Title: "Information Processing in Biochemical Networks"

Peter Thomas

Salk Institute

Abstract:

From the early days of information theory, biologists have borrowed its ideas for the analysis of communication between living cells, particularly within networks of nerve cells in the brain. Applications have included Barlow's theory of redundancy reduction (whitening) for transmission of visual information through the optic nerve, sparse representations of statistically typical visual stimuli via information maximization, and investigation of "the nature of the neuronal code" i.e whether sensory information in the brain is carried only in the firing rates of neurons or also in the timing of individual action potentials.

Nerve cells are but one class of cells that utilize chemical signaling to receive and relay sensory information. From the immune system to the regulation of tissue growth, biological cells rely on complex networks of biochemical interactions to extract information about their environments and determine responses to chemical messages. We have begun the systematic analysis of an elementary biochemical signal-transduction network viewed as a communications channel. Two fundamental processes limit the transmission of information via chemical means. Chemical signals typically reach their targets via diffusion. Once in the vicinity of a cell equipped to receive it, a signaling molecule binds to a receptor protein spanning the cell membrane. By repeatedly binding and unbinding to signaling molecules the receptor can signal to the cell interior the local concentration of signaling molecule. Together, diffusion and the stochastic receptor-binding process attenuate high-frequency signals, nonlinearly limit high-amplitude signals, and introduce intrinsic non-Gaussian channel noise. We describe the challenges in analysing this system as a novel kind of nonlinear, nonGaussian communications channel, and indicate some of our progress to date.