EPFL.hoc import

[Michael]

I'm trying to import in my Netpyne model neuron from EPFL .hoc/.nml
https://bbp.epfl.ch/nmc-portal/downloads.html
Are there already available means to do that together with morphology, ions, mechanisms etc?
 
If not, I am trying to write my own simple converter from .hoc to .json and then use
netParams.loadCellParamsRule(label='Erule', fileName='test.json')
netParams.cellParams['Erule']['conds'] = {'cellType': ['E']}
 
Would it support all .hoc parameters needed for simulation? I see there are some equations in mechanisms and not sure .json would support such parameters

[Joe Graham]

Hello,

You should be easily able to import cell models from .hoc, .json, and .py files.  Please see here for details and examples: http://netpyne.org/advanced.html#importing-externally-defined-cell-models

If your model has mechanisms defined in a .mod file, you will need to compile them in the directory where you are running your NetPyNE simulations.

Please give it a try and let us know if you have any problems.

Best wishes,

Joe

[Michael]

Good Day! Thank you for your answer. I am trying to launch it with this code

from netpyne import specs, sim, conversion

netParams = specs.NetParams()  # object of class NetParams to store the network parameters

cellRule = netParams.importCellParams(

        label='PYR_HH3D_hoc',

        conds={'cellType': 'PYR', 'cellModel': 'HH3D'},

        fileName='template.hoc',

        cellName='bAC217_L4_BP_d04c4872bd',

        importSynMechs=True)

I still get errors

/nrml_3d/L4_BP_bAC217_1$ python pynml.py 

1 

1 

NEURON: NaTs2_t is not a MECHANISM

 in biophysics.hoc near line 50

  insert NaTs2_t

                ^

        xopen("biophysics.hoc")

      execute1("{xopen("bi...")

    load_file("biophysics.hoc")

  xopen("template.hoc")

and others

0 

NEURON: ProbGABAAB_EMS is not a template

 in synapses.hoc near line 274

             section.sec synapse = new ProbGABAAB_EMS(seg_x)

                                                           ^

        xopen("synapses.hoc")

      execute1("{xopen("sy...")

    load_file("synapses/s...")

  xopen("template.hoc")

and others

0 

NEURON: init not enough arguments

 near line 0

 ^

        bAC217_L4_BP_d04c4872bd[0].init()

Traceback (most recent call last):

  File "pynml.py", line 40, in <module>

    importSynMechs=True)

  File "/home/ubuntu/.local/lib/python2.7/site-packages/netpyne/specs/netParams.py", line 401, in importCellParams

    secs, secLists, synMechs, globs = conversion.importCell(fileName, cellName, cellArgs, cellInstance)

  File "/home/ubuntu/.local/lib/python2.7/site-packages/netpyne/conversion/neuronPyHoc.py", line 280, in importCell

    cell = getattr(h, cellName)(**cellArgs)  # create cell using template, passing dict with args

RuntimeError: hoc error

Even after commenting all problem strings. Which I believe damaged synapses data

/nrml_3d/L4_BP_bAC217_1$ python pynml.py 

1 

1 

NEURON: NaTs2_t is not a MECHANISM

 in biophysics.hoc near line 50

  insert NaTs2_t

                ^

        xopen("biophysics.hoc")

      execute1("{xopen("bi...")

    load_file("biophysics.hoc")

  xopen("template.hoc")

and others

0 

1 

NEURON: init not enough arguments

 near line 0

 ^

        bAC217_L4_BP_d04c4872bd[0].init()

Traceback (most recent call last):

  File "pynml.py", line 40, in <module>

    importSynMechs=True)

  File "/home/ubuntu/.local/lib/python2.7/site-packages/netpyne/specs/netParams.py", line 401, in importCellParams

    secs, secLists, synMechs, globs = conversion.importCell(fileName, cellName, cellArgs, cellInstance)

  File "/home/ubuntu/.local/lib/python2.7/site-packages/netpyne/conversion/neuronPyHoc.py", line 280, in importCell

    cell = getattr(h, cellName)(**cellArgs)  # create cell using template, passing dict with args

RuntimeError: hoc error

[Michael]

still error

/nrml_3d/L4_BP_bAC217_1$ nrnivmodl ./mechanisms

..

