Lung cancer is the leading cause of cancer-related mortality for men and women in the United States, accounting for 28% of all cancer deaths in 2004. Non-small cell lung cancer (NSCLC) is the most common histologic type of lung cancer. A detailed knowledge of the spectrum of mutations driving NSCLC is crucial for the development of rational based therapeutics and diagnostics. However, NSCLC cells are notoriously aneuploid and showed extensive genomic instability resulting in numerous recurrent cytogenetic aberrations that are thought to play important roles in tumorigenesis. Molecular analyses of lung adenocarcinoma have begun to identify a small number of well-known lesions such as activated oncogenes (K-ras, c-myc and c-erbB2/neu) and tumor suppressor genes (p53, p16 and Rb). Based on the numerous recurrent alterations observed by conventional low-resolution cytogenetic analyses, we believe that NSCLC cells harbor many additional, undiscovered lung cancer-relevant lesions. The comprehensive high-resolution genome-wide search for recurrent regional alterations, and ultimately, the identification of the targets of such alterations will have an enormous impact on our ability to understand the process of lung tumorigenesis, to generate preclinical models, and to develop sensitive diagnostics and effective therapies. To discover these lung cancer genes, we propose to perform a systemic genome-wide analysis for small recurrent regional loss and gains that would identify loci harboring candidate lung cancer relevant genes in 30 human NSCLC cells lines and 70 primary human NSCLC using the high density oligonucleotide array comparative genomic hybridization (CGH) platform. After a highly selective bioinformatic filtering process, a limited selection of candidate lung cancer relevant genes will be assayed for their cancer–relevant activities in vitro in a series of cell-based assays.