Using global data

[jsy...@york.ac.uk]

Hi,

Is there a way to get a reference of a specific agent in
updateSpAgentState() function?
I'm trying to get a distance between every cell of one type and a specific
cell every loop.
Is using global data the only way? If so, how do I store data types like
SpAgentState?
Thanks for your help.

[seunghwa.kang]

Hello,

If I am understanding your question correctly,

you want to evaluate one to all interactions.

For example, you may want to find the distance between one immune cell and every bacteria cell in the simulation domain.

First Biocellion is not designed to support this type of computing. Biocellion assumes cells directly interact with only neighboring cells and interaction among distant cells occur only indirectly through the surrounding environment or other cells in between.

I may be able to find a better way if you explain why you want to do this but without this knowledge, I can suggest few tricks... but if evaluating one to all interactions is critical, Biocellion may not be a good fit for your case.

First, you can update global data only in the initGlobal function, so using global data for this purpose is not advisable... If you're an advanced C programmer and you don't care about running Biocellion on distributed memory clusters and you know how to handle data synchronization, you can use global data for this purpose... but it's more like hacking Biocellion rather than properly using Biocellion.

Second, you can do post processing using Paraview. I am not an expert in Paraview, but it's pretty likely that you can post process simulation output using Paraview to compute one to all distances; you may need to write a custom Paraview filter. If Paraview does not support this, you may need to write your own program for this.

Third, you can use Biocellion's summary mechanism (Look for SummaryOutputInfo, updateSummaryOutputInfo, and getRecentSummaryRealVal in the Biocellion 1.2 user manual). You may define three summary variables for x, y, z coordinate values for the cell you want to find one to all distances, set summary values to the position of the cell in the unit box containing the cell and 0 for all the other boxes, and set the reduction type to sum. (this may not be the most efficient way but computing one to all interactions becomes inefficient anyway if your simulation domain becomes really big).

I think you can take the second or the third options, but I am really curious why you need to do this; If evaluating one to all interactions is critically important to model many biological systems, I need to think about updating Biocellion to support evaluating one to all interactions.

Thank you very much,

-seunghwa