bertram_2004

Calcium and Glycolysis Mediate Multiple Bursting Modes in Pancreatic Islets - Bertram, Satin, Zhang, Smolen, Sherman - 2004 ==================================================================================

Original model status:

Not curated, does not run in PCEnv simulator.


Model Analysis:


Compared with XPPAUT .ode file, has some differently named variables, also has some equations which aren't in the .ode file and others which are different. Recoding the model to match the .ode file is required before asking for any help from the authors - the best help possible is available: the original source code.

Have tried to identify how the model code fits with what is in the .ode file and have failed. Most likely easier and faster to simply recode the model based on the .ode file.

  • PCEnv Error:

Error: Model is underconstrained.

The following variables couldn't be defined:
  • fraction_state
  • W00000
  • W11100
  • ATP4
  • g_K_ATP_
  • O_infinity_ADP_ATP
  • sum_state
  • R_PFK
  • ATP
  • AMP
Following variables are defined in (out) component "glycolysis":
  • fraction_state
  • W00000
  • w11100
  • sum_state
  • glycolysis
Following variables are defined in component "nucleotide_sensitive_potassium_current":
Following variables are defined in (out) component "adenine_nucleotides":
  • ATP
  • AMP

Work done on model:

Bertram2004_XPPAUT .ode file

Attempted to translate the .ode file into CellML. For some reason PCEnv is telling me that the model is overconstrained, and is pointing to one of the math elements involved in the iterative calculation of PFK (phosphofructokinase). File works on XPPAUT...

Will send this issue to the mailing list to see if anyone can help.

Update: Andre pointed out an error, which has allowed me to fix the model. It now integrates and produces nice output. For some reason, however, the integrator refuses to go past a certain point.

Having fixed the model tt now produces the correct output for one cycle of bursting. After once cycle all of the parameters that are not constants suddenly turn into NaN's. We suspected this might be something to do with the few square roots that are in the model, so we change the equation to root the absolute value of the variable (fbp) being rooted. This worked in that the integrator was now no longer stopping, but then after a few cycles of bursting, the output went totally awry. No idea how to fix this, will probably leave it for later.

Curation Status:

Model is running and producing the right output. However PCEnv is having trouble integrating it and only one cycle of bursting is being produced. Version 01 Variant 01 replaced with working model. Version 01 Variant 02 has been uploaded, with the sqrt(abs(fbp)) addition, allowing a few bursting cycles to be simulated before the system breaks down.

Futher work required:

Find out why output is turning chaotic after several cycles.


Return to PMR Curation Wiki

subtopics: