Cancer is fundamentally a genetic disease caused by genomic and transcriptomic abnormalities that deregulate pathways controlling cell proliferation and survival. In fact, the detection of somatic mutations in DNA sequencing data has become a popular approach for the genomic profiling of cancer samples. In the first part of the talk, I will describe how we developed a pipeline to detect a wide range of somatic mutations from DNA sequencing data of cancer samples with high validation rates. We combined two standard tools (GATK and MuTect) to create the GATK-LODN method, which was able to increase the sensitivity and specificity of GATK, while preserving mutations not detected by MuTect. In the second part of the talk, I will show how we investigated the transcriptional landscape of 2,378 cancer samples across 11 tumor types in the search for multi-tumor drug targets. We approached the problem with a network analysis approach in which we combined gene expression profiles with protein-protein interaction and cancer pathway annotations. We found gene signatures characterizing three groups of tumors and we validated their importance through the analysis of mutational data, clinical information, and in vitro experiments. Finally, we proposed a new drug combination for cancer treatment that showed promising results in two different cancer models.