Vahid compiled
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:36
thanks!
Travis Jacobs
@travs
Jun 21 2015 16:37
and here is a chart we have been updating, which lists the relevant figures we will be grabbing and running through the channelworm pipeline/digitizer
@stripathy no sweat!
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:37
how can I help with this?
Chee Wai Lee
@cheelee
Jun 21 2015 16:38
https://trello.com/c/SL6y8dHH looks pretty interesting from my perspective - is it waiting on any dependencies? I cannot tell.
Stephen Larson
@slarson
Jun 21 2015 16:38
@stripathy maybe you should say more about what’s possible with channel data from literature relative to what you’ve alrady done?
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:39
ok sure - is this list of papers that you've put together more-or-less comprehensive? are you looking for more papers that may contain usefull stuff?
Travis Jacobs
@travs
Jun 21 2015 16:40
@stripathy What we're basically doing is going through each paper in that list at a high level, and grabbing plots that might be useful for building/optimizing our channel models
Stephen Larson
@slarson
Jun 21 2015 16:40
@cheelee no dependencies I think — several issues could get started with right away like openworm/muscle_model#37 or openworm/muscle_model#36
H
@ahrasheed
Jun 21 2015 16:40
@pgleeson @slarson would it be premature to work on a script like https://github.com/openworm/CElegansNeuroML/blob/master/CElegans/pythonScripts/SpreadsheetDataReader.py to accomplish reading data from PyOpenworm for this openworm/CElegansNeuroML#27
Travis Jacobs
@travs
Jun 21 2015 16:41
@stripathy It is not comprehensive, but is basically a place for us to get started
Chee Wai Lee
@cheelee
Jun 21 2015 16:41
@slarson cool, I'll go take a look!
Travis Jacobs
@travs
Jun 21 2015 16:41
@stripathy One place to contribute may be to pick up where Vahid and I have left off with noting the interesting plots, since you have experience with these types of papers
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:41
@travs where do the plots that you've grabbed go? into the channelworm user interface?
Stephen Larson
@slarson
Jun 21 2015 16:41
@ahrasheed Not premature at all — do check out the approach in openworm/CElegansNeuroML#11 and openworm/CElegansNeuroML#12 though as it was heading down a reasonable path
Travis Jacobs
@travs
Jun 21 2015 16:42
@stripathy Yes (eventually). We're actaully just listing them now and then we'll go back and grab them and run them through the pipeline
Stephen Larson
@slarson
Jun 21 2015 16:43
Yeah where’s that spreadsheet @travs ?
Travis Jacobs
@travs
Jun 21 2015 16:44
Joe Bowen
@joebowen
Jun 21 2015 16:44
Ok, #94 is solved.
Travis Jacobs
@travs
Jun 21 2015 16:44
Hey Joe! Awesome :)
Joe Bowen
@joebowen
Jun 21 2015 16:45
:-)
Looks like while web_app/wsgi.py wasn't needed for openshift, it was needed for local dev... oops. :-)
Travis Jacobs
@travs
Jun 21 2015 16:47
@stripathy I've updated #8 with links to the sheet and folder containing our ion channel paper/figure lists
Are you able to access both of those locations?
Stephen Larson
@slarson
Jun 21 2015 16:48
Thanks for the spreadsheet @travs — so @stripathy yeah; right now we are grabbing figures like this. Also worth mentioning that in addition to this @VahidGh is interested in also getting out specific parameters, which make up the input fields in the ChannelWorm app
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:49
@slarson @VahidGh can you list which parameters you'd like to get?
(just a few examples are good)
Stephen Larson
@slarson
Jun 21 2015 16:49
@ahrasheed you’ll notice the pull request is very old. it is based on the earlier version of PyOpenWorm — my expectation is that it is broken at this point
@ahrasheed so it needs someone to go in and do it correctly based on the latest version, since we just released this week
Travis Jacobs
@travs
Jun 21 2015 16:50
@stripathy check out this model for patch clamp experiments
Stephen Larson
@slarson
Jun 21 2015 16:50
@ahrasheed but definitely use that as a starting point
H
@ahrasheed
Jun 21 2015 16:50
@slarson ok .. gonna try and run it and make corrections according to the latest pyopenworm docs
Stephen Larson
@slarson
Jun 21 2015 16:51
awesome
Travis Jacobs
@travs
Jun 21 2015 16:51
@stripathy the variable names on the left should be all of them
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:52
@travs yeah this looks like a great start
Travis Jacobs
@travs
Jun 21 2015 16:52
:D
let me know if you have any questions!
