New Approaches for the Molecular Profiling of Human Cancers through Omics Data Analysis

In this thesis, we present three studies in which we applied ad hoc computational methods for the molecular profiling of human cancers using omics data. In the first study our main goal was to develop a pipeline of analysis able to detect a wide range of single nucleotide mutations with high validation rates. We combined two standard tools to create the GATK-LODN method, and we applied our pipeline to exome sequencing data of hematological and solid tumors. We created simulated datasets and performed experimental validation to test the pipeline sensitivity and specificity. In the second study we characterized the gene expression profiles of 11 tumor types aiming the discovery of multi-tumor drug targets and new strategies of drug combination and repurposing. We clustered tumors and applied a network-based analysis to integrate gene expression and protein interaction information. We defined three multi-tumor gene signatures, characterized by the following categories: NF-KB signaling, chromosomal instability, ubiquitin-proteasome system, DNA metabolism, and apoptosis. We evaluated the gene signatures based on mutational, pharmacological and clinical evidences. Moreover, we defined new pharmacological strategies validated by in vitro experiments that showed inhibition of cell growth in two tumor cell lines. In the third study we evaluated thyroid gene expression profiles of normal, Papillary Thyroid Carcinoma (PTC) and Anaplastic Thyroid Carcinoma (ATC) samples. The samples grouped in a progressional trend according to tissue type and the main biological processes affected in the normal to PTC transition were related to extracellular matrix and cell morphology; and those affected in the PTC to ATC transition were related to the control of cell cycle. We defined signatures related to each step of tumor progression and mapped the signatures onto protein-protein interaction and transcriptomical regulatory networks to prioritize genes for following experimental validation.