In the post-genome era, high throughput gene expression profiling has been successfully used to develop genomic biomarker panels (GBP) that can be integrated into clinical decision making. The development of GBPs in the context of personalized medicine is a scientifically challenging and resource-intense process. It needs to be accomplished in a systematic phased approach to address biological variation related to a clinical phenotype (e.g. disease etiology, gender, etc.) and minimize technical variation (noise). Here we present the methodological aspects of GBP development based on the experience of the Cardiac Allograft Rejection Gene Expression Observation (CARGO) study, a study that lead to the development of a molecular classifier for rejection screening in heart transplant patients.