MSc & MSc Funded Theses
Permanent URI for this collection
Browse
Browsing MSc & MSc Funded Theses by Course name "Applied Bioinformatics"
Now showing 1 - 3 of 3
Results Per Page
Sort Options
Item Open Access Classification of endocrine resistant breast cancers from transcriptomic datasets using multi-gene signatures(Cranfield University, 2012-09) Larionov, Alexey; Cameron, David; Morgan, SarahBreast cancer is the most frequent cancer in women in developed countries. Endocrine treatment is indicated to the majority of breast cancer patients. However, in some cases it does not work despite the current clinical indications. Eventually the resistance may develop in many of those who initially respond. Re-analysis of available breast cancer transcriptomic datasets using new multi-gene signatures associated with endocrine resistance may help to understand and overcome endocrine resistance. The goal of this project was to develop a bioinformatics pipeline to (i) select endocrine resistant cases from the available breast cancer datasets and (ii) classify the selected cases by multiple multi-gene signatures. The pipeline has been successfully designed and applied for classification of endocrineresistant samples from 9 breast cancer datasets using 7 transcriptional signatures. The obtained results have been presented in a dedicated web site. The pipeline consists of: Procedures for a manually curated selection of relevant datasets and signatures; Procedures for semi-automatic data pre-processing, allowing cross-platform analysis; A new, fully automated, classification algorithm (Iterative Consensus PAM). The main features of the developed classification algorithm include: It is based on un-supervised partitioning; It allows for “non-classifiable” samples; The procedure does not require a training set; The procedure can be used in a cross-platform context (Affymetrix & Illumina). The developed pipeline and web site may constitute a prototype for a future web-hub collecting (i) data on endocrine-resistant breast cancer specimens, (ii) collecting multigene signatures relevant to endocrine resistance and (iii) providing tools to apply the signatures to the data. The web-repository could provide a tool to integrate the data and signatures and to produce new clinical and biological knowledge about endocrine resistance in breast cancer.Item Open Access Creation of a software tool for browsing genome variation(Cranfield University, 2011-10) Horner, Neil; Larcombe, Lee D; Clark, TaaneThe advent of next generation sequencing has led to an explosion of the amount of DNA sequences in public databases. A challenge is now to find tools that are able to make it easier for researchers to browse and make sense of this data. One organism that has recently been subject to extensive sequencing is Plasmodium falciparum, a devastating pathogen that infects hundreds of millions of people annually. The first goal of this project was to create a new desktop genome variation browser that can quickly handle large amounts of data from sequencing projects involving numerous isolates. The second aim was to use the new tool to analyse recently-sequenced strains of P. falciaprum in order to identify polymorphisms that may be involved in antibiotic resistance. The variation browser described here was written in C++ and the Qt graphical framework in order to make an easy to use and fast tool that can visualise data from variant call format (VCF) files, which is now a de facto standard for storing polymorphism data. The user is able to browse a VCF file to gain a graphical representation of the variation among multiple samples. For rapid identification of relevant polymorphisms, the user is able to filter variant positions using several criteria including mapping quality, sample group membership, and whether the mutations alter the amino acid sequence of a gene. Some basic statistical analysis was incorporated to help identify selective pressures acting on polymorphic sites. The usefulness of the program was ascertained by analysing 75 isolates of P. falciparum from Africa and Asia. Mutations were identified in the chloroquine resistance marker protein, PI4-K, and a putative ubiquitin carboxyl hydrolase, which are potentially involved in antibiotic resistance.Item Open Access A systems biology approach to target discovery in regulatory T cells(Cranfield University, 2011-08) Weston, Marie C.; Cauchi, Michael; Page, MattRegulatory T cells (Tregs) have a central role in the maintenance of tolerance to self- antigens and the prevention of autoimmune disease. This study used an integrative systems biology approach to identify tolerogenic genes in Tregs which could potentially serve as novel therapeutic targets for immunological disorders. A consensus Treg gene signature was generated by comparing gene expression in Treg vs naïve or conventional T cells across multiple public studies. Ingenuity Pathway Analysis software was then used to expand the Treg consensus gene list to include interacting proteins accessible to intervention by antibody therapeutics. Many viruses co-opt genes for host proteins that modulate the host’s immune system. It is hypothesized that some viruses may have co-opted genes that can induce tolerance, allowing the virus to evade elimination by the host’s immune system. Putative tolerogenic genes were therefore selected for further investigation based upon their presence in viral genomes. The presence of human genes in viral genomes was investigated by performing a batch reciprocal BLAST search. The biological significance of the human vs viral alignments was evaluated by manual inspection of the alignments and searching for the presence of shared motifs and protein family domains in the viral and human sequences. A final list of ten putative tolerogenic genes included genes known to be associated with immune function and some already established therapeutic targets for autoimmune diseases, as well as four potentially novel therapeutic targets. The biological rationale for the putative targets’ involvement in tolerance was explored in the context of Treg gene expression and protein-protein interaction (PPI) network topology. A PPI network was generated and annotated with confidence scores for each of the interactions. The Cytoscape plugin JActiveModules was used to find putative functional network modules.