These are chat archives for openworm/ChannelWorm

2nd
Jun 2015
Stephen Larson
@slarson
Jun 02 2015 15:26
@VahidGh and @miladjafary -- @travs and I are wanting to make improvements to Neurotune code but don't want to be redundant with what you are doing. Would you be comfortable pointing us to the current state of your code so we can see? Otherwise I fear we will re-invent the wheel
Vahid Ghayoomie
@VahidGh
Jun 02 2015 19:13
data_vs_candidate-IV.png
@slarson, Actually I've already finished the code. Resolving some bugs and playing with parameters.
As mentioned before I'm using the Optimization class, from Neurotune in addition to some other methods for simulation and calculating cost function for our purposes.
It just needs to be integrated with some data layer code (directly from GUI or pyOW) to initialize parameters, and the final NeuroML file generator.
I've tested the code for EGL-19 with hard-coded values and it's lovely I/V curve :) (with no I- or V-clamp plot), the result was not bad for the first try, and not so good with just one I/V plot (but small range of min_max values) for the evaluation!
Anyway, trying to test it with some current- or voltage clamp data and would upload it today if there was no bug during these new tests.
Of course we would need more help and investigation afterward...
Travis Jacobs
@travs
Jun 02 2015 21:54
@VahidGh Looking good Vahid!
Looking forward to seeing the code and helping push development of the optimization suite forward!
Vahid Ghayoomie
@VahidGh
Jun 02 2015 22:58
@travs, Thanks, of course. Working to resolve some minor problems...