Difference between revisions of "Team:Manchester/Model/Analyse"

 
(10 intermediate revisions by 2 users not shown)
Line 101: Line 101:
 
</style>
 
</style>
  
−
<h1 class="title11">Result Analysis</h1>
+
<h1 class="title11">Analyse the Results</h1>
  
 
<div class="team">
 
<div class="team">
  
−
<h1 class="bigtitle">Making ensemble outputs</h1>
+
<!---------------------------------------------------beer's---------------------------------------------------->
  
−
<p> Contents </br>
 
−
<a href="#heading1">Probability Outputs</a></br>
 
−
<a href="#heading2">What can be concluded?</a></br>
 
−
<a href="#heading3">Human Practice Link: Police Usage</a></br>
 
−
<a href="#heading4">Units and Beer's law?</a></br>
 
−
 
−
</p<
 
−
<p id="heading1" class="title2" style="font-size:25px;">Probability Outputs</p>
 
  
 +
<h1 id="heading4" style="margin-top:90px;" class="bigtitle">Units and Beer’s law</h1>
 
<p style="font-size:17px;">
 
<p style="font-size:17px;">
−
From all your data you can output predictions for specific observables for your system. These outputs will be a pdf, more simply a histogram will do. The observable we used was concentration for a time close to steady state. We also produced a ensemble plot of concentration vs time for oxidised abts.     
 
−
</p>
 
  
−
<p id="heading2" class="title2" style="font-size:25px;">What can concluded from these things</p>
+
The experimental team used a plate reader, which measured the changes of absorbance over time in optical density values. To convert these into concentrations of ABTS, for comparison with our model predictions, we used Beer’s law and a calibration experiment. </br></br> This allows the experimental results to be directly compared with the model predictions and hence <a href="https://2016.igem.org/Team:Manchester/Model/result"> further analysis</a> to be undertaken.
  
−
<p style="font-size:17px;">
 
−
From these graphs questions about your system can be answered. In our case does GDL matter and how does ABTS concentration matter. This can be done by comparing the results from the lab for these observables with model predictions. For example for the GDL question we would compare the outputs of uni-bi bi-uni with revmm - bi uni etc. In this case if the histogram for revmm - bi uni was far from the observed values and uni-bi bu uni was close to the observed values you could conclude GDL does make a difference. See results for this analysis.
 
 
</p>
 
</p>
  
−
<p id="heading3" class="title2" style="font-size:25px;">Human practice link: Police usage</p>
+
<p class="title2" style="font-size:25px;">Calibration Experiment</p>
−
 
+
−
 
+
−
 
+
 
<p style="font-size:17px;">
 
<p style="font-size:17px;">
−
The cost of the patch can be estimated from the data, setting a maximum time for expression , minimum expression amount and assuming the cost of a patch material is negligible so the total cost is dependant only on the enzyme costs. The cost to produce a patch for all enzyme ratios can be generated. This is then comparable with the current cost of systems like breathalysers (HP - Police interview) to determine if the alcopatch would be relevant to specific industries like the police
+
The calibration experiment was done with ABTS, H2O2, and HRP. There was an excess of H2O2, so that the endpoint of the reaction was such that we could assume all the ABTS was oxidised. We did this experiment for a range of ABTS concentrations. </br></br>
−
</p>
+
There were some extra complications because ABTS decays and the other reagents cause absorbance; these issues were taken into account initially, but the influence on the results was negligible so this assumption could be relaxed. This analysis allowed us to convert OD values to concentrations in μg ml<sup>-1</sup> ABTS, using Beer’s law, which states that absorbance measured in OD values at a given path length is proportional to the concentration of the absorbing compound (dye). Using the known molecular weight of ABTS, this could then be converted to concentrations in mM, which could be compared to our model predictions.
−
 
+
</p> </br>
−
<p class="title2" style="font-size:25px;">How code works</p>
+
<img src="https://static.igem.org/mediawiki/2016/f/f8/T--Manchester--modelling_graph_8.png"></img>
−
 
+
<p> </p>
−
 
+
−
<p style="font-size:17px;">
+
−
<a href="https://github.com/Manchester-iGem-2016/UoMiGem2016">Link to github homepage</a>  
+
−
<br /><br />
+
−
The codes for these outputs are short and understandable, also there is no guarantee your relevant outputs will be the same. So the files are only included for reference.  
+
−
</p>
+
−
 
+
−
 
+
−
 
+
−
<!---------------------------------------------------beer's---------------------------------------------------->
+
−
 
+
−
 
+
−
<h1 id="heading4" style="margin-top:90px;" class="bigtitle">Units and beer’s law</h1>
+
−
<p class="title2" style="font-size:25px;">Why units must be same?</p>
+
−
<p style="font-size:17px;">
+
−
In the lab we could produce concentration vs time graphs, these were simulated using our model.
+
−
We required the units for both to be the same so we could compare.
+
−
</p>
+
−
 
