Biomarker discovery: identification of a growth factor gene signature

Gene expression signatures can be developed as comprehensive pathway readouts and used as pharmacodynamic or patient-stratification biomarkers. While a consensus on the best practices for selecting gene expression signatures from microarray data is evolving, we have developed basic guidelines to ensure consistency and quality. Here we illustrate these guidelines through the identification of a growth factor gene expression signature that is responsive to phosphatidylinositol 3-kinase (PI3K) pathway perturbations in vitro and related to phosphatase and tensin homolog (PTEN) deregulation in vivo.