User:J. C. MartinezGarcia/Notebook/HMS Activities/2008/11/13
The Sontag´s help  Main project page Previous entry Next entry 
Semi white box modelThe last days I worked in the analysis of the secondorder model proposed by Eduardo Sontag (Eduardo's Sontag website). He is now in a sabbatical year at MIT (working in the Electrical Engineering Department). The model that he proposed to us, and that I implemented in the computer, is the following one:
where x corresponds to the variable which dynamics drives the biestable behavior and y corresponds to the variable involved directly with the expression of the AKT protein. For the function f(x) there exist several possibilities. The first one (implemented in the computer) is a cubic function: f(x) = Vmax x(t)/(Km+xx(t))  l x(t), i.e a MichaelisMenten function with degradation of x(t). I took, because of certain conditions involving the behavior of s(x):=a/(K+k*yx(t)), the following values for the parameters:
For the input u(t), this corresponds to the Zstk drug. As far as k is concerned, this is a variable associated to the cut of the feedback acting from the second equation to the first one. In the first version of the computer implementation I chose k to be a logical variable, being just equal to zero when no feedback is considering in this equation. It must be pointed out that the second equation ensures that:
The question now is how to justify in biological terms the chosen functions. In particular, the choice of f(x(t)) must be justified. Eduardo Sontag is proposing a mechanism not based in cooperativity. I will consider this question now. Some words on robustnessI printed some papers by Eduardo SOntag considering robustness measures. These are the following ones:
I will read them later and I will cultivate the contact con Sontag (that is important to improve the impact of my research in systems biology from a theoretical point of view).
