Team:Evry/model integration

From 2012.igem.org

(Difference between revisions)
m
 
(10 intermediate revisions not shown)
Line 3: Line 3:
<h1>Model integration - Epilogue</h1>
<h1>Model integration - Epilogue</h1>
-
Modelling an organism on multiple scales is not an easy task. It requires usually to create various models in order to capture different kind of features and to take into account specific aspects related to the considered scale.
+
<p>Modelling an organism on multiple scales is not an easy task. It requires usually to create various models in order to capture different kind of features and to take into account specific aspects related to the considered scale.
<br/><br/>
<br/><br/>
Nevertheless, many multi-scale models can represent fairly well what is happening at various scales without really being connected together. Our team's endeavour is aiming at proposing an integration across various scales and modelling solutions.  
Nevertheless, many multi-scale models can represent fairly well what is happening at various scales without really being connected together. Our team's endeavour is aiming at proposing an integration across various scales and modelling solutions.  
<br/><br/>
<br/><br/>
In this we try to show that composite models, with each bit representing a single aspect of a problem, have to be designed having a bigger picture in mind. The general objective we pursued was to really be able to link all our models to the bench-work, to experimental data. Of course, achieving such a big project can not totally be done on the iGEM time scale because too many specific experiments would be required in order to calibrate properly all the pieces of our "big picture". But what is achievable, although much demanding, in the course of a summer is to come up with a very precise plan and all the models ready, and fitting together. Were future competitors willing to build upon this work and carry over the suggested experiments, they would end up with a very useful tool for science and for designing complex constructs in a multicellular organism.
In this we try to show that composite models, with each bit representing a single aspect of a problem, have to be designed having a bigger picture in mind. The general objective we pursued was to really be able to link all our models to the bench-work, to experimental data. Of course, achieving such a big project can not totally be done on the iGEM time scale because too many specific experiments would be required in order to calibrate properly all the pieces of our "big picture". But what is achievable, although much demanding, in the course of a summer is to come up with a very precise plan and all the models ready, and fitting together. Were future competitors willing to build upon this work and carry over the suggested experiments, they would end up with a very useful tool for science and for designing complex constructs in a multicellular organism.
-
 
+
</p>
Line 16: Line 16:
</center>
</center>
 +
 +
<h2> Details on integration </h2>
 +
 +
<center>
 +
<img width="800px"src="http://2012.igem.org/wiki/images/8/81/Integration1.png" alt="Details part one"/>
 +
</center>
 +
 +
<center>
 +
<img width="800px"src="http://2012.igem.org/wiki/images/b/b7/Integration2.png" alt="Details part one"/>
 +
</center>
 +
 +
<h1> The new Integrated Model </h1>
 +
 +
<h2>The Theory</h2>
 +
 +
The integrated-model looks like the following scheme:
 +
 +
<center>
 +
  <img src="http://2012.igem.org/wiki/images/a/aa/Integrated-model.png" alt="integrated model" />
 +
</center>
 +
 +
<p>
 +
On this scheme, an arrow from one model (or block) to another means that the output (the curve) from the one is the input of the second. The integration sign is in fact a scaling method that allows us to have result on an higher scale. For example, integrating the behaviour of one cell at the tissue scale as presented above. The output produced by this model is simply a curve of the quantity of GFP as a function of time and model the behaviour of our auxin degradation system.
 +
</p>
 +
 +
<br/>
 +
The important points of this integrated-model are:<br/>
 +
 +
<ul>
 +
  <li>It is composed of multiple sub-models</li>
 +
  <li>The way each models are wired is important</li>
 +
  <li>Changing one model can either add: a change in behaviour or a gain (or loss) of precision</li>
 +
</ul>
 +
<br/>
 +
<p>
 +
The key mechanisms in this models are not on the models but on the transitions: integrating and derivating. The integration happens when an input model is on a lower scale than its successor and allows to simulate the behaviour of a homogeneous higher system composed of multiple instances of this model. For example, integrating the behaviour a creation model at the cell scale can be used to simulate the behaviour of a tissue. On the other hand, derivating a model means interpreting it at each time step and downsiding it of one scale. For example, from the transport model to the degradation model.
 +
</p>
 +
<h2>The implementation</h2>
 +
You can find<a href="http://2012.igem.org/File:Integrated-model.zip">here</a> the whole model. You just have to open the file complete.m and follow its instructions.
 +
 +
<h2> Limitations </h2>
 +
 +
There exists some limitations with this model: you can't do feedback loop or make two sub-models interacts.
<script type="text/javascript">writeFooter()</script>
<script type="text/javascript">writeFooter()</script>
</html>
</html>

Latest revision as of 03:59, 27 October 2012

Model integration - Epilogue

Modelling an organism on multiple scales is not an easy task. It requires usually to create various models in order to capture different kind of features and to take into account specific aspects related to the considered scale.

Nevertheless, many multi-scale models can represent fairly well what is happening at various scales without really being connected together. Our team's endeavour is aiming at proposing an integration across various scales and modelling solutions.

In this we try to show that composite models, with each bit representing a single aspect of a problem, have to be designed having a bigger picture in mind. The general objective we pursued was to really be able to link all our models to the bench-work, to experimental data. Of course, achieving such a big project can not totally be done on the iGEM time scale because too many specific experiments would be required in order to calibrate properly all the pieces of our "big picture". But what is achievable, although much demanding, in the course of a summer is to come up with a very precise plan and all the models ready, and fitting together. Were future competitors willing to build upon this work and carry over the suggested experiments, they would end up with a very useful tool for science and for designing complex constructs in a multicellular organism.

The global modelling picture
The big picture

Details on integration

Details part one
Details part one

The new Integrated Model

The Theory

The integrated-model looks like the following scheme:
integrated model

On this scheme, an arrow from one model (or block) to another means that the output (the curve) from the one is the input of the second. The integration sign is in fact a scaling method that allows us to have result on an higher scale. For example, integrating the behaviour of one cell at the tissue scale as presented above. The output produced by this model is simply a curve of the quantity of GFP as a function of time and model the behaviour of our auxin degradation system.


The important points of this integrated-model are:
  • It is composed of multiple sub-models
  • The way each models are wired is important
  • Changing one model can either add: a change in behaviour or a gain (or loss) of precision

The key mechanisms in this models are not on the models but on the transitions: integrating and derivating. The integration happens when an input model is on a lower scale than its successor and allows to simulate the behaviour of a homogeneous higher system composed of multiple instances of this model. For example, integrating the behaviour a creation model at the cell scale can be used to simulate the behaviour of a tissue. On the other hand, derivating a model means interpreting it at each time step and downsiding it of one scale. For example, from the transport model to the degradation model.

The implementation

You can findhere the whole model. You just have to open the file complete.m and follow its instructions.

Limitations

There exists some limitations with this model: you can't do feedback loop or make two sub-models interacts.