Full sensitivity functions over time

[micronaut]

Date: Sun, 12 Aug 2012 11:56:27 -0400
Subject: Re: Full sensitivity functions over time

Hi,

Sorry for dredging up an old forum post. I was interested to know what the difference would be between running sensitivity analysis on the steady-state vs the end point subtasks, if the system reaches steady state by the end of the time-course?

I have a model that does this, but the sensitivity results are very different between steady state and time-course subtasks.

I'm wondering what the modelling difference would be between sensitivity analysis performed on the 2 subtasks?

Thanks for any assistance.

> Hello!
> Can you please tell me how to generate the plot (report?) of full sensitivity functions over time, as the ODE solns are generated? I've done all this:
>
> "Please select 'Sensitivities' as the subtask and add the duration of the time course as the scan parameter. To select the duration you will have to switch to the expert mode and choose:
> TaskListTime-Course.(Problem)Time-Course.Duration"
>
> but I don't know what kind of the output should I choose. Is it possible to get a plot of the sensitivity of one chosen species over time (each curve = each parameter)?
>
> > Sincerely,
> > Kate

[shoops]

Date: Mon, 13 Aug 2012 10:54:34 -0400
Subject: Re: Re: Full sensitivity functions over time

Hello Micronaut,

>
> I'm wondering what the modelling difference would be between sensitivity analysis performed on the 2 subtasks?

There are several reasons for possible differences:
1) A steady-state is not guaranteed to be reachable on a trajectory from the initial time point.

2) The numerics are different and the accuracy in which you are calculating the steady state.

In my experience the best way to perform sensitivity analysis on a steady-state is to deselect the Newton method and just use forward integration.

Thanks,
Stefan

[Diána Philipp]

Hi!

I'm interested in this old post, which asks for full sensitivity function over time.

Could you please tell me how to generate a plot or report of full sensitivity function over time?

  I've done all this:

 

"Please select 'Sensitivities' as the subtask and add the duration of the time course as the scan parameter. To select the duration you will have to switch to the expert mode and choose:

TaskListTime-Course.(Problem)Time-Course.Duration"

 

And which output should I choose? Is it possible to get a plot of the sensitivity of one chosen species over time (each curve = each parameter)?

And I also get a warning every time when I try to do the full sensitivity function over time:

CSteadyState (1): The model is explicitly time dependent. Therefore, the calculation of a steady state using the Newton method is not very meaningful.

Where can I change steady state or where can I deselect the Newton method?

 

Thank you!

Sincerely,

Diana

[Stefan Hoops]

Hell Diana,

On Mon, 26 Oct 2015 14:17:13 -0700 (PDT)
Diána Philipp <philip...@gmail.com> wrote:

>
> Hi!
> I'm interested in this old post, which asks for full sensitivity
> function over time.
>
> Could you please tell me how to generate a plot or report of full
> sensitivity function over time?
>
> I've done all this:
>
>
>
> "Please select 'Sensitivities' as the subtask and add the duration of
> the time course as the scan parameter. To select the duration you
> will have to switch to the expert mode and choose:
>
> TaskListTime-Course.(Problem)Time-Course.Duration"
>
>
>
> And which output should I choose? Is it possible to get a plot of the
> sensitivity of one chosen species over time (each curve = each
> parameter)?
>
> And I also get a warning every time when I try to do the full
> sensitivity function over time:
>
> CSteadyState (1): The model is explicitly time dependent. Therefore,
> the calculation of a steady state using the Newton method is not very
> meaningful.

You may have forgotten to select Time Course as the subtask of the
sensitivities.

>
> Where can I change steady state or where can I deselect the Newton
> method?

This can be done in the steady-state task.

Thanks,
Stefan



--
Stefan Hoops, Ph.D.
Research Assistant Professor
Virginia Bioinformatics Institute
Virginia Tech
1015 Life Science Circle (0477)
Blacksburg, Va 24061, USA

Phone: (540) 231-1799
Fax: (540) 231-2606
Email: sho...@vbi.vt.edu

[Diána Philipp]

 Dear Stefan,

 Thank you for the fast answer! A corrected my mistakes, but I still don't know a few things.

 I would like to make a plot where the x axis is time, and the y axis is d(one specific species)/d(all the parameters), so I could see the sensitivity in time. How can I make this plot for that one specific species (so e.g. with 8 parameters I would get 8 curves, each parameter = each curve)?

 Sincerely,

 Diana

[Stefan Hoops]

Hello Diana,

Create a plot and create a new curve. Select time as the x-axis and
select the sensitivity matrix as the y-axis, You will be prompted to
which element you want to draw.

Thanks,
Stefan