Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


tez070213.final.pdf (11.03 MB)

ETD Abstract Container

Abstract Header

METABOLIC NETWORK-BASED ANALYSES OF OMICS DATA

Cicek, A. Ercument
http://rave.ohiolink.edu/etdc/view?acc_num=case1372866879

Abstract Details

2013, Doctor of Philosophy, Case Western Reserve University, EECS - Computer and Information Sciences.
Omics studies have shed light to human metabolism for decades, with the advancement of high throughput screening technologies. Each domain provides traces of information that may lead to solving mechanisms of complex diseases. However the number of candidate hypotheses is usually exponential with respect to the number of variables. Thus, there is a need for algorithms that explore the search space efficiently while reasoning in a biologically sound way. Biological networks have been used as a roadmap to investigate the interplays between the variables. As they provide apriori knowledge on the biological problem at hand, they enable algorithms to prune the search space and consider only related variables together. In this thesis, we describe algorithms and systems that use metabolic networks in order to analyze omics data. First, we address the problem of predicting changes in the metabolite levels based on a condition, such as a genetic perturbation. We propose ADEMA, An Algorithm to Determine Expected Metabolite Level Changes Using Mutual Information. ADEMA detects related metabolites in the metabolic network, and then using mutual information, it calculates the expected metabolite level changes based on the given condition. Next, we briefly describe SMDA Algorithm (Steady-state Metabolic Network Dynamics Analysis Algorithm). The algorithm starts by classifying each measurement as low, normal or high based on the concentrations published in the literature. However, usually there is no consensus on what is normal and what is not, which leads algorithm to end up with an empty result set. We propose semi-automated approaches to locate and fix misclassifications due to such discrepancies. Then, we describe MIRA, Mutual Information-based Reporter Algorithm. MIRA’s goal is to find the metabolites around which transcriptional regulation is centered, based on the transcriptome and the metabolic network of the organism. MIRA is an improvement over the current z-score based reporter algorithm, as it considers expressions of multiple genes at a time. Finally, we describe ReconModels.org, which is an online database of published genome-scale reconstructed metabolic networks of the organisms with browsing, querying, visualizing, edit and comparing metabolic networks.
Gultekin Ozsoyoglu (Advisor)
Mehmet Koyuturk (Committee Member)
Mitchell Drumm (Committee Member)
Jing Li (Committee Member)
Z. Meral Ozsoyoglu (Committee Member)
Ivan Iossifov (Committee Member)
178 p.
Bioinformatics; Computer Science
Metabolomics; Transcriptomics; Metabolic Network; Gene Expression; Knowledge Discovery

Recommended Citations

Citations

  • Cicek, A. E. (2013). METABOLIC NETWORK-BASED ANALYSES OF OMICS DATA [Doctoral dissertation, Case Western Reserve University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=case1372866879

    APA Style (7th edition)

  • Cicek, A. Ercument . METABOLIC NETWORK-BASED ANALYSES OF OMICS DATA. 2013. Case Western Reserve University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=case1372866879.

    MLA Style (8th edition)

  • Cicek, A. Ercument . "METABOLIC NETWORK-BASED ANALYSES OF OMICS DATA." Doctoral dissertation, Case Western Reserve University, 2013. http://rave.ohiolink.edu/etdc/view?acc_num=case1372866879

    Chicago Manual of Style (17th edition)

case1372866879
1,718
© 2013, all rights reserved.
This open access ETD is published by Case Western Reserve University School of Graduate Studies and OhioLINK.