Quantifying stability in gene list ranking across microarray derived clinical biomarkers

Our method projects genomic tumor expression data to a lower dimensional space representing the main variation in the data. Some information regarding the phenotype resides in this low dimensional space, while some information resides in the residuum. We then introduce an information ratio (IR) as a metric defined by the partition between projected and residual space. Upon grouping phenotypes such as tumor tissue, histological grades, relapse, or aging, we show that higher IR values correlated with phenotypes that yield less robust biomarkers whereas lower IR values showed higher transferability across studies. Our results indicate that the IR is correlated with predictive accuracy. When tested across different published datasets, the IR can identify information-rich data characterizing clinical phenotypes and stable biomarkers.The IR presents a quantitative metric to estimate the information content of gene expression data with respect to particular phenotypes.The challenge to identify stable tumor prognosis and predictive outcome markers remains critical in clinical cancer research. Many studies rely on microarrays to determine which genes are predominantly indicative of clinical cancer phenotypes or prognosis. However, biological and technical variations across samples and studies make it challenging to identify true, predictive clinical biomarkers [1,2]. Identification of stable gene expression signatures can facilitate the classification of clinical phenotypes and their associated physiological states. Histologic tumor grade, ER (estrogen receptor) status and predicted risk of relapse are among the currently used labels to distinguish prognosis and treatment regimes. Our motivation in this study was to determine when stable predictive biomarkers can be identified from multiple microarray studies or meta-analyses.Results from microarray experiments can be arranged as an n by p matrix with n being the number of samples and p the number of measured features or pro