2009 Math Biology Seminar - 06

Cellular decisions rely upon a cell making some measurement of its surroundings and then regulating its behaviour based on this measurement. For many cellular processes this decision is regulated by the transcriptional output of a gene which is regulated by a single input which may exist in several possible states. Given the noise inherent in the input signal and the chemical processes coupling the input to the output, how well can the input states be measured by the genetic network and is the genetic network optimized to maximize the likelihood of determining the correct input state? Recently, this problem has been analyzed in the context of mutual information. I will discuss the application of mutual information to the problem of morphogen readout in developing organisms, and will compare the predicted optimal morphogen gradient for the early drosophila factor Bcd to its experimental profile. Lastly, I will show how optimizing mutual information for an organism that needs to infer a two state environment which varies, leads to an optimal transcriptional response with biologically realistic kinetic parameters.