Questions about running Stochastic Simulation Algorithm (SSA) in COPASI

[Konstantia Georgouli]

Hello,

My name is Konstantia Georgouli and I am a postdoctoral researcher in Lawrence Livermore National Lab. 

COPASI is a great work and I am really interested in learning some more information regarding this simulation software tool.

I have checked your github repo but I still have some questions.

Specifically, it would be great if you could answer the following questions:

 

1. Is there any limit to the number of reactants in case of running Stochastic Simulation Algorithm (SSA)?
2. Moreover, in SSA, is it allowed any arbitrary formula?
3. If yes, how do you support this? And do you require a specific way from user perspective?

 

I am looking forward to hearing from you.

Thank you. 

Kind regards,

Konstantia

[Hoops, Stefan (sh9cq)]

Hello Konstantina,

Please find my answers below

On Mon, 2021-02-08 at 06:05 -0800, Konstantia Georgouli wrote:
> Hello,
> My name is Konstantia Georgouli and I am a postdoctoral researcher in
> Lawrence Livermore National Lab. 
> COPASI is a great work and I am really interested in learning some
> more information regarding this simulation software tool.
> I have checked your github repo but I still have some questions.
> Specifically, it would be great if you could answer the following
> questions:
>  

>    1. Is there any limit to the number of reactants in case of

> running Stochastic Simulation Algorithm (SSA)?

I have several interpretations of this question which I try to answer
all:
1) You can have an arbitrary number of reactants R1 + R2 + ... Rn
2) The stoichiometry must be a positive integer
3) There is a maximum number of particles a reactant may have it must
fit into a 64 bit signed integer

>    2. Moreover, in SSA, is it allowed any arbitrary formula?

Yes, though the rate law should not be explicitly time dependent since
that would violate the SSA assumption that the propensities do not
change between reaction events.

>    3. If yes, how do you support this? And do you require a specific
> way from user perspective?
The only thing which we optionally do is to correct for higher order
reactions e.g. 2 A -> B will lead in to A * (A-1) * k where k is rate
constant. This does not change whether k is a constant or a function of
A. The user can turn this behavior on or of for the whole model if they
do not want this behavior.

Thanks,
Stefan


--
Stefan Hoops, Ph.D.
Research Associate Professor
Biocomplexity Institute & Initiative
University of Virginia
995 Research Park Boulevard
Charlottesville, VA 22911

Phone: +1 540 570 1301
Email: sho...@virginia.edu

[Konstantia Georgouli]

Dear Stefan,

Thank you very much for your answers. Following your reply, I would like to ask you one more question. 

Does COPASI rely on the generation of machine code using a certain method such as LLVM, or on AST 

traversal to enable arbitrary expressions of reaction rate formulas?

I am looking forward to hearing from you.

Thank you.

Kind regards,

Konstantia

[Hoops, Stefan (sh9cq)]

Hello Konstantia,

The currently released versions are using an AST to process arbitrary
functions. However the next release will use JIT compiled assembler
code for the 64 bit platforms.

Thanks,
Stefan