Information in cells
In order to survive and reproduce, cells must sense the changes in their environment and react appropriately. Typically, they will do so by exposing on their membranes a repertoire of receptor proteins that are sensitive to the concentrations of various chemicals, temperature, pH etc. As it reacts with its environment, the structure of the receptor may change and thus initiate a cascade of chemical reactions that transduces the signal into the cell interior. Often, the ultimate response of the cascade will be a change in the concentrations of transcription factors, special proteins that up- or down-regulate expression levels of structural or metabolic genes. For instance, on detecting lactose, a food source for the bacterium E. coli, the cell might start expressing proteins that allow it to digest this particular sugar, while simultaneously directing its flagella to swim up the gradient of nutrients where food is abundant.
Cells are chemical computers: they represent information about the environment and their internal state as concentrations of various signaling compounds, such as transcription factors and activated signaling proteins. Very specific chemical reactions between these species lead to useful computations that can integrate signals from the environment to direct metabolism, to control the division cycle, or to build up a multi-cellular organism. Since proteins are metabolically costly and take time to manufacture, the concentrations of signaling molecules can be very low and subject to small number fluctuations. Cells therefore have to compute reliably in the presence of noise.
We have used information theory to ask: How much information about the environment can cells process and, given a limit to the number of signaling molecules available, how should they process it in the most optimal fashion? Is this theoretical maximum consistent with what we see cells doing in nature?