School of BioSciences - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemExploring the cancer transcriptome with novel bioinformatics approaches( 2022)Currently three out of every 10 deaths within Australia will be a direct consequence of cancer. Cancer is a complex and genetically heterogeneous disease that is, as a consequence, effectively unique to each individual. However, there are common driving events, phenotypes, and risks that can segregate cancer within tumour types and subtypes. These groupings are beneficial as they can both inform treatment regimes and yield new targets for pharmaceutical development. Next Generation Sequencing (NGS) of RNA has enabled measurement of the abundance and makeup of a sample’s transcriptome, which through bioinformatics analysis, can reveal the rich interplay between genetic mutations and their functional and phenotypic consequences. This thesis focuses on three key transcriptome projects. The first project developed the ALLSorts software which is the first publicly available and open-source classifier for determining subtypes of B-Cell Acute Lymphoblastic Leukemia (B-ALL). The purpose of this tool is to provide researchers with an accurate method for using transcriptome data to quickly label B-ALL samples according to 18 subtypes. Subtyping is becoming part of clinical standard-of-care, informing targeted pharmaceutical treatment and/or treatment intensity. The second project, Slinker, is a publicly available and open source visualisation tool that can be applied to any gene that highlights splicing variation between a case and controls. Novel splicing is regularly observed across a variety of diseases, including cancer, and can lead to a significant alteration of the final transcript, possibly transforming it into a pathogenic driver. Slinker is novel in that it utilises the superTranscritome method to create succinct visualisations by removing redundant features. The final project in this thesis is an analysis of the utility of long read transcripts as a transcriptomic reference, specifically within a spatial context. Three references were compared: the hg38 reference transcriptome, the long reads themselves as a reference, and both combined. Each had gene expression quantified through highly accurate, short read technology. The combined reference resulted in both a higher mapping rate and novel expressed sequences, of which one belongs to a gene that is a known prognostic marker for the oropharyngeal head and neck cancers that this method was applied to.