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:55
@travs can any of those fields be null? seems a bit restrictive as is. Also, you'll need to extract out some basic experimental condition info at some point, like recording temperature
@travs @VahidGh would someone mind walking me through the plot digitization process? or really the workflow from paper -> digitized data?
Travis Jacobs
@travs
Jun 21 2015 16:59
@stripathy yes, I believe they are able to be empty if needed. And I agree on the condition info; the idea was that we will grab that afterward, but you're right, this needs to be incorporated into the model
@stripathy I can run you through the workflow right now in the google hangout if you'd like
Shreejoy Tripathy
@stripathy
Jun 21 2015 16:59
yeah, that'd be awesome
H
@ahrasheed
Jun 21 2015 17:02
@pgleeson thanks, got it. Understand the concern about openworm/PyOpenWorm#152 and looking forward to its resolution. For now using mwatt's script to see if it still works
Richard C Gerkin
@rgerkin
Jun 21 2015 17:07
OK back
@pgleeson I didn't quite follow your comment on my pull request. You want to leave Python 2.6 support in pyNeuroML?
Stephen Larson
@slarson
Jun 21 2015 17:09
@rgerkin he just put in the chat he needs to step away for a minute btw ...
Richard C Gerkin
@rgerkin
Jun 21 2015 17:11
OK. @slarson What do you think about Python 2.6 support in general for OpenWorm related repos? This may not be quite the right domain: http://astrofrog.github.io/blog/2015/05/09/2015-survey-results/, but impression is that no one is using 2.6.
Travis Jacobs
@travs
Jun 21 2015 17:11
@joebowen :'(
Not sure, maybe we're doing an unintentional ddos
Stephen Larson
@slarson
Jun 21 2015 17:12
@rgerkin are we talking 2.6 vs. 2.7 or 2.6 vs. 3?
Richard C Gerkin
@rgerkin
Jun 21 2015 17:12
So I'm generally trying to make sure what I write is both 2.7 and 3.4 compatible.
And not worrying about 2.6
Joe Bowen
@joebowen
Jun 21 2015 17:13
@travs sorry, that's me trying to fix the makemigrations script
Stephen Larson
@slarson
Jun 21 2015 17:13
Can you link to your pull request so I can get the context too?
Richard C Gerkin
@rgerkin
Jun 21 2015 17:13
NeuroML/pyNeuroML#8
Travis Jacobs
@travs
Jun 21 2015 17:13
@stripathy #71 is the collecting of plots and #72 is the actual digitization
Joe Bowen
@joebowen
Jun 21 2015 17:13
@travs, trying to fix VahidGh/ChannelWorm#90
Travis Jacobs
@travs
Jun 21 2015 17:13
@joebowen ahhh ok, that's fine as long as we know what's causing it :)
Joe Bowen
@joebowen
Jun 21 2015 17:15
@travs, so the issue is that the fields in ion_channel/models.py are "non-nullable fields" but don't have defaults. So when you do a makemigrations, it's prompts the cron script (which is the true problem, cause cron scripts aren't "promptable") to provide a one-off default. The real fix would be to give the fields defaults.
or make them nullable fields
Stephen Larson
@slarson
Jun 21 2015 17:16
Personally I think 2.7 support is all that is needed @rgerkin — I would reply in the pull request though to verify though just to be sure that’s what @pgleeson thinks too
Richard C Gerkin
@rgerkin
Jun 21 2015 17:16
@slarson Larger context is: NML2ChannelAnalysis was one giant function (intended for CLI via main) which produced a bunch of output, and I have been refactoring it so that individual steps are accessible to other python functions so make tighter SciUnit integration possible.
Stephen Larson
@slarson
Jun 21 2015 17:16
(as in 3.4 is good too, but 2.7 is the minimum)
Richard C Gerkin
@rgerkin
Jun 21 2015 17:17
I do all my work in 3.4 so I'll get the 3.4 for free... then I make sure 2.7 works as well.
