| Summary: | A well-established approach for detecting genes involved in tumorigenesis due to copy number alterations (CNAs) is to assess the recurrence of the alteration across multiple samples. Expression data can be used to filter this list of candidates by assessing whether the gene expression significantly differs between tumors depending on the copy number status. A drawback of this approach is that it may fail to detect low-recurrent drivers. Furthermore, this analysis does not provide information about expression changes for each gene as compared to the whole data set and does not take into consideration the expression of normal samples. Here we describe a novel method (Oncodrive-CIS) aimed at ranking genes according to the expression impact caused by the CNAs. The rationale of Oncodrive-CIS is based on the hypothesis that genes involved in cancer due to copy number changes are more biased towards misregulation than are bystanders. Moreover, to gain insight into the expression changes caused by gene dosage, the expression of samples with CNAs is compared to that of tumor samples with diploid genotype and also to that of normal samples. Oncodrive-CIS demonstrated better performance in detecting putative associations between copy-number and expression in simulated data sets as compared to other methods aimed to this purpose, and picked up genes likely to be related with tumorigenesis when applied to real cancer samples. In summary, Oncodrive-CIS provides a statistical framework to evaluate the in cis effect of CNAs that may be useful to elucidate the role of these aberrations in driving oncogenesis. An implementation of this method and the corresponding user guide are freely available at http://bg.upf.edu/oncodrivecis.
|
|---|
