DescriptionUnderstanding individualized breast cancer treatment options can help physicians care for their patients by careful selection of personalized therapies. The first steps towards this goal have already been taken by clinicians, with the frequent use of molecular and genetic biomarkers to classify breast cancer into categories which direct treatment. This thesis will propose new therapeutic targets for different breast cancer subtypes, as well as a new set of biomarkers that more efficiently predict hormone resistance in estrogen positive (ER+) breast tumors. A novel methodology for therapeutic target prediction will be proposed, based on a new paradigm called “gene centrality”. In addition to being over-expressed, good therapeutic targets should have a high degree of connectivity in the tumor network. Gene centrality encompasses this concept by measuring the connectivity of genes in a network in which each edge is weighted by the level of over-expression of the target gene. Using this method, a series of high centrality SRC proto-oncogenes (LYN, YES1, HCK, FYN, and LCK) were identified in subsets of Basal-like and HER2+ breast cancers. The hypothesis that YES1 is a therapeutic target in breast cancer was experimentally tested. We found that Basal-like breast tumor cell lines showed a significant decrease in fitness upon silencing the expression of YES1. Another validated therapeutic target in breast cancer is the estrogen receptor ESR1, targeted by drugs such as Tamoxifen. However, a significant fraction (~30%) of ER+ cases doesn’t respond well to this therapy. A novel outlier analysis method was applied to gene expression data from ER+ breast cancer patients to identify genes highly associated with Tamoxifen resistance. These included cell cycle genes as well as several chromosomal amplification sites. In addition to the well known HER2 amplicon on 17q12, we discovered that amplicons in 8q24.3, 8p11.2 and 17q21.33-q25.1 correlate strongly with early distant metastasis and poor long term survival. As independent biomarkers for Tamoxifen resistance, together these chromosomal regions are predictive for ~75% of patients that suffer early disease relapse.