Interactive visualization and model-based analysis of genomics data.

摘要

Biotechnologies, such as DNA sequencing and microarray technology, have transformed research in molecular biology from being "gene-centric" to "genome-centric", opening new avenues in the areas of drug discovery, clinical diagnostics and agriculture. There is a pressing need for development of new computational and visualization techniques to gain biological knowledge from massive amounts of heterogeneous and complex genomics data. Computational methods transform biological questions into mathematical problems to produce large quantities of numerical results that require biological interpretation. Visual presentation of such data combined with interactive exploration will allow biologists to comprehend underlying biology. Motivated by these challenges of bridging the gaps between data processing by computational scientists and its interpretation by life scientists, this dissertation presents visualization tools for genomics data with easy-to-use interfaces.;Understanding the process of gene regulation is one of the most important and challenging questions for which the scientific community is increasingly looking for an answer in genomics data. Different types of genomics data shed light on different aspects of gene regulation making integration of data essential to get a handle on the whole process of gene regulation. To facilitate such integration, this dissertation presents a model of gene regulation that leads to a graph theoretic structure, which provides an invariant view of regulation from both the sequence and gene expression. Methods to obtain approximations to such a structure from gene expression data and DNA-protein interaction data are presented.