Team:ETH Zurich/Modeling/Stochastic

From 2011.igem.org

(Difference between revisions)
(Method)
(Analysis of Results)
 
(70 intermediate revisions not shown)
Line 3: Line 3:
{{:Team:ETH Zurich/Templates/SectionStart}}
{{:Team:ETH Zurich/Templates/SectionStart}}
= Stochastic Analysis =
= Stochastic Analysis =
-
'''Besides the parameter space search, we performed another type of analysis in order to verify that our system is robust and that it is not bistable. We were especially interested in the GFP band, whether it is always present and whether the amount of GFP produced in the cells for different concentrations of Acetaldehyde has large fluctuations. Because gene expression is an intrinsically stochastic process, we performed stochastic simulations to see how our system reacts to noise and how it responds to perturbations.'''  
+
'''Besides the parameter space search, we performed another type of analysis in order to verify that our system is robust and that it is monostable (i.e. it has only one steady state). We were especially interested in the GFP band, whether it is always present and whether the amount of GFP produced in the cells for different concentrations of Acetaldehyde has large fluctuations. Because gene expression is an intrinsically stochastic process, we performed stochastic simulations to see how our system reacts to noise and how it responds to perturbations.'''  
<br> <br>
<br> <br>
-
'''The following results show that our system is indeed robust and it does not have bifurcations.'''
+
'''The following results show that our system is indeed robust to noise and it does not have bifurcations.'''
{{:Team:ETH Zurich/Templates/SectionEnd}}
{{:Team:ETH Zurich/Templates/SectionEnd}}
{{:Team:ETH Zurich/Templates/SectionStart}}
{{:Team:ETH Zurich/Templates/SectionStart}}
= Method =
= Method =
-
For the stochastic simulations we used the [http://magnet.systemsbiology.net/software/Dizzy/ Dizzy] software, to which we gave as an input an SBML file with the description of our system. Moving from deterministic to stochastic simulations we had to change several things in the SBML file. Firstly, we exported the SBML file to a CMDL format.Next, we converted the species values as well as the production rates and repression coefficients from concetrations (in &mu;M) to numbers of molecules. Since our initial system was described in terms of ODEs, we also had to change this by separating the system into distinct reaction channels (i.e. separate degradation from activation/repression) and specifying the reaction rates.
+
For the stochastic simulations we used the [http://magnet.systemsbiology.net/software/Dizzy/ Dizzy] software, to which we gave as an input an SBML file with the description of our system. Moving from deterministic to stochastic simulations we had to change several things in the SBML file. Firstly, we exported the SBML file to a CMDL format.Next, we converted the species values as well as the production rates and repression coefficients from concentrations (in &mu;M) to number of molecules. Since our initial system was described in terms of ODEs, we also had to change this by separating the system into distinct reaction channels (i.e. separate degradation from activation/repression) and specifying the reaction rates.
-
 
+
<br> <br>
-
With the new description of the system, we could start the simulations. We started with a deterministic simulation to take the deterministic steady states of the species, after 5000 min(the acetaldehyde amount was the one that leads to the maximal GFP value, i.e the peak of the band). Starting from deterministic steady states, we performed a stochastic simulation with Gillespie-direct algorithm for 10000 min in order to get the stochastic steady states. After that, we restarted the stochastic simulation having the stochastic steady state values as our initial and simulated for 100 000 more minutes (storing the GFP values every minute). We collected all the GFP values from the last simulation and plotted them in a histogram (150 bins).
+
With the new description of the system, we could start the simulations. For analyzing the robustness of the system, we chose 5 strategic Acetaldehyde concentrations (see Figure 1), and for each of them we performed stochastic simulations. We selected the following Acetaldehyde concentrations:
-
 
