A Strategy for Identifying Putative Causes of Gene Expression Variation in Human Cancer

There is often a need to predict the impact of alterations in one variable on another variable. This is especially the case in cancer research, where much effort has been made to carry out large-scale gene expression screening by microarray techniques. However, the causes of this variability from one cancer to another and from one gene to another often remain unknown. In this study we present a systematic procedure for finding genes whose expression is altered by an intrinsic or extrinsic explanatory phenomenon. The procedure has three stages: preprocessing, data integration and statistical analysis. We tested and verified the utility of this approach in a study, where expression and copy number of 13,824 genes were determined in 14 breast cancer samples. The expression of 270 genes could be explained by the variability of gene copy number. These genes may represent an important set of primary, genetically “damaged” genes that drive cancer progression