+
−
<p class="title2" style="font-size:25px;">How we did this for the model?</p>
+
−
<p style="font-size:17px;">
+
−
For our model our concentrations should be measured in units like mM i.e a measure of the amount of compounds per volume.
+
−
The model units come from the units used in the initial conditions, we used mM by having the concentrations in ug/ml from the lab and dividing by the molecular weight.
+
−
 
+
−
</p>
+
  
 
<p class="title2" style="font-size:25px;">How we did this for the experiments?</p>
 
<p class="title2" style="font-size:25px;">How we did this for the experiments?</p>
Line 170: Line 131:
  
  
−
<p class="title2" style="font-size:25px;">Calibration experiment</p>
+
<p class="title2" style="font-size:25px;">Our OD-to-concentration converter software
−
<p style="font-size:17px;">
+
−
The calibration experiment was done with Abts h202 and hrp there was excess H202 so that the equilibrium was such that we could assume all the abts was oxidised. We did this for a range of initial conditions. We now knew the od values at steady state and the concentrations at steady state. And hence we could calibrate.
+
−
<br /><br />
+
−
There were some extra complications because abts decays and the other reagents cause absorbance, these were taken into account initially but the change was negligible so it was ignored, this assumption could be relaxed in the future. This analysis allowed us to convert od values to ug/ml and hence mM.
+
−
 
+
 
</p>
 
</p>
−
 
−
 
−
 
−
<p class="title2" style="font-size:25px;">Theory of beer's law</p>
 
 
<p style="font-size:17px;">
 
<p style="font-size:17px;">
−
Beer’s law simply states that concentration in a unit like mM is proportional to absorbance measured in od value, a result from fluid mechanics leads to this phenomena specifically with the od scale.
+
The code for these calculations is available on our [Github page]. Absorbance values in OD are read into MATLAB, together with the concentrations of oxidised ABTS from the calibration experiment. A linear regression line is then fit to the data. Using the equation of the resulting calibration line, it is now possible to convert absorbance values (OD) to concentrations (μg ml<sup>-1</sup>.) Dividing the result by the molecular weight of the compound (g mol<sup>-1</sup>) results in the concentration (μmol ml<sup>-1</sup> = mM).
 
</p>
 
</p>
−
 
−
<center>
 
−
<img class="width50" src="https://static.igem.org/mediawiki/2016/f/f8/T--Manchester--modelling_graph_8.png" alt="graph 8" />
 
−
</center>
 
−
 
−
<p style="font-size:17px;">The graph is for the described experiment .The dotted line uses the gradient and intercept from fitting. With these you can easily convert between the two.
 
−
 
−
</p>
 
−
 
−
 
−
<p class="title2" style="font-size:25px;">How code works and Git</p>
 
−
 
−
 
−
<p style="font-size:17px;">
 
−
<a href="https://github.com/Manchester-iGem-2016/UoMiGem2016">Link to github homepage</a>
 
−
<br /><br />
 
−
This code is a simple one. Absorbance values in od are read into matlab as well as the concentrations of abts oxidised from the calibration experiment. A straight line fit is then made. Now from this straight line equation given one variable e.g. absorbance values (od) you can calculate the concentration in (ug/ml.) You could divide by mr to get the answer in mM.   
 
−
 
−
</p>
 
−
 
  
  

Latest revision as of 22:28, 19 October 2016

Manchester iGEM 2016

Analyse the Results

Units and Beer’s law

The experimental team used a plate reader, which measured the changes of absorbance over time in optical density values. To convert these into concentrations of ABTS, for comparison with our model predictions, we used Beer’s law and a calibration experiment.

This allows the experimental results to be directly compared with the model predictions and hence further analysis to be undertaken.

Calibration Experiment

The calibration experiment was done with ABTS, H2O2, and HRP. There was an excess of H2O2, so that the endpoint of the reaction was such that we could assume all the ABTS was oxidised. We did this experiment for a range of ABTS concentrations.

There were some extra complications because ABTS decays and the other reagents cause absorbance; these issues were taken into account initially, but the influence on the results was negligible so this assumption could be relaxed. This analysis allowed us to convert OD values to concentrations in μg ml-1 ABTS, using Beer’s law, which states that absorbance measured in OD values at a given path length is proportional to the concentration of the absorbing compound (dye). Using the known molecular weight of ABTS, this could then be converted to concentrations in mM, which could be compared to our model predictions.


How we did this for the experiments?

The experimental team used a plate reader. The units showing concentration over time were optical density values, these can be converted into a concentration using beers law and a calibration experiment.

Our OD-to-concentration converter software

The code for these calculations is available on our [Github page]. Absorbance values in OD are read into MATLAB, together with the concentrations of oxidised ABTS from the calibration experiment. A linear regression line is then fit to the data. Using the equation of the resulting calibration line, it is now possible to convert absorbance values (OD) to concentrations (μg ml-1.) Dividing the result by the molecular weight of the compound (g mol-1) results in the concentration (μmol ml-1 = mM).

Sponsors

Follow us on