|
|
Post by Stefan on Jun 13, 2021 7:41:44 GMT
Hello,
What is the following warning supposed to mean?
Warning: during the simulation some reaction probabilities were greater than 1. You man want to return the simulation with the WARNINGS BLOCK enable to get more detail.
Does this influences simulation result?
Thank you.
Stefan
|
|
|
|
Post by Adam on Jun 14, 2021 13:17:49 GMT
Hello Stefan,
thank you for your question.
Yes, this warning influences the simulation result.
This is related to the probability of a reaction happening when two molecules collide.
When it is higher than one, this means that some reactions will be missed and
the simulation result will be wrong.
If you run Blender/CellBlender from a terminal (command line) and run the simulation,
you will be able to see which specific reaction caused this warnin (MCell prints it to the terminal).
Another reason might be due to surface molecule+surface molecules reactions.
Even throug the individual reaction probability is <= 1, a surface molecule first diffuses to the destination location and
thenall neighbors are checked for reaction. In the case the sum of all possible reaction probabilities might be > 1.
A solution might be to decrease the time step of the simulation or one can also set a custom time step for the molecules
Best regards,
Adam
In the MCell 3.5.1 version, some of these examples print this warning (this was fixed but the fixed examples were not released yet).
|
|
|
|
Post by Ron on Jul 28, 2021 14:47:10 GMT
Hello Adam, thanks for your explanation.
I've got a similar warning message during running my simulation while i'm trying to count reactions between transmitted volume molecules and surface molecules that randomly located on a circle surface. Changing the time step indeed improve the error probabilities, but it didn't solve it and i still have probabilities exceeded 1. I also changed the Surface Grid Density parameter and it doesn't have a significant influence. Are there any additional parameters recommended to deal with ?
I was thinking about :
1. INTERACTION_RADIUS - change the reaction volume (distance) - is there any way to change it ?
2. random distribution of the molecules on the circle surface - maybe it is associated with the way I fill the circle surface - the fill type (Fan/NGON...?) or the number of subdivision in order to make the molecules uniformly distributed on the circle surface ? - what's the correct way to make it ?
Any idea how can i deal with that ?
Thank you,
Ron
|
|
|
|
Post by Adam on Aug 3, 2021 16:11:53 GMT
Hi Ron,
so if I understand correctly, the warning is printed for the volume molecule - surface molecule reaction, right?
This is how the probability factor for the reaction is computed:
pb_factor = 1.0e11 * surface_grid_density / (2.0 * N_AV) * sqrt(PI * t_step / D_tot);
prob = pb_factor * rate;
t_step - time step of the volume molecule (global time step if the molecule has no custom time step)
D_tot - diffusion constant of the volume molecule
(please note that some of the values are scaled in MCell into internal units so you won't get the same result if you would try to compute this manually)
When a volume molecule hits a tile and the tile contains a surface molecule, then a possible reaction between these two molecules is evaluated.
So (1) INTERACTION_RADIUS won't gave any impact here (it is used only for volume-volume reactions), otherwise it can be set in "Run Simulation" -> "Advanced options".
The only way to lower the probability of the reaction further is to decrease the time step or set a smaller custom time step for the volume molecule.
With MCell4, there is an option under "Run Simulation" -> "Notifications" -> "Species and Reactions Report" that enables generation of file reports/rxn_report.txt where the probabilities of reactions are logged.
The distribution of molecules on surface (2) won't have any effect either.
If I should explain anything better, please let me know.
Adam
|
|
