Please use this identifier to cite or link to this item:

Integrative Transcriptomic Analysis of Long Intergenic Non-Coding RNAs in Cancer.

File Size Format  
2017-05-phd-ching.pdf 41.83 MB Adobe PDF View/Open

Item Summary

Title:Integrative Transcriptomic Analysis of Long Intergenic Non-Coding RNAs in Cancer.
Authors:Ching, Travers H.-m.
Contributors:Molecular Biosciences & Bioeng (department)
Date Issued:May 2017
Publisher:University of Hawaiʻi at Mānoa
Abstract:Long non-coding RNAs (lncRNA) are a relatively new and mysterious class of RNA molecules
that are transcribed in eukaryotic cells. They are di erentiated from mRNA transcripts in that
they do not code for proteins and are much larger than small RNA species, such as microRNAs.
Long intergenic non-coding RNAs (lincRNAs) are a subclass of lncRNAs that appear outside
the boundaries of known genes. At the current state, little is de nitively known about lincRNAs.
LincRNAs have a diverse range of functions, such as providing molecular sca olding for
chromatin remodeling, acting as molecular sponges for microRNAs, or directly interacting with
promoter and enhancer regions to promote or downregulate gene expression.
In humans, there are a huge number of lincRNA genes, more than the number of genes that are
protein coding; it has been estimated that 80% of the human genome is transcribed, yet only
2-3% is translated. There is active debate in the eld as to what proportion of those transcripts
are biologically relevant, as the alternative is that some of those transcripts are meaningless
noise, due to leaky RNA polymerases.
There is an increasing number of lincRNAs that are known to be functionally relevant to cancer
such as XIST, MALAT1, HOTAIR and PCAT1. XIST generally acts to silence one copy of the
X-chromosome in women; in breast cancer, it is found to be downregulated. HOTAIR, within
the HOX locus, is deregulated in aggressive metastatic tumors. HOTAIR expression is increased
in metastatic cancer and is a biomarker for poor prognosis. MALAT1 is enriched in the nucleus,
regulates cell motility and is also implicated in metastasis. PCAT1 is implicated in disease
progression in prostate cancer. However, the functions and mechanisms of most lincRNAs are
not de nitively known.
This dissertation focuses on elucidating the roles of lincRNAs in relation to cancer pathogenesis.
The focus is on identifying lincRNA biomarkers in cancer and to further elucidate clinically
relevant lincRNA mutations. Using bioinformatics and computational biology approaches to
analyze lincRNA expression and mutation pro les, I will attempt to determine which lincRNAs
are relevant to tumorigenesis and progression and how mutation data correlates with expression
and clinical phenotypes.
Information generated from this investigation will provide knowledge on the role of non-coding
RNAs in the development and progression of cancer. It will also help to elucidate the application
of machine learning methods to cancer and non-coding gene research domains. Most importantly,
it will push forward the translational and clinical applications of lincRNAs as potential
cancer biomarkers and therapeutic targets.
In chapter 1, I further explain the technical background relevant to the projects contained in this
dissertation. Chapter 2 is a lincRNA review paper published in BioData Mining, focusing on
the upcoming computational challenges related to lincRNA research. Chapter 3 is an analysis of
RNA-Seq di erential expression methods published in RNA; computational approaches in order
to nd upregulated or downregulated lincRNAs. Chapter 4 is an exploration of the expression
landscape of lincRNA across 12 cancer types, published in eBioMedicine. Chapter 5 and 6 are
applications of machine learning methods to high dimensional biological data. In Chapter 5,
I explore a neural network-cox regression machine learning hybrid model, in order to predict
patient survival, and to elucidate the biological pathways relevant to each patient. In chapter
6, I elucidate the somatic mutation landscape in lincRNAs across 12 cancer types. I quantify
which molecular features are correlated with lincRNA mutation probability, and I show that
these results could be used to provide more robust subtyping and clustering of tumor samples.
Finally, in Chapter 7, I discuss what these research projects have accomplished in the grand
scheme of the lincRNA research eld, and explain what further work needs to be accomplished
to follow up.
Description:Ph.D. Thesis. University of Hawaiʻi at Mānoa 2017.
Rights:All UHM dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.
Appears in Collections: Ph.D. - Molecular Biosciences and Bioengineering

Please email if you need this content in ADA-compliant format.

Items in ScholarSpace are protected by copyright, with all rights reserved, unless otherwise indicated.