Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Mar 31 13:39
  • Aug 27 2018 14:25
    gsarma reopened #148
  • Aug 27 2018 14:25
    gsarma reopened #147
  • Aug 27 2018 14:25
    gsarma reopened #146
  • Aug 27 2018 14:25
    gsarma reopened #145
  • Aug 27 2018 14:25
    gsarma reopened #144
  • Aug 27 2018 14:25
    gsarma reopened #143
  • Aug 27 2018 14:25
    gsarma reopened #141
  • Aug 27 2018 14:25
    gsarma reopened #138
  • Aug 27 2018 14:24
    gsarma reopened #134
  • Aug 27 2018 14:24
    gsarma reopened #126
  • Aug 27 2018 14:24
    gsarma reopened #125
  • Aug 27 2018 14:24
    gsarma reopened #114
  • Aug 27 2018 14:24
    gsarma reopened #113
  • Aug 27 2018 14:24
    gsarma reopened #108
  • Aug 27 2018 14:24
    gsarma reopened #106
  • Aug 27 2018 14:24
    gsarma reopened #97
  • Aug 27 2018 14:24
    gsarma reopened #89
  • Aug 27 2018 14:23
    gsarma reopened #80
  • Aug 27 2018 14:23
    gsarma reopened #68
Stephen Larson
@slarson
Welcome @krisdamato
krisdamato
@krisdamato
thanks!
Stephen Larson
@slarson
@shah-n — Can you explain the problems you are running into in here?
Neil Shah
@shah-n
I received run time errors when running the python files which build the EGL-19 and SLO-2 ion channels.
Here is the EGL-19 error:
Traceback (most recent call last):
File "EGL-19.py", line 66, in <module>
best_candidate = candidates.optimize(do_plot=True, seed=1234)
File "/home/computer/git/channelworm/src/neurotune/neurotune/optimizers.py", line 142, in optimize
individuals_file=ind_file)
File "/home/computer/.local/lib/python2.7/site-packages/inspyred/ec/ec.py", line 434, in evolve
initial_fit = evaluator(candidates=initial_cs, args=self._kwargs)
File "../../../channelworm/fitter/evaluators.py", line 141, in evaluate
self.mySimulator = Simulator(self.sim_params,cand_var,cell_var,self.bio_params['gate_params']).patch_clamp()
File "../../../channelworm/fitter/simulators.py", line 22, in init
self.pc_type = sim_params['pc_type']
KeyError: 'pc_type'
The optimization for SLO-2 runs successfully, but the following run-time error gets produced.
Traceback (most recent call last):
File "SLO-2.py", line 222, in <module>
myModelator.compare_plots(sampleData,bestSim,show=True)
File "../../../channelworm/fitter/modelators.py", line 192, in compare_plots
x_var = sampleData['VClamp']['x_var']
KeyError: 'x_var'
Neil Shah
@shah-n
The EGL-36 ion channel gets built, but the file is recognized as a text file instead of an XML file on Ubuntu.
Vahid Ghayoomie
@VahidGh
@shah-n, Thanks for checking this. The example file generating EGL-36 model, is using the latest version of cwFitter that getting inputs from DB instead of files (compare the initiators of EGL-36 and SLO-2).
Other examples should be updated as well based on these changes.
We have a plan to complete data and build models for ion channels with digitized data from patch-clamp experiments.
Feel free to play with scripts and models and let me know if there is any question!
Stephen Larson
@slarson
@VahidGh Any idea how @shah-n can fix the error?
Stephen Larson
@slarson
@VahidGh — looks like EGL-19-2.py works where EGL-19.py doesn’t. Can we replace the one that doesn’t work?
Stephen Larson
@slarson
@VahidGh — we are running into problems that there is no method called “ss_plots” in modelators.py -- this is being called in EGL-19-2.py and SLO-2.py -- do you know why there is no ss_plots? Maybe we are not importing something we need?
Vahid Ghayoomie
@VahidGh
@slarson, Since there were some examples to show how to use cwFitter, I didn't go further to update all examples with new changes.
We were going to integrate the Django app with cwFitter, that I switched to the predictor project.
Stephen Larson
@slarson
@VahidGh Gotcha. Well @shah-n is aiming to go through and read EGL-36.py to best understand what is doing, and to add comments to it. Then, he’ll read through EGL-19-2.py, comment it, and understand how they are similar / different. Your perspectives on what are the pros / cons of each will help. We’d like to help out meaningfully to be able to convert all ion channel data into neuroml files. Does this seem reasonable? How would you advise us to proceed?
Vahid Ghayoomie
@VahidGh
Great, I suggest starting with a new channel based on EGL-36.
For example SHK-1, with ID=16.
Stephen Larson
@slarson
Nice :thumbsup: I think that will be a great one once we can make sure that EGL-19-2.py SLO-2.py is also working again.