+
* 0 &mu;M (0 molecules). At this concentration the GFP output should approach 0 &mu;M.
-
[[File:GFPmlabel.png|thumb|900px|center|'''Figure 1''': GFP versus Acetaldehyde. The red squares denote the 5 Acetaldehyde concentrations for which we simulated the time course of GFP steady state in the stochastic mode.]]
+
* 25 &mu;M (30110 molecules). At this concentration GFP rises quickly, so it should be very sensitive to noise.
 +
* 46 &mu;M (55402 molecules). This is the concentration that gives the peak GFP value.
 +
* 100 &mu;M (120440 molecules). At this concentration GFP is decreasing, so it should be sensitive to noise.
 +
* 300 &mu;M (361320 molecules). At this concentration the GFP output should again approach 0 &mu;M.  
 +
<br> <br>
 +
[[File:GFPmlabel.png|thumb|900px|center|'''Figure 1''': GFP versus Acetaldehyde. The red squares denote the 5 Acetaldehyde concentrations for which we simulated the time course of GFP steady state in stochastic mode.]]
 +
In order to be sure that we are indeed simulating the time course of the "stochastic steady state", and that we are not in the transient regime, we implemented the following approach:
 +
* We ran a deterministic simulation for 5000 minutes (~ 83 hours) to get to the deterministic steady state. We took over the concentration values for the last time point as new initial conditions for the next simulation.
 +
* We ran a stochastic simulation for 10000 minutes (~ 7 days), to make sure that we are somewhere around the “stochastic steady state”. Once again we took over the concentration values for the last time point as new initial conditions for the next simulation.
 +
* We ran a long stochastic simulation for 100000 minutes (~ 70 days) storing the molecule numbers of the species every minute.
 +
<br>
 +
For the stochastic simulations we used the "Gillespie-direct" option, since this gives an exact solution, compared to the other solvers which only approximate the solution.
{{:Team:ETH Zurich/Templates/SectionEnd}}
{{:Team:ETH Zurich/Templates/SectionEnd}}
Line 21: Line 32:
= Simulations =
= Simulations =
 +
After we obtained the samples with Dizzy, we loaded the data in Matlab to analyze it. To get an initial feeling of how the system behaves in the stochastic regime, we plotted the GFP molecule numbers as function of time. In Figure 2 we can see that for the peak of the band (Acetaldehyde = 46 &mu;M), GFP values fluctuate around a mean value which is about 23500 molecules.
-
{| border="0"
+
[[File:Gfpfl50.png|thumb|900px|center|'''Figure 2''': Stochastic simulation of GFP steady state for an Acetaldehyde concentration of 46 &mu;M.]]  
-
|- valign="top"
+
Next, in order to see how the values are distributed and whether there are multiple steady states, we plotted the normalized histograms for the 5 data sets. We used 150 bins for plotting the histograms for the 3 intermediate Acetaldehyde values and 28 and 61 bins for the first and last Acetaldehyde value, respectively. In these last 2 cases we had to reduce the number of bins because the range of the GFP molecule values was narrower.
-
|[[File:Gfp0.png|thumb|400px|center|'''Figure 10''': Exploring the parameter space of TetR degradation rate]]
+
-
 
+
-
|[[File:Gfp25.png|thumb|400px|center|'''Figure 11''': Exploring the parameter space of CI degradation rate]]
+
-
|-
+
-
|colspan="2"|
+
-
 
+
 +
{| border="0" width="100%"
|- valign="top"
|- valign="top"
-
|[[File:Gfpfl50.png|thumb|400px|center|'''Figure 10''': Exploring the parameter space of TetR degradation rate]]
+
|[[File:Gfp0.png|thumb|400px|center|'''Figure 3''': Distribution of the number of GFP molecules for an Acetaldehyde input of 0 &mu;M]]
-
 
+
|[[File:Gfp25.png|thumb|400px|center|'''Figure 4''': Distribution of the number of GFP molecules for an Acetaldehyde input of 25 &mu;M]]
-
|[[File:Gfp50.png|thumb|400px|center|'''Figure 11''': Exploring the parameter space of CI degradation rate]]
+
|-
|-
|colspan="2"|
|colspan="2"|
Line 40: Line 46:
|- valign="top"
|- valign="top"
-
|[[File:Gfp100.png|thumb|400px|center|'''Figure 10''': Exploring the parameter space of TetR degradation rate]]
+
|[[File:Gfp50.png|thumb|400px|center|'''Figure 5''': Distribution of the number of GFP molecules for an Acetaldehyde input of 46 &mu;M]]
-
 
