Cell Type-specific Analysis of Human Interactome and Transcriptome

Shahin Mohammadi, Purdue University

Abstract

Cells are the fundamental building block of complex tissues in higher-order organisms. These cells take different forms and shapes to perform a broad range of functions. What makes a cell uniquely eligible to perform a task, however, is not well-understood; neither is the defining characteristic that groups similar cells together to constitute a cell type. Even for known cell types, underlying pathways that mediate cell type-specific functionality are not readily available. These functions, in turn, contribute to cell type-specific susceptibility in various disorders. In this dissertation, I propose a novel measure of similarity between cells and utilize it to identify de novo cell types. I show that my method allows us to uncover novel cancer subtypes. Furthermore, by constructing underlying pathways that drive progression of these subtypes, I show that we can pinpoint diagnostic biomarkers and potential therapeutic targets. Then, I develop a method to dissect the cell type composition of complex tissues. Using this snapshot of what tissues/cell types look like, I create a framework for constructing tissue/cell type-specific interactomes to shed light on the systems-level understanding of cellular functions. I use these networks to uncover brain-specific pathways that are involved in Alzheimer's and Parkinson's diseases. Finally, I provide evidence for the conservation of these interactomes across distant species, even down to unicellular organisms, such as yeast.

Degree

Ph.D.

Advisors

Szpankowski, Purdue University.

Subject Area

Biology|Computer science

COinS