Stephen Larson
@slarson
Jun 21 2015 17:17
that seems reasonable
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:21
Ok, I'm back...
@stripathy, We are not still stable on the digitization process. One issue related to this is #46, that made to start working on a new approach. If succeed, the process of the digitization and model generation would be much easier.
Travis Jacobs
@travs
Jun 21 2015 17:22
@joebowen ahhh, ok, this is something Shreejoy and I just talked about. We should make the fields nullable for now, and make defaults if appropriate later
@joebowen raising an issue right now on this
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:22
figure_1.png
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:22
@VahidGh no worries - digitization is a pain (why we don't do it yet for neuroelectro)
@VahidGh if you guys come up with a good solution for digitization that'd be really awesome for everyone
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:23
Right now working on some workflow that using the "automatic digitization" of the ditigizer
Joe Bowen
@joebowen
Jun 21 2015 17:24
@travs, sounds like a good plan. @all, It looks there are some ForeignKeys that are set to nullable (Graph.experiment and Graph.patch_clamp). Was that intentional? Or was the intention to make those non-nullable?
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:24
and having some digitized plot (dots) like the pic I posted, I can match the best model to the digitized points
Stephen Larson
@slarson
Jun 21 2015 17:24
the workflow that started here though, while not fully part of the pipeline, is at least usable
@joebowen maybe link directly to code for your question?
Milad Jafary
@miladjafary
Jun 21 2015 17:25
@joebowen yes they could be nullable
Joe Bowen
@joebowen
Jun 21 2015 17:25
@miladjafary ok, I'll go ahead and make the rest of the items that don't have defaults nullable
@slarson, not sure how to link to a specific line, but not sure it's needed now. Do you have a ref for how to do that in future?
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:26
the yellow point is the simulated current and the red and black are two different estimation that trying to match the best using different distance measurements
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:27
@VahidGh - what's the issue? that the image resolution quality is low?
Joe Bowen
@joebowen
Jun 21 2015 17:27
@slarson cool, thanks
Richard C Gerkin
@rgerkin
Jun 21 2015 17:27
blob
@vahidgh @travs Do you guys know why the model and data are so far out of alignment in https://github.com/VahidGh/ChannelWorm/blob/tests/tests/IVCurve_Model.ipynb (see bottom figure), also show above:
Black is data, red is model
Travis Jacobs
@travs
Jun 21 2015 17:28
@joebowen tracking what you're doing in nulling the model fields in #95 now
Stephen Larson
@slarson
Jun 21 2015 17:28
@joebowen also if you use that in the future, it is also helpful to know that typing ‘y’ when looking at code in github will switch to the permanent link to that specific version — even better b/c otherwise the code will change as revisions come in.
Joe Bowen
@joebowen
Jun 21 2015 17:29
@slarson ok, I'll keep that in mind.
Travis Jacobs
@travs
Jun 21 2015 17:29
@rgerkin I had the same question. Have not had time to run your notebook yet but it may have something to do with the scaling being done (I'll link the lines)
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:30
@stripathy, most of the figures in papers we are working on are not in a perfect resolution or size
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:30
@VahidGh yah... ok. =\
Travis Jacobs
@travs
Jun 21 2015 17:31
@rgerkin ok I cant link lines in a scipy notebook it seems, but is it maybe in block 14?
Richard C Gerkin
@rgerkin
Jun 21 2015 17:32
@travs Those are factors of 1000 to convert from mV to V, etc. but I think we are looking at about a factor of 10 here.
Travis Jacobs
@travs
Jun 21 2015 17:32
because the plot has a "dip" in the right location on the x-axis, it's just way too small
ok
I'm not sure then :\
Stephen Larson
@slarson
Jun 21 2015 17:34
@rgerkin do we have an issue to work through that question?
