Reverse-Engineering Gene Regulatory Networks from Microarray Data

Date of Award

2004

Document Type

Thesis

Degree Name

Bachelors

Department

Natural Sciences

First Advisor

Clore, Amy

Keywords

Microarray, Gene Network, Reverse-Engineering

Area of Concentration

Biology

Abstract

Though microarrays were developed nearly a decade ago, they have only begun to fulfill their potential. Advances in reproducibility and error control are slowly bringing a longtime goal � extracting causal regulatory information from gene expression data � within reach. Major computational hurdles currently prevent the application of mathematically precise methods to this task, but a wide-ranging collection of new and adapted systems from computational modeling are already achieving notable successes. The next task is to compare objectively the effectiveness of these methods in forming regulatory models from various types of data. This study aims to judge, partly by comparing false- negative and false-positive error rates in predicting known datasets, the relative efficacy of methods ranging from the well-established, technical differential equation models, to the relatively informal, ad-hoc event and edge detection methods. These algorithms are applied to synthetic gene expression profiles and yeast datasets in the Stanford Microarray Database (http://genorne-www.stanford.edu/microarray/) and Gene Expression Omnibus (http://www.ncbi.nlm.nih-gov/geo/) to define this effectiveness across various applications. In accord with previous studies, several methods retrieved synthetic gene connections quite well, but the percentage of known yeast transcriptional regulators recovered from real-world data remained at chance levels.

Rights

This bibliographic record is available under the Creative Commons CC0 public domain dedication. The New College of Florida, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.

This document is currently not available here.

Share

COinS