+
|[[File:Gfp100.png|thumb|400px|center|'''Figure 6''': Distribution of the number of GFP molecules for an Acetaldehyde input of 100 &mu;M]]
-
|[[File:Gfp300.png|thumb|400px|center|'''Figure 11''': Exploring the parameter space of CI degradation rate]]
+
|-
|-
|colspan="2"|
|colspan="2"|
-
 
-
We see in Figure 1 that we obtain a Gaussian distribution of GFP number of molecules at the steady state, with a mean of 23491 molecules and standard deviation of 667 molecules. If we convert this value to concentration we get 19.5046 &mu;M, which is very close to what we obtained deterministically. We see that the system keeps fluctuating very closely to its steady state, with no large jumps away from it. The fact that the Gaussian distribution is unimodal tells us that our system is monostable. From this analysis we confirmed what we concluded before and that is the robustnes of our system, especially of the GFP band. We can now be sure that for a certain acetaldehyde concentration once the band appears, it won't start fading away.
 
|}
|}
 +
 +
[[File:Gfp300.png|thumb|400px|center|'''Figure 7''': Distribution of the number of GFP molecules for an Acetaldehyde input of 300 &mu;M]]
 +
{{:Team:ETH Zurich/Templates/SectionEnd}}
{{:Team:ETH Zurich/Templates/SectionEnd}}
{{:Team:ETH Zurich/Templates/SectionStart}}
{{:Team:ETH Zurich/Templates/SectionStart}}
= Analysis of Results =
= Analysis of Results =
-
''' Insert overview of page here '''
+
From the time course plot (Figure 2) we can already see that the values always fluctuate around one mean value which leads us to believe that the system is stable. Since the numbers of GFP molecules do not jump between two or more distinct values we can assume that the system is not bistable. The plot alone however does not constitute conclusive evidence, so for this reason we generated and analyzed the distributions of GFP values.
 +
<br> <br>
 +
All of the 5 distributions have a Gaussian-like shape, which means that the GFP values cluster around one mean value. The distributions are relatively narrow, which means there are no large fluctuations and there are unimodal, which corresponds to a having no multiple stead states in the deterministic regime.
 +
 
 +
{|style="margin: 1em auto 1em auto; width: 80%;" class="wikitable sortable"
 +
!Acetaldehyde concentration
 +
!GFP mean (no. of molecules)
 +
!GFP mean (&mu;M)
 +
!Deterministic GFP value (&mu;M)
 +
!GFP standard deviation (no. of molecules)
 +
!GFP standard deviation (&mu;M)
 +
!GFP relative standard deviation (%)
 +
|-
 +
|0 &mu;M
 +
|12.2903
 +
|0.0102 &mu;M
 +
|0.01 &mu;M
 +
|3.5473
 +
|0.0029 &mu;M
 +
|28.86 %
 +
|-
 +
|25 &mu;M
 +
|1954.4
 +
|1.6227 &mu;M
 +
|1.7157 &mu;M
 +
|544.6184
 +
|0.4522 &mu;M
 +
|27.86 %
 +
|-
 +
|46 &mu;M
 +
|23491
 +
|19.5046 &mu;M
 +
|19.6046 &mu;M
 +
|669.5291
 +
|0.5559 &mu;M
 +
|2.85 %
 +
|-
 +
|100 &mu;M
 +
|2981.7
 +
|2.4757 &mu;M
 +
|2.4830 &mu;M
 +
|243.6979
 +
|0.2023 &mu;M
 +
|8.17 %
 +
|-
 +
|300 &mu;M
 +
|40.0056
 +
|0.0332 &mu;M
 +
|0.0332 &mu;M
 +
|7.1426
 +
|0.0059 &mu;M
 +
|17.85 %
 +
|}
 +
 