I also want to loop in @rgerkin here to say that in addition to this, we want to flesh out the SciUnit aspect of these that is exemplified in the iPython notebook he made here: https://github.com/openworm/ChannelWorm/blob/0b1c1f49b71785f5333e3c86d2ed97832d75a8f5/tests/SCU_IVCurve_Model.ipynb
That example could use some updating, since the latest fits from the fitter project are much better than what is shown in that notebook
Vahid Ghayoomie
@VahidGh
@slarson, We can also start with some intros to the fitting process.
I would be glad to have a meeting with @shah-n to discuss more.
Stephen Larson
@slarson
OK great @VahidGh — I’m sending you an invite for next Friday for getting together with Neil (anyone else welcome). Let me know if that time doesn’t work for you
Vahid Ghayoomie
@VahidGh
@slarson, Works for me, thanks.
Stephen Larson
@slarson
@VahidGh and @shah-n ; maybe let’s just chat on here?
Vahid Ghayoomie
@VahidGh
Hi everyone
Stephen Larson
@slarson
Hey Vahid - -thanks for trying to join
If you give us a few minutes; we are pushing out what neil has done to github
then we will send you a link and you can comment on it and we can discuss
Neil Shah
@shah-n
Here is the link for the comments I currently have on EGL-36.py:
Stephen Larson
@slarson
@VahidGh — check these out — @shah-n is going through the code to check his understanding — can you see if those comments agree with your understanding?
Neil Shah
@shah-n
I haven't yet commented this line
Neil Shah
@shah-n
@VahidGh: When running any of the ion channel files (example: EGL-36.py) in Windows, the import files are not found. I do not have this problem when using Ubuntu. Do you know how I can avoid the import issues when using Windows?
Vahid Ghayoomie
@VahidGh
@shah-n, I haven't tested it on Windows yet. If the error occurs when importing files from relative paths, then it could be due to fore-slash/back-slash issue in paths. We should consider using os.path.join() instead of fore-slash directory separators.
However, if the error occurs when trying to import installed packages, maybe using Virtualenv, Ananconda, etc... could help!
Thanks for mentioning this issue.
Neil Shah
@shah-n
@VahidGh: For EGL-36.py, when I change the value of POV_ID from 11 to 8 (changing the input graph), I get the same NML output file. Is that expected?
Vahid Ghayoomie
@VahidGh
@shah-n, Have you changed the optimization method (opt parameter)?
By default it's value is None, so the optimized values would be those hard-coded ones. If you choose one of the available options like -pso, then you will notice that it takes some time for fitter to optimize the curves with new values and generate new plots.
Also, make sure to change parameters like file_name, etc. for generating a new .nml file.
Neil Shah
@shah-n
@VahidGh: Changing the optimization method is what I needed to do. Thanks.
Neil Shah
@shah-n
@VahidGh: For EGL-19-2.py, can you please tell me where the figure ID's are assigned and the NML file is created?
Vahid Ghayoomie
@VahidGh
@shah-n, This file is using an old version of fitter that getting data from file instead of DB.
The digitizer app has also a feature to export data to csv files after digitization. So, if you need to work independent of this DB you can use this method.
Vahid Ghayoomie
@VahidGh
NML file generation process could be added too, but the results would be the same as EGL-19.py, so didn't need to.
Neil Shah
@shah-n
@VahidGh: Do these comments agree with your understanding of EGL-36.py? https://github.com/openworm/ChannelWorm/compare/master...shah-n:master
Vahid Ghayoomie
@VahidGh
@shah-n, Yeah, it's OK for understanding the big picture. We need to get a bit deeper into the codes to understand the values and parameters we used. If you think would be better to have a chat/Hangout for this conversation, please let me know.
Neil Shah
@shah-n
@VahidGh: Thank you for offering to help. Please let me know whenever you are available for a discussion.
Stephen Larson
@slarson
@shah-n @VahidGh just pinging you both post end of the calendar year-- have you had a chance to reconnect?
Neil Shah
@shah-n
@slarson: We did not get a chance to reconnect.
Vahid Ghayoomie
@VahidGh
@shah-n, I've sent you a message with a Doodle link some time ago.
Would you please check if you can see my message in Gitter private messages.
Neil Shah
@shah-n
@VahidGh : Sorry, I did not notice your private message. I have now responded to it.
Vahid Ghayoomie
@VahidGh
@shah-n, No problem. Updated the poll, would you please check again.
Brain302
@Brain302
Good Day! Just a concept of an experiment 3D (worm scheme2D vs time1D) chart. Sorry if something like this already exists. Wonder if something like this can be added to experiment output diagrams. And also which functionalities can be added to this. Currently it does slices and tries to find possible cascades from neuron https://github.com/Brain302/somebrowser/blob/master/README.md