Oral Cancer is a disease related to multiple exposures. In the United States approximately 50000 people are diagnosed with cancer of the head and neck each year. Unfortunately even though there have been some improvements in care, 10000 head and neck cancer patients die each year. Identifying those patients that would respond the best to treatment through quality of care guidelines, genetic signatures, and identifying genetic targets for treatment through machine learning analysis are the focus of this dissertation. The work outlined in this dissertation specifically addresses the differences in survival between patients that are meeting NCCN guidelines for recommendation of chemotherapy and receive recommended therapy and those that do not. This study is a well powered analysis of 37,985 patients pulled from the California cancer registry. It was found that patients have significantly improved survival when there provider prescribes chemotherapy when recommended by NCCN.
In Chapter 2 gene expression signatures are made to predict patient response to treatment. An aggregate signature was identified using a high dimensional dataset with relatively low number of observations (n = 257). By permuting the dataset 100 times via Monte Carlo cross validation and then performing differential expression analysis between treatment responders and non-responders within each permuted dataset, this study was able to identify genes that were differentially expressed across multiple permutations and utilize those gene expression values within a final aggregated signature predicting treatment response.
Chapter 3 utilizes the same gene expression data in a different way by applying machine learning methods known as random forest to rank influential genes and evaluate the pathways within which those genes reside. Integrated with this machine learning analysis is the application of chemical informatics to identify those small molecules in FDA approved drug database and Traditional Chinese Medicine database that meet similarity criteria when measure against a reference ligand known to bind to a drug target site.
|Commitee:||Norden-Krichmar, Trina, Ziogas, Argyrios|
|School:||University of California, Irvine|
|School Location:||United States -- California|
|Source:||DAI-B 80/01(E), Dissertation Abstracts International|
|Subjects:||Genetics, Bioinformatics, Epidemiology, Oncology|
|Keywords:||Bioinformatics, Monte Carlo cross validation, NCCN guidelines, Oral cancer, Random forest, Treatment response|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be