channel quantization framework for multiple antenna broadcast scheduling

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2010.Cataloged from PDF version of thesis.Includes bibliographical references (p. 137-143).This thesis presents the development of a new microfluidic method for separating and characterizing cells based upon differences in their electrical properties. The objective of this work is to obtain a genome-wide mapping of genotype to electrical phenotype in the budding yeast Saccharomyces cerevisiae. Towards this end, we present: (1) the development of a novel equilibrium gradient separation method, called isodielectric separation (IDS); (2) the development of physical theories describing how interactions between particles effect microscale separations; and (3) the application of IDS to a screen for electrical phenotypes in the yeast deletion library. Despite the variety of technologies available for cell sorting, a relative lack of intrinsic separation methods - those which separate cells according to their natural, unmodified characteristics - persists. To address this need, we have developed isodielectric separation. IDS separates cells according to differences in their intrinsic electrical properties. Using dielectrophoresis in a medium with spatially varying electrical conductivity, IDS drives cells to the locations where their polarization charges vanish, spatially resolving cells with different electrical properties. Our implementation of IDS offers label-free, continuous-flow separation, and is capable of resolving graded differences in electrical properties. Additionally, we demonstrate the ability to extract quantitative information from samples during separation, establishing IDS as an analytical technique as well as a preparative one. Any platform for performing genetic screens must have high throughput. Although satisfying this requirement would be greatly facilitated by using high cell densities, physical interactions between cells under these conditions can affect the performance of devices used for screening. Although pervasive, interactions between cells or particles are challenging to describe quantitatively, especially in the confined environments typical of microfluidic devices. By studying the effects of electrostatic and hydrodynamic interactions between particles in a microfluidic device, we have found that ensembles of interacting particles exhibit emergent behaviors that we are able to predict through numerical simulations and a simple analytic model based on hydrodynamic coupling. Applying these theoretical tools to IDS and other microfluidic separation methods has provided insight into how particle interactions can profoundly influence separation performance in counterintuitive ways. Having demonstrated the performance metrics necessary for a genetic screen, we apply IDS to the genome-wide analysis of electrical properties in the budding yeast S. cerevisiae. Although others have studied changes in electrical properties induced by drastic changes in gene expression (e.g.in differentiation) or by specific mutations in a small number of genes, a systematic and comprehensive analysis of the relationship between genotype and electrical phenotype has yet to be performed. Using IDS, we have screened the -5000 strains in the yeast deletion collection for altered electrical phenotype. This work has identified a number of genes associated with distinct electrical properties, and, by analyzing known interactions and correlations between these genes, we have identified pathways and morphologies that appear to be primary determinants of electrical phenotype.by Michael D. Vahey.Ph.D