Successfully created x86_64/special

/nrml_3d/L4_BP_bAC217_1$ python pynml.py 

1 

1 

1 

1 

NEURON: init not enough arguments

 near line 0

 ^

        bAC217_L4_BP_d04c4872bd[0].init()

Traceback (most recent call last):

  File "pynml.py", line 40, in <module>

    importSynMechs=True)

  File "/home/ubuntu/.local/lib/python2.7/site-packages/netpyne/specs/netParams.py", line 401, in importCellParams

    secs, secLists, synMechs, globs = conversion.importCell(fileName, cellName, cellArgs, cellInstance)

  File "/home/ubuntu/.local/lib/python2.7/site-packages/netpyne/conversion/neuronPyHoc.py", line 280, in importCell

    cell = getattr(h, cellName)(**cellArgs)  # create cell using template, passing dict with args

RuntimeError: hoc error

/nrml_3d/L4_BP_bAC217_1$ ./run_hoc.sh

..successfully loads gui

/nrml_3d/L4_BP_bAC217_1$ python pynml.py 

absolutely same error

[Joe Graham]

Whenever you get an error message that includes "NEURON: <mechName> is not a MECHANISM", that is typically going to mean that you either forgot to compile your mod files or a mod file didn't compile correctly...

One big problem I see in your error messages is that you're using Python 2.  This has been deprecated since January 2020, and we no longer actively support Python 2.  Please upgrade to Python 3 and see if you still get these errors (after compiling the mod files).

If you still get these errors with Python 3, it would be easiest if you could provide a link directly to the files you are using and your NetPyNE code (a GitHub repo is ideal).

Best wishes,

Joe

[Michael]

error was here template.hoc. strangely it worked with nrngui

proc init() { local synapses_enabled

    ///synapses_enabled = $1

MOD files were recompiled with nrnivmodl

now it loads and netParams.save('wow.json') writes the data OK. However 3D plot doesn't show this neuron. and netParams.popParams['PYR'] = {'cellType': 'PYR', 'numCells': 5, } also gives some error

my GitHub

https://github.com/MP1233/bg/blob/main/L4_BP_bAC217_1/pynml2.py

log so far is. probably some synapses refer to invalid mechanisms. however their nrngui example worked ok. investigating. probably commenting out invalid synapses will solve the problem and make model more robust

Added inhibitory synapse 39 originating from cell 149929 of m-type L6_LBC on basal section 5(0.446000) and dep 370.000000

Added inhibitory synapse 40 originating from cell 149929 of m-type L6_LBC on basal section 5(0.457000) and dep 1076.000000

 Could not create 

Saving netParams to wow.json ... 

Start time:  2021-09-26 21:13:44.886235

Creating network of 2 cell populations on 1 hosts...

ERROR: Some mechanisms and/or ions were not inserted (for details run with cfg.verbose=True). Make sure the required mod files are compiled.

  Number of cells on node 0: 6 

  Done; cell creation time = 0.13 s.

Making connections...

  Number of connections on node 0: 0 

  Done; cell connection time = 0.00 s.

Adding stims...

  Number of stims on node 0: 1 

  Done; cell stims creation time = 0.00 s.

