Home Team Project Mathematical Modelling Logbook Result Sponsors

The Result

Here are some results from our model.

correct input
  • det adj N V det adj N stop
  • det N V adj N stop (see first figure)
incorrect input
  • det adj N V N V
  • N V det adj adj (see second figure)

MATLAB Simulation

Based on our model design described in Mathematical Modeling page; we simulate our network with giving both correct and incorrect order of inputs which corresponds for correct and incorrect sentence. First the correct sentence is tested with following order: “det, N, V, adj, N, Stop” all the protein levels for states and intervals represent good response in the simulated network and finally the correct protein level shows that the input sentence was right.

Cor det N V adj N.png

Then for liability test, incorrect sentence was given with following order: “N, V, det, adj, adj” as seen before all protein levels are working fine before we reach the point which model detects an error in the order of inputs and automatically goes to incorrect state. By that point no more input is given as model is going back to initial state. This is also the case for the correct sentence.

Incor N V det adj adj.png

The reset system (Not shown) put the model to its original initial interval and ready for the new input. This is implemented in model in order to make it more stable and interactive as well as possible to be more realistic while implemented in the real cell.

The Suggested Plasmid

In order to carry out such a model we need to design input and internal plasmids. Designing input plasmid requires 5 exogenous activators which are labeled as X 1…X5 in the model. Metal ion seems to be good candidates as well as pathways such as Methionine biosynthesis and Galactose degradation pathways. Let’s consider Iron ion as the first reagent .Induction of iron leads to regulation of fnr gene which acts as positive regulatory element for FdrA.


This is the input plasmid involving one state which can be seen as S1 in graphical representation of our linguistic model. It should be able to inhibit 9 other genes selected for other 9 states. In order to implement internal plasmid based on our model each gene need to be inhibited by specific repressors for example, in case of fdrA the inhibitor Narl which needs specific binding site.


Along with this, the gene should have the flexibility to be released from inhibition meaning that there must be a way to remove the repressor. Proteolysis could be a possible example. As we reach the last input reagent and it regulates the final input plasmid it should ultimately activate one of the 2 final GFP or RFP tagged genes, based on the semantically correct or incorrect sentence we pursue followed by the specific pathway. Implementation of such a genetic network needs extensive research upon existing genetic pathways in E.coli that would complement our mathematical model. But due to time constrain it was not possible to actually design such a synthetic network. Thus, it shall be our future work!!