Despite vast difference in their physical scales and constituent components, there are fundamental functional similarities among metabolic, genetic, and neuronal networks. They must measure their environment, process the measured information, make estimations of the quantities of interest, and respond appropriately. This project will study the ability of biological networks to perform information processing tasks efficiently and robustly, in order to identify their common design principles. It will study these principles in the context of biological networks of sufficient complexity to capture information processing qualities, yet of sufficient simplicity to yield generalizable insights.
Specific Aims: 1. To define information theoretic quantities measuring efficiency of a network in processing information in a given environment, and to quantify the functioning of simple, experimentally relevant genetic network topologies in this fashion. 2. To investigate maximizing, extracting, and efficiently representing the predictive information [11] in input data as a biological design principle. 3. To discover, by stochastic simulation of simple genetic network topologies, architectures of biological networks which support multiple functions and therefore are in principle evolvable.
Partnering with (i) Leading Edge Partners, an organization bringing professionals to deliver seminars to New York City (overwhelmingly minority) public high school students and (ii) Columbia University's Science Honors Program, coordinating lectures to New York City k-12 students as well as their families.
