Rule mining on microRNA expression profiles for human disease understanding

Field	Value	Language
dc.contributor.author	Song, Renhua
dc.date.accessioned	2016-09-21T00:47:03Z
dc.date.available	2016-09-21T00:47:03Z
dc.date.issued	2016
dc.identifier.uri	http://hdl.handle.net/10453/52935
dc.description	University of Technology Sydney. Faculty of Engineering and Information Technology.	en_AU
dc.description.abstract	This research employs rule mining methods to study the important roles of miRNAs in human diseases. From past experience and from reviewing the literature, rule mining is a widely used data mining technique for the discovery of interesting relationships in large data sets. MicroRNAs (miRNAs) are endogenous and highly conserved non-coding RNA molecules. They can inhibit and/or promote the post-transcriptional expression of target messenger RNAs (mRNAs). miRNAs thus play a pivotal role in a cell’s differentiation, proliferation, growth, mobility, and apoptosis, as well as in viral replication and proliferation. This has inspired many research works aimed at detecting miRNAs’ functions in human disease. However, with the current deluge of miRNA data, previous works have suffered from limitations in terms of handling the relationship between various molecules. Firstly, they usually identify single miRNAs as biomarkers, and always produce low sensitivity and specificity. Secondly, intensive research largely depends on the inverse expression relationships between miRNAs and mRNAs to discover miRNA-mRNA regulatory modules. Finally, the miRNA-miRNA co-regulations and miRNA self-regulations have not been well investigated. As a result, rule mining is a powerful new technology with great potential to help researchers focus on the most important miRNAs for understanding human diseases. This thesis reports our past and current research outcomes in this area. The contributions of the thesis are as follows: • A novel rule mining method is proposed to detect the significant miRNA biomarkers. • A “change to change” method is proposed to mine both positive and negative regulatory relationships from paired miRNA and mRNA expression data sets. • A progressive data refining approach is proposed to identify the lung cancer miRNA-miRNA co-regulation network. • A novel framework is proposed to detect the self-regulation miRNAs. The research was conducted through four case studies. (1) The first case study was on lung squamous cell carcinoma for accurate diagnosis of this disease through the reliable miRNA biomarkers identified by a novel rule discovery method. (2) The second case study was on paired miRNA and mRNA expression data of HCV patients to detect both positive and negative regulatory modules. (3) The third case study was on lung cancer data sets for the computational methods to identify miRNA-miRNA co-regulation networks and miRNA-miRNA co-regulatory relationships. (4) The fourth case study was on multiple data types to infer self-regulation miRNAs in humans through an integrative rule mining framework and approach. All the results have been verified by the existing literature and databases.	en_AU
dc.format	Thesis (PhD)
dc.language.iso	en_AU	en_AU
dc.relation	https://opus.lib.uts.edu.au/bitstream/10453/52935/2/02whole.pdf
dc.rights	info:eu-repo/semantics/openAccess
dc.rights	au.edu.uts.lib/ppc
dc.rights	The author owns the copyright in this thesis including all reproduction and reuse rights for the work. The work may not be altered without the permission of the copyright owner. Attribution is essential when quoting or paraphrasing from this thesis.
dc.subject	RNA molecules.	en
dc.subject	MicroRNAs (miRNAs) biomarkers.	en
dc.subject	Data mining technique.	en
dc.subject	Rule mining methods.	en
dc.subject	Self-regulation miRNAs.	en
dc.title	Rule mining on microRNA expression profiles for human disease understanding	en_AU
dc.type	Thesis	en_AU
utslib.copyright.status	open_access

Abstract:

This research employs rule mining methods to study the important roles of miRNAs in human diseases. From past experience and from reviewing the literature, rule mining is a widely used data mining technique for the discovery of interesting relationships in large data sets. MicroRNAs (miRNAs) are endogenous and highly conserved non-coding RNA molecules. They can inhibit and/or promote the post-transcriptional expression of target messenger RNAs (mRNAs). miRNAs thus play a pivotal role in a cell’s differentiation, proliferation, growth, mobility, and apoptosis, as well as in viral replication and proliferation. This has inspired many research works aimed at detecting miRNAs’ functions in human disease. However, with the current deluge of miRNA data, previous works have suffered from limitations in terms of handling the relationship between various molecules. Firstly, they usually identify single miRNAs as biomarkers, and always produce low sensitivity and specificity. Secondly, intensive research largely depends on the inverse expression relationships between miRNAs and mRNAs to discover miRNA-mRNA regulatory modules. Finally, the miRNA-miRNA co-regulations and miRNA self-regulations have not been well investigated. As a result, rule mining is a powerful new technology with great potential to help researchers focus on the most important miRNAs for understanding human diseases. This thesis reports our past and current research outcomes in this area. The contributions of the thesis are as follows: • A novel rule mining method is proposed to detect the significant miRNA biomarkers. • A “change to change” method is proposed to mine both positive and negative regulatory relationships from paired miRNA and mRNA expression data sets. • A progressive data refining approach is proposed to identify the lung cancer miRNA-miRNA co-regulation network. • A novel framework is proposed to detect the self-regulation miRNAs. The research was conducted through four case studies. (1) The first case study was on lung squamous cell carcinoma for accurate diagnosis of this disease through the reliable miRNA biomarkers identified by a novel rule discovery method. (2) The second case study was on paired miRNA and mRNA expression data of HCV patients to detect both positive and negative regulatory modules. (3) The third case study was on lung cancer data sets for the computational methods to identify miRNA-miRNA co-regulation networks and miRNA-miRNA co-regulatory relationships. (4) The fourth case study was on multiple data types to infer self-regulation miRNAs in humans through an integrative rule mining framework and approach. All the results have been verified by the existing literature and databases.

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

http://hdl.handle.net/10453/52935