Team:ETH Zurich/human/essay/project

From 2014.igem.org

(Difference between revisions)
(Complexity in our project)
Line 1: Line 1:
<html><article></html>
<html><article></html>
== Complexity in our project ==
== Complexity in our project ==
 +
;'''Wet lab'''
 +
 +
At the very beginning the work in the wet lab seemed straightforward. The way we planed our time was ambitious, but soon after the start we faced first problems. In different experiments cross-talk between our different quorum sensing molecules was observed. The next steps focused on quantifying this cross-talk, this was never achieved completely. Identifying interactions on the subparts (molecules) level of the cell in order to understand the whole system behavior.
 +
One of the central questions in our project was the observation of emergence. The techniques of rapid prototyping and 3D-printing followed by PDMS molding allowed us to use custom-designed millifluidic chips. The chip with the grid on it allowed us to observe the phenomenon of guided emergence. In this case we do not have a simplification but a contextualisation. In the design of our experiments, we took into account the different factors.
 +
Living beings, even if not multicellular have naturally emergent properties.
 +
While facing problems with our own constructs, especially to mention the integrases, we decided to attempt a reproduction of the logic XOR implemented in Bonnet paper.[9]. We faced many problems and on the way of trying to reproduce the experiments they did. On this way our team had many interesting discussions concerning the reproducibility of data in general and in this specific case.
 +
Another important lesson we have learned in the lab is the debugging constructs. Many times the biology did at first glance not work as expected. Further examination has then shown mistakes in the design, mistakes of the person conducting the experiment and also taught us about the influence of parameters that have a priori not been taken into account. The cell as open system which interacts with its environment.
 +
 +
 +
;'''Modeling'''
 +
: A model is always a simplistic representation of reality. It contains assumptions. Using standard descriptions, like chemical reactions and mass action, we analytically derived each formula we wanted to fit. This derivation was only possible thanks to some assumptions. As these [http://2014.igem.org/Team:ETH_Zurich/modeling/int page] shows, every assumption was carefully made and thought about. The process of simplification was put into question because our human practice project tends to investigate why simplification is powerful but not enough to understand the surrounding world. In our case, simplifying was necessary. That's why we tried to understand what every assumptions could make sense in a biological way.
 +
 +
: Our modeling part is focused on parameter fitting (see our [http://2014.igem.org/Team:ETH_Zurich/modeling/parameters parameter page]). We use a classical deterministic model and tried to fit it to the experiments. Matching the reality level with the description level is a complicated task, as there is no optimal match. Differences and similarities between experiments and simulations give insights on where emergent phenomena could happen.
 +
 +
: As randomness is a property of complex systems, it motivated us to derive a stochastic model. Some biological events, like binding, are typically stochastic phenomena. The human practice part inspired us in a certain sense.
 +
 +
: We opted for an engineered representation (see the [http://2014.igem.org/Team:ETH_Zurich/project/infopro information processing page]). It was natural to divide our systems into submodules. Thus, we took into account the fact that different levels of description can give several insights on the system works. The decomposition into interacting submodules (see the [http://2014.igem.org/Team:ETH_Zurich/modeling/overview modeling overview page]) was crucial to face this complex problem.
 +
<html></article></html>
<html></article></html>

Revision as of 01:20, 18 October 2014

Complexity in our project

Wet lab

At the very beginning the work in the wet lab seemed straightforward. The way we planed our time was ambitious, but soon after the start we faced first problems. In different experiments cross-talk between our different quorum sensing molecules was observed. The next steps focused on quantifying this cross-talk, this was never achieved completely. Identifying interactions on the subparts (molecules) level of the cell in order to understand the whole system behavior. One of the central questions in our project was the observation of emergence. The techniques of rapid prototyping and 3D-printing followed by PDMS molding allowed us to use custom-designed millifluidic chips. The chip with the grid on it allowed us to observe the phenomenon of guided emergence. In this case we do not have a simplification but a contextualisation. In the design of our experiments, we took into account the different factors. Living beings, even if not multicellular have naturally emergent properties. While facing problems with our own constructs, especially to mention the integrases, we decided to attempt a reproduction of the logic XOR implemented in Bonnet paper.[9]. We faced many problems and on the way of trying to reproduce the experiments they did. On this way our team had many interesting discussions concerning the reproducibility of data in general and in this specific case. Another important lesson we have learned in the lab is the debugging constructs. Many times the biology did at first glance not work as expected. Further examination has then shown mistakes in the design, mistakes of the person conducting the experiment and also taught us about the influence of parameters that have a priori not been taken into account. The cell as open system which interacts with its environment.


Modeling
A model is always a simplistic representation of reality. It contains assumptions. Using standard descriptions, like chemical reactions and mass action, we analytically derived each formula we wanted to fit. This derivation was only possible thanks to some assumptions. As these page shows, every assumption was carefully made and thought about. The process of simplification was put into question because our human practice project tends to investigate why simplification is powerful but not enough to understand the surrounding world. In our case, simplifying was necessary. That's why we tried to understand what every assumptions could make sense in a biological way.
Our modeling part is focused on parameter fitting (see our parameter page). We use a classical deterministic model and tried to fit it to the experiments. Matching the reality level with the description level is a complicated task, as there is no optimal match. Differences and similarities between experiments and simulations give insights on where emergent phenomena could happen.
As randomness is a property of complex systems, it motivated us to derive a stochastic model. Some biological events, like binding, are typically stochastic phenomena. The human practice part inspired us in a certain sense.
We opted for an engineered representation (see the information processing page). It was natural to divide our systems into submodules. Thus, we took into account the fact that different levels of description can give several insights on the system works. The decomposition into interacting submodules (see the modeling overview page) was crucial to face this complex problem.