Richard C Gerkin
@rgerkin
Jun 21 2015 17:34
FYI, you can always link to a heading in an IPython notebook by slugifying the name of the header and appending it to URL, e.g. #Compare-the-predicted-to-the-observed added to URL name goes to heading "Compare the predicted to the observed"
@slarson No we do not. Creating now
Stephen Larson
@slarson
Jun 21 2015 17:34
Cool
Richard C Gerkin
@rgerkin
Jun 21 2015 17:36
Just noticed that the header linking does not work with new GitHub IPython notebook rendering -- only works when viewing in NBViewer or actual IPython notebook server
Travis Jacobs
@travs
Jun 21 2015 17:36
@rgerkin yeah :(
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:41
@travs is there any effort currently to encode the ion channel kinetics info too, e.g. Fig 4D here: https://drive.google.com/file/d/0B4qffTA1q81remJzZWtoWlFWY0U/edit
Stephen Larson
@slarson
Jun 21 2015 17:42
@stripathy ooh — I didn’t even realize folks plot that stuff…!
@stripathy I rarely see that in papers — but it seems valuable!
Travis Jacobs
@travs
Jun 21 2015 17:43
@stripathy oh awesome! Yeah, we should certainly record those in the sheet to refer back to
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:44
@slarson @travs yah, my sense too is that it'll be rare but will be there occasionally
Richard C Gerkin
@rgerkin
Jun 21 2015 17:44
Tufte's first law: for any two variables, they have been plotted against each other.
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:45
@slarson @travs there'll also be more non-uniformity in how they're calculated plotted unfortunately
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:46
@stripathy, We have a plan to also include those figures, but for now the priority is with VClamp, IClamp and I/V figures
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:46
@VahidGh is there a "curation guidelines" document anywhere? even if in beta?
Vahid Ghayoomie
@VahidGh
Jun 21 2015 17:46
@rgerkin, indeed :)
@stripathy, not yet
Stephen Larson
@slarson
Jun 21 2015 17:49
@VahidGh @stripathy closest thing to that is here tho: https://channelworm.readthedocs.org/en/latest/information-management/
and these docs are … wait for it … editable on github :)
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:50
@slarson @VahidGh haha ok, fair enough
@slarson did you hear back from the channelpedia folks?
Stephen Larson
@slarson
Jun 21 2015 17:51
@stripathy yes — Rajnish got back

So Yes, we have been redoing data collection for Channelpedia and I as of now our current status is following.

  1. We have ~90 stable cell lines expressing different ion channel genes and we have characterized complete (~40) Kv family (Kv1- Kv12) of Rat ion channel genes in CHO cell line.
  2. Automated patch clamp experiments were done at three different temperatures (15c, 25c, 35c) in a very controlled and consistent environment.
  3. We have electrophysiology data from ~11000 cells and out of them ~5000 we can really use for analysis and modeling. All the data I have recorded will be made available through Channelpedia.

We are currently working on an article to summarize our results on characterization of all Kv-ion channels. Till now the main conclusions are

  1. Each Kv-ion channel has a distinct electrical identity. (Ex: Kv_Ranjan_Emmanuelle_2015.png)
  2. Temperature has a very significant effect on ion channels and this effect is very rarely reported in literature and often ignored in models. (Ex: Kv1_1Model.png)
  3. Some ion channel responses are very variable where as some are very consistent. These observations stand against genes from different species (Mouse, Human) and different host cell lines (HEK, CV1).

We hope to publish it very soon. Please stay in touch and let me know if I can be of any help to your project.

Shreejoy Tripathy
@stripathy
Jun 21 2015 17:54
@slarson did he say anything about contributing more literature data to channelpedia?
Stephen Larson
@slarson
Jun 21 2015 17:54
for now it sounds like We do want to make Channelpedia opensource but right now the main focus is to publish the results.
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:55
@slarson sounds reasonable, thanks
Stephen Larson
@slarson
Jun 21 2015 17:55
but I was glad you reminded me about him
we had actually exchanged notes back in 2012, i only noticed when i searched my email :)
Shreejoy Tripathy
@stripathy
Jun 21 2015 17:56
@VahidGh @travs one thing I'm having trouble with with going through these articles is that i want to know what kinds of things I should be on the look out for - I/V vs G/Gmax vs V/I vs control vs KO
Richard C Gerkin
@rgerkin
Jun 21 2015 17:58
@stripathy @slarson Sharon tells me that they have a commitment to her to have every channel model available in NeuroML
Vahid Ghayoomie
@VahidGh
Jun 21 2015 18:00
@stripathy, @travs We can have a separate meeting about this, and take a look at some of the articles deeper
Stephen Larson
@slarson
Jun 21 2015 18:02
@rgerkin great!
