Our study design anticipates configuring all the diverse classes of data described to address the biological and clinical challenges posed by lung cancer. We call our approach integrative to emphasize the inclusion of behavior (e.g., nicotine dependency, smoking intensity, depression, anxiety, and other psychological traits) and outcome (e.g., survival from lung cancer) with the traditional molecular epidemiology framework which uses biomarkers to elucidate the biological relationships between exposure, genes, and diseases. Moreover, the integrative approach allows cross-sectional analyses of multiple factors (e.g., germline genetic variation, somatic mutations and gene expression in relation to lung cancer risk or progression) [21]. This approach provides several key advantages over more fragmented designs: 1) It is highly efficient and cost effective, since information collected for one purpose can be leveraged for another, instead of each goal requiring independent planning, infrastructure and data collection; 2) Because the design includes diverse study domains, diverse questions can be addressed that are inaccessible to more constricted designs. For example, depression (a behavior) is known to be related to smoking (an exposure). However, only an integrated design can establish
