Team:Warsaw/ExpressionAdaptors/Solution

From 2011.igem.org

(Difference between revisions)
 
(7 intermediate revisions not shown)
Line 12: Line 12:
<h2>Our Solution</h2>
<h2>Our Solution</h2>
-
 
-
<!-- <div class="note">Here goes something about our project</div> -->
 
-
 
-
<br />
+
<br>
<div align="justify">
<div align="justify">
Line 27: Line 24:
<li>Promoter is not a part of expression adapter because it is a DNA sequence and is not translated to mRNA</li>
<li>Promoter is not a part of expression adapter because it is a DNA sequence and is not translated to mRNA</li>
</ul>
</ul>
 +
</div>
 +
<br>
 +
<h2>Computational design of the expression adapters</h2>
 +
<br>
 +
<div align="justify">
 +
Expression adapters were created by a genetic algorithm that we designed specifically for this task. It generates random 5 bp spacers and 10 amino acid long proteins. Proteins were constructed out of only 11 amino acids, those that does not direct to the N- degradation pathway [1]. Genetic algorithm was used because there is too many combinations of random spacers and 10 aa proteins to computationally test them all. It would take around 2 years and we wouldn't make it for iGEM 2011;)
 +
Genetic algorithm looked  like this:
 +
<br />
 +
<img src="https://static.igem.org/mediawiki/2011/4/40/Genetic.png"></img>
 +
<br />
-
<br /><br />
+
First we generated set of random expression adapters, then we in silico fused them with several fluorescent proteins, namely GFP, SF-GFP, YFP, mORANGE and RFP - to model their behavior with different proteins. We submited the sequences to a locally installed version of the RBS calculator – an algorithm developed by Christopher A. Voigt [2]. The calculator considers parameters like fold of the protein or complementarity with the ribosomal RNA and returns the infomation of the protein expression strength.  
-
Expression adapters were created by a genetic algorithm that we designed specifically for this task. It generates random 5 bp spacers and 10 amino acid long proteins. Proteins were constructed out of only 11 amino acids, those that does not direct to the N- degradation pathway [1]. Genetic algorithm was used because there is too many combinations of random spacers and 10 aa proteins to test them all. First generation of spacers and proteins is submitted to several rounds of random mutations and recombination. The outcome sequences from all the rounds of genetic alterations were then gathered in one list. <br /><br /> After removal of the duplicates sequences were scored using RBS calculator – an algorithm developed by Christopher A. Voigt [2]. The calculator considers parameters like fold of the protein, complementarity with the ribosome and... Each adapter sequence is modeled with several fluorescent proteins, namely GFP, SF-GFP, YFP, mORANGE and RFP. Those with the best score are then processed with another software tool that chooses adapters with the best (highest) expression levels and lowest deviations between different proteins. <br /><br />
+
In genetic algorithm we assessed (out fitness function took into account) two parameters:
-
From the final forty sequences six adapters providing six different levels of expression were chosen manually and synthesized in Laboratory of DNA Sequencing and Oligonucleotide Synthesis in the Institute of Biochemistry and Biophysics, Polish Academy of Sciences (Warsaw, Poland). The final step was to test them in the wet lab.  </div>
+
<ul>
 +
<li> the standard deviation of the strength (as computed by RBC calculator)for one expression adapter and various proteins
 +
Low standard deviation means that this particular expression adapter strongly determines the fold of the mRNA around an RBS part.
 +
</li>
 +
<li> strength (as computed by RBC calculator) - we evolved sequences to get the highest expression.
 +
</li>
 +
</ul>
 +
The best adapters were chosen. Then random mutations and recombination were introduced to sequences and the population was assessed again.
 +
From all generated sequences six adapters providing six different levels of expression and having standard deviation close to 0 (gave the same results for all tested proteins) were chosen and synthesized. The final step was to test them in the wet lab.  </div>
<br />
<br />
</div>
</div>

Latest revision as of 12:24, 21 September 2011

Example Tabs

Our Solution


To make truly standardize strength of RBS parts we propose a brand new part in synthetic biology - we call expression adapters. They are short sequences consisting of RBS, 5 bp spacer and a short (10 amino acid long) beginning of a protein. An expression adapter look like this:

Why are they constructed this way:
  • RBS-spacer-and-the-beginning-of-a-protein determin the fold of the mRNA around the RBS. This way the fold, and subsequently the strength of the RBS is independent of the protein used.
  • Spacer is a piece of DNA that is between RBS (e.g. B0034) and ORF (open reading frame). It keeps the optimal distance between those two parts - needed for efficient protein expression
  • Promoter is not a part of expression adapter because it is a DNA sequence and is not translated to mRNA

Computational design of the expression adapters


Expression adapters were created by a genetic algorithm that we designed specifically for this task. It generates random 5 bp spacers and 10 amino acid long proteins. Proteins were constructed out of only 11 amino acids, those that does not direct to the N- degradation pathway [1]. Genetic algorithm was used because there is too many combinations of random spacers and 10 aa proteins to computationally test them all. It would take around 2 years and we wouldn't make it for iGEM 2011;) Genetic algorithm looked like this:

First we generated set of random expression adapters, then we in silico fused them with several fluorescent proteins, namely GFP, SF-GFP, YFP, mORANGE and RFP - to model their behavior with different proteins. We submited the sequences to a locally installed version of the RBS calculator – an algorithm developed by Christopher A. Voigt [2]. The calculator considers parameters like fold of the protein or complementarity with the ribosomal RNA and returns the infomation of the protein expression strength. In genetic algorithm we assessed (out fitness function took into account) two parameters:
  • the standard deviation of the strength (as computed by RBC calculator)for one expression adapter and various proteins Low standard deviation means that this particular expression adapter strongly determines the fold of the mRNA around an RBS part.
  • strength (as computed by RBC calculator) - we evolved sequences to get the highest expression.
The best adapters were chosen. Then random mutations and recombination were introduced to sequences and the population was assessed again. From all generated sequences six adapters providing six different levels of expression and having standard deviation close to 0 (gave the same results for all tested proteins) were chosen and synthesized. The final step was to test them in the wet lab.



  1. The N-end rule pathway of protein degradation. Varshavsky A. Genes Cells. 1997 Jan;2(1):13-28.
  2. Automated design of synthetic ribosome binding sites to control protein expression Howard M. Salis, Ethan A. Mirsky & Christopher A. Voigt, Nature Biotechnology 27, 946 - 950 (2009)

Retrieved from "http://2011.igem.org/Team:Warsaw/ExpressionAdaptors/Solution"