Recording 1 traces of 1 types on node 0

NEURON: Arg out of range in user function

 near line 0

 ko0_k_ion = 2.5

                ^

        PtrVector(0)

Traceback (most recent call last):

  File "pynml2.py", line 97, in <module>

    sim.createSimulateAnalyze(netParams = netParams, simConfig = simConfig)    

  File "/home/ubuntu/.local/lib/python3.6/site-packages/netpyne/sim/wrappers.py", line 202, in createSimulateAnalyze

    (pops, cells, conns, stims, rxd, simData) = sim.create(netParams, simConfig, output=True)

  File "/home/ubuntu/.local/lib/python3.6/site-packages/netpyne/sim/wrappers.py", line 54, in create

    simData = sim.setupRecording()             # setup variables to record for each cell (spikes, V traces, etc)

  File "/home/ubuntu/.local/lib/python3.6/site-packages/netpyne/sim/setup.py", line 434, in setupRecording

    setupRecordLFP()

  File "/home/ubuntu/.local/lib/python3.6/site-packages/netpyne/sim/setup.py", line 327, in setupRecordLFP

    cell.imembPtr = h.PtrVector(nseg)  # pointer vector

RuntimeError: hoc error

[Michael]

Thanks! Now it works

Also while plotting there was a little bug in

analysis/lfp.py 

                     for j in range(i,i+nseg):  del trSegs[j]

replaced:

                     for j in range(i,i+nseg): 

                            if (i+nseg < len(trSegs)-1) : del trSegs[j]

just in case someone needs working .py code for EPFL example:

from netpyne import specs, sim, conversion

# Network parameters

netParams = specs.NetParams()  # object of class NetParams to store the network parameters

cellRule = netParams.importCellParams(

        label='PYR_HH3D_hoc',

        conds={'cellType': 'PYR', 'cellModel': 'HH3D'},

        fileName='createsimulation.hoc',

        cellName='bAC217_L4_BP_d04c4872bd',

        importSynMechs=True)

## Population parameters

netParams.popParams['PYR'] = {'cellType': 'PYR', 'numCells': 5, 'yRange': [700,800]}

netParams.cellParams['PYR_HH3D_hoc']['conds'] = {'cellType': ['PYR']}

## Synaptic mechanism parameters

netParams.synMechParams['exc'] = {'mod': 'Exp2Syn', 'tau1': 0.8, 'tau2': 5.3, 'e': 0}  # NMDA synaptic mechanism

# Stimulation parameters

netParams.stimSourceParams['bkg'] = {'type': 'NetStim', 'rate': 50, 'noise': 0.0}

netParams.stimTargetParams['bkg->all'] = {'source': 'bkg', 'conds': {'cellType': ['PYR']}, 'weight': 10.0, 'sec': 'soma', 'delay': 15, 'synMech': 'exc'}

# Simulation options

simConfig = specs.SimConfig()        # object of class SimConfig to store simulation configuration

simConfig.duration = 0.05*1e3           # Duration of the simulation, in ms

simConfig.dt = 0.1                # Internal integration timestep to use

simConfig.verbose = False            # Show detailed messages

simConfig.recordTraces = {'V_soma':{'sec':'soma','loc':0.5,'var':'v'}}

simConfig.recordStep = 1             # Step size in ms to save data (eg. V traces, LFP, etc)

simConfig.filename = 'cell_lfp'  # Set file output name

simConfig.recordLFP = [[x, y, 35] for y in range(280, 1000, 150) for x in [30, 90]]

simConfig.analysis['plotTraces'] = {'include': [('E',0)], 'oneFigPer':'cell', 'overlay': True, 'figSize': (5,3),'saveFig': True}      # Plot recorded traces for this list of cells

simConfig.analysis['plotLFP'] = {'includeAxon': False, 'plots': ['timeSeries',  'locations'], 'figSize': (5,9), 'saveFig': True}

simConfig.analysis['getCSD'] = {'spacing_um': 150, 'vaknin': True}

#simConfig.analysis['plotCSD'] = {'timeRange': [10,45],'saveFig': True}

#sim.analysis.getCSD(spacing_um= 150, vaknin= True)

#simConfig.analysis['plotCSD'] = {'saveFig': True}

# Create network and run simulation

sim.createSimulateAnalyze(netParams = netParams, simConfig = simConfig)    

sim.analysis.plotCSD()

# check model output

sim.checkOutput('cell_lfp')