Richard C Gerkin
@rgerkin
Jun 21 2015 18:03
FYI, no-jinx-fingers-crossed, but I think our R01 is getting funded so I should be able to spend a non-trivial amount of time on this stuff now
Travis Jacobs
@travs
Jun 21 2015 18:03
@stripathy right, we've been getting I/V curves, as well as I/time and V/time if they are present, and only for control. We've also noted whether the plots are whole-cell or per ion channel in the comments
@VahidGh We can do that.
Shreejoy Tripathy
@stripathy
Jun 21 2015 18:06
@travs @VahidGh yes, that sounds great
Vahid Ghayoomie
@VahidGh
Jun 21 2015 18:06
Great, @travs , would you please you make it? :D
Joe Bowen
@joebowen
Jun 21 2015 18:09
I'm working on getting the server back up and running, but I'm not sure what the default for patchclamp.deltat should be. makemigrations is asking for a one-off default, should I set it to 0?
Stephen Larson
@slarson
Jun 21 2015 18:09
@/all FYI all we should get back to the hangout to wrap up in 20 minutes — half past.
Travis Jacobs
@travs
Jun 21 2015 18:13
@joebowen I think null? Unless it has to be numeric
Joe Bowen
@joebowen
Jun 21 2015 18:13
it can't be null
I was going to set them just to be nullable but I guess there's a reason for not doing that.
Shreejoy Tripathy
@stripathy
Jun 21 2015 18:15
@joebowen just make them 0 or -1 for now?
@joebowen i can't tell you how many times me and @rgerkin came across this issue with neuroelectro and django migrations
Joe Bowen
@joebowen
Jun 21 2015 18:17
@stripathy ok, I'll set it to 0 and we can always change it later.
Vahid Ghayoomie
@VahidGh
Jun 21 2015 18:17
@joebowen, the problem is more general, and by changing every model, we still have the issue.
We can ignore the migrate for now and replacing it with the solution of syncing dbs between repos, till finding a better solution
Joe Bowen
@joebowen
Jun 21 2015 18:18
Ok, patch_clamp is back up.
Travis Jacobs
@travs
Jun 21 2015 18:18
@joebowen sweet!
Joe Bowen
@joebowen
Jun 21 2015 18:19
@VahidGh ok, we can look into it. @all, I'm actually gonna be heading out. Lots of fun! Lets do this again :-)
Travis Jacobs
@travs
Jun 21 2015 18:19
@VahidGh @stripathy Sent the invite to your inboxes, hopefully those times work :)
Joe Bowen
@joebowen
Jun 21 2015 18:19
@VahidGh I'll work on a method of syncing db's later this week.
Stephen Larson
@slarson
Jun 21 2015 18:19
@joebowen thanks !!
Joe Bowen
@joebowen
Jun 21 2015 18:19
@VahidGh Do we have a repo that we want to sync to already setup?
Travis Jacobs
@travs
Jun 21 2015 18:19
@joebowen Thanks Joe! We'll be in contact :)
Vahid Ghayoomie
@VahidGh
Jun 21 2015 18:20
@joebowen, great, not yet, but we can have a different branch for this purpose
Joe Bowen
@joebowen
Jun 21 2015 18:21
@VahidGh ok, sounds like a plan
Vahid Ghayoomie
@VahidGh
Jun 21 2015 18:22
@travs, yeah thanks, works for me
Travis Jacobs
@travs
Jun 21 2015 18:22
Excellent
Shreejoy Tripathy
@stripathy
Jun 21 2015 18:26
@VahidGh @travs I put some of my questions/concerns in a comment on #8 .
Travis Jacobs
@travs
Jun 21 2015 18:27
@stripathy Perfect, we'll refer to that in our meeting, thanks!
Vahid Ghayoomie
@VahidGh
Jun 21 2015 18:28
@stripathy, would be helpful, thanks
Travis Jacobs
@travs
Jun 21 2015 18:32
@/all ok folks, let’s get back into the hangout. If you didn’t make it in originally there are some spots open now… first come first served!