In some ways, a cell is like a large, noisy, neural net. Molecules, or segments of molecules – genes, transcription factors, and the rest of the cell’s regulatory apparatus – act as nodes in the network. Interactions between molecules can be thought of as connections between the nodes. The propensity of a pair of molecules to interact can be identified with the weight of a connection. The pattern of response of a gene to the binding of a transcription factor seems analogous to the response function of a node. Either thing could in principle be modeled, somewhat imperfectly, with a huge, interconnected system of stochastic differential equations.
This mathematician’s way of looking at cells – a direct intellectual descendant of Stuart Kauffman’s original ‘random Boolean network’ model of gene regulation – makes them seem rather like complex analog computers. And yet many biologists – including Kauffman himself – are skeptical of this further step. Are cells really just ‘processing information’? Are computers really a good model for cells? Isn’t life something rather different from computation? When we make this sort of idealized mathematical model, aren’t we greatly oversimplifying what is, in reality, a very complex physical system? Is the information, as opposed to the molecules we mentally associate it with, really there at all?
Continue reading "Biological Information and Natural Selection" by Daniel Cloud, Ph.D. 