 +
In the table above we can see that the mean values of the distributions of GFP molecule numbers (in concentration) obtained from the stochastic simulations are very similar to the deterministic ones obtained previously by simulating the system with an ODE solver. This confirms once again the fact that the deterministic steady state corresponds to the stochastic one, thus it sees that we do not have bifurcations..
 +
<br> <br>
 +
Analyzing the relative standard deviations (RSD), we notice that the highest value was obtained for an input concentration of acetaldehyde equal to 0 &mu;M. This is due to the fact that in this case GFP is repressed, so there is only leaky GFP expression. As the GFP output is on the order of a few up to tens of molecules, the ratio between the value of the mean and the amplitude of the fluctuations is much higher than the absolute value of the fluctuations.
 +
<br> <br>
 +
The lowest RSD corresponds to the acetaldehyde concentration that gives the GFP peak (46 &mu;M). This is in accordance to the central limit theorem, since for higher values of the mean the relative value of the fluctuations decreases. We can notice that for 46 &mu;M of acetaldehyde the standard deviation is on the same order of magnitude as for 25 and 100 &mu;M of acetaldehyde, however the RSD is much smaller for the peak value.
 +
<br> <br>
 +
The RSD is higher in the region where the GFP increases compared the region where it decreases despite the fact the value of the mean is comparable. This can be explained by taking into account the steepness of the slope which is higher for 25 than for 100 &mu;M of acetaldehyde. On the left side of the peak the GFP value changes more quickly than on the right side, meaning that it is more prone to larger fluctuations.
 +
 
{{:Team:ETH Zurich/Templates/SectionEnd}}
{{:Team:ETH Zurich/Templates/SectionEnd}}
{{:Team:ETH Zurich/Templates/HeaderNewEnd}}
{{:Team:ETH Zurich/Templates/HeaderNewEnd}}

Latest revision as of 22:27, 28 October 2011

Can you feel the smoke tonight?
 

Contents

Stochastic Analysis

Besides the parameter space search, we performed another type of analysis in order to verify that our system is robust and that it is monostable (i.e. it has only one steady state). We were especially interested in the GFP band, whether it is always present and whether the amount of GFP produced in the cells for different concentrations of Acetaldehyde has large fluctuations. Because gene expression is an intrinsically stochastic process, we performed stochastic simulations to see how our system reacts to noise and how it responds to perturbations.

The following results show that our system is indeed robust to noise and it does not have bifurcations.

Method

For the stochastic simulations we used the [http://magnet.systemsbiology.net/software/Dizzy/ Dizzy] software, to which we gave as an input an SBML file with the description of our system. Moving from deterministic to stochastic simulations we had to change several things in the SBML file. Firstly, we exported the SBML file to a CMDL format.Next, we converted the species values as well as the production rates and repression coefficients from concentrations (in μM) to number of molecules. Since our initial system was described in terms of ODEs, we also had to change this by separating the system into distinct reaction channels (i.e. separate degradation from activation/repression) and specifying the reaction rates.

With the new description of the system, we could start the simulations. For analyzing the robustness of the system, we chose 5 strategic Acetaldehyde concentrations (see Figure 1), and for each of them we performed stochastic simulations. We selected the following Acetaldehyde concentrations:

  • 0 μM (0 molecules). At this concentration the GFP output should approach 0 μM.
  • 25 μM (30110 molecules). At this concentration GFP rises quickly, so it should be very sensitive to noise.
  • 46 μM (55402 molecules). This is the concentration that gives the peak GFP value.
  • 100 μM (120440 molecules). At this concentration GFP is decreasing, so it should be sensitive to noise.
  • 300 μM (361320 molecules). At this concentration the GFP output should again approach 0 μM.



Figure 1: GFP versus Acetaldehyde. The red squares denote the 5 Acetaldehyde concentrations for which we simulated the time course of GFP steady state in stochastic mode.

In order to be sure that we are indeed simulating the time course of the "stochastic steady state", and that we are not in the transient regime, we implemented the following approach:

  • We ran a deterministic simulation for 5000 minutes (~ 83 hours) to get to the deterministic steady state. We took over the concentration values for the last time point as new initial conditions for the next simulation.
  • We ran a stochastic simulation for 10000 minutes (~ 7 days), to make sure that we are somewhere around the “stochastic steady state”. Once again we took over the concentration values for the last time point as new initial conditions for the next simulation.
  • We ran a long stochastic simulation for 100000 minutes (~ 70 days) storing the molecule numbers of the species every minute.


For the stochastic simulations we used the "Gillespie-direct" option, since this gives an exact solution, compared to the other solvers which only approximate the solution.


