This project will develop novel statistical methodology to help address the following three questions:
(i) What are the signalling pathways and regulatory modules (transcription factors and miRNAs) disrupted in a given cancer type?
(ii) How does this pattern of alteration relate to existing molecular classifications?
(iii) Can we use this inferred pattern of pathway and regulatory module alteration to provide a clinically more relevant classification of cancer?

The aim is to apply novel methods using ideas from network theory to multi-dimensional cancer genomic data (mRNA expression, SNP, miRNA, copy-number, mutational spectra, DNA methylation) data to help address the following problems in cancer genomics:
(i) Identification of the genomic aberrations that are most likely to be drivers of the carcinogenic process and that thus play a key role in characterising the cancer phenotype spectrum.
(ii) To explore the statistical properties of cancer gene networks and investigate the relationship of these network properties with cancer phenotypes.

This project is geared towards analysing mRNA expression data to
(i) investigate the similarities between the molecular classifications of different cancers (breast, prostate and lung cancer).
(ii) to identify novel prognostic subclasses and molecular markers within these clinically heterogeneous cancers.

This project is aimed at developing novel statistical methodology for the analysis of mRNA expression data to help elucidate the regulatory modules and signalling pathways that
(i) differentiate stem cell like and multipotent progenitor cell populations from their commited differentiated counterparts.
(ii) differentiate normal stem cells from transformed cancer stem cells.

The aim is to develop novel statistical methods for the analysis of large scale DNA methylation data (Illumina Infinium, MeDIP-seq) in cancer to more robustly identify regions undergoing cancer-specific hyper/hypo methylation and that may serve as diagnostic or prognostic biomarkers.