Most of our knowledge of the genetics of gene expression has been derived through studies conducted in cell lines and normal primary tissues. Understanding the germline contribution to gene expression levels in tumors is confounded by the acquisition of complex somatic and epigenetic alterations, as well as a dearth of datasets measuring this information on a common set of samples. Although eQTL studies using tumors have been reported, the effects of somatic genetic and epigenetic factors have not been addressed (Grisanzio et al., 2012; Kristensen et al., 2006; Pomerantz et al., 2010). Our method provides a practical solution to account for the multiple factors that determine gene expression levels. Moreover, the method is easily expandable to accommodate other factors, such as somatically acquired single nucleotide variants, that may influence gene expression and can be applied to the growing availability of genome wide tumor datasets.
