Fluctuation theory as a basis for experimental design

25 March 2011

Johan Paulsson
Department of Systems Biology
Harvard Medical School


Mathematics traditionally affects experimental biology in two ways. Dynamic models quantify mechanistic assumptions and make testable predictions about system properties, while statistics is used to analyze the conclusiveness of the results. In individual cells, statistical patterns also arise spontaneously because chemical dynamics is inherently probabilistic. This in principle introduces more testable outputs, such as variances or distributions rather than just averages, but at the same time creates an explosion in the number of free parameters. In this talk I will show how seemingly straightforward experimental strategies systematically fail to separate between completely different assumptions, but also how stochastic theory could be rigorously used to greatly facilitate experimental design, particularly in complex or poorly characterized systems.

current theory lunch schedule