Simulations

After we obtained the samples with Dizzy, we loaded the data in Matlab to analyze it. To get an initial feeling of how the system behaves in the stochastic regime, we plotted the GFP molecule numbers as function of time. In Figure 2 we can see that for the peak of the band (Acetaldehyde = 46 μM), GFP values fluctuate around a mean value which is about 23500 molecules.

Figure 2: Stochastic simulation of GFP steady state for an Acetaldehyde concentration of 46 μM.

Next, in order to see how the values are distributed and whether there are multiple steady states, we plotted the normalized histograms for the 5 data sets. We used 150 bins for plotting the histograms for the 3 intermediate Acetaldehyde values and 28 and 61 bins for the first and last Acetaldehyde value, respectively. In these last 2 cases we had to reduce the number of bins because the range of the GFP molecule values was narrower.

Figure 3: Distribution of the number of GFP molecules for an Acetaldehyde input of 0 μM
Figure 4: Distribution of the number of GFP molecules for an Acetaldehyde input of 25 μM


Figure 5: Distribution of the number of GFP molecules for an Acetaldehyde input of 46 μM
Figure 6: Distribution of the number of GFP molecules for an Acetaldehyde input of 100 μM
Figure 7: Distribution of the number of GFP molecules for an Acetaldehyde input of 300 μM


Analysis of Results

From the time course plot (Figure 2) we can already see that the values always fluctuate around one mean value which leads us to believe that the system is stable. Since the numbers of GFP molecules do not jump between two or more distinct values we can assume that the system is not bistable. The plot alone however does not constitute conclusive evidence, so for this reason we generated and analyzed the distributions of GFP values.

All of the 5 distributions have a Gaussian-like shape, which means that the GFP values cluster around one mean value. The distributions are relatively narrow, which means there are no large fluctuations and there are unimodal, which corresponds to a having no multiple stead states in the deterministic regime.

Acetaldehyde concentration GFP mean (no. of molecules) GFP mean (μM) Deterministic GFP value (μM) GFP standard deviation (no. of molecules) GFP standard deviation (μM) GFP relative standard deviation (%)
0 μM 12.2903 0.0102 μM 0.01 μM 3.5473 0.0029 μM 28.86 %
25 μM 1954.4 1.6227 μM 1.7157 μM 544.6184 0.4522 μM 27.86 %
46 μM 23491 19.5046 μM 19.6046 μM 669.5291 0.5559 μM 2.85 %
100 μM 2981.7 2.4757 μM 2.4830 μM 243.6979 0.2023 μM 8.17 %
300 μM 40.0056 0.0332 μM 0.0332 μM 7.1426 0.0059 μM 17.85 %

In the table above we can see that the mean values of the distributions of GFP molecule numbers (in concentration) obtained from the stochastic simulations are very similar to the deterministic ones obtained previously by simulating the system with an ODE solver. This confirms once again the fact that the deterministic steady state corresponds to the stochastic one, thus it sees that we do not have bifurcations..

Analyzing the relative standard deviations (RSD), we notice that the highest value was obtained for an input concentration of acetaldehyde equal to 0 μM. This is due to the fact that in this case GFP is repressed, so there is only leaky GFP expression. As the GFP output is on the order of a few up to tens of molecules, the ratio between the value of the mean and the amplitude of the fluctuations is much higher than the absolute value of the fluctuations.

The lowest RSD corresponds to the acetaldehyde concentration that gives the GFP peak (46 μM). This is in accordance to the central limit theorem, since for higher values of the mean the relative value of the fluctuations decreases. We can notice that for 46 μM of acetaldehyde the standard deviation is on the same order of magnitude as for 25 and 100 μM of acetaldehyde, however the RSD is much smaller for the peak value.

The RSD is higher in the region where the GFP increases compared the region where it decreases despite the fact the value of the mean is comparable. This can be explained by taking into account the steepness of the slope which is higher for 25 than for 100 μM of acetaldehyde. On the left side of the peak the GFP value changes more quickly than on the right side, meaning that it is more prone to larger fluctuations.


Back to iGEM Our Sponsors
ETHZ-BASF.png ETH Zurich Logo.png ETHZ-Lonza.png ETHZ-Merck Serono.png
ETHZ-Novartis.png ETHZ-Roche.png ETHZ-Syngenta.png DSM MasterLogo.png