Finding common aberrations in array CGH data

Array comparative genomic hybridization (aCGH) technology provides high-resolution measurements of DNA aberrations at tens of thousands of locations distributed throughout the genome. These genomic aberrations cause alterations in gene expression patterns which, in turn, are a common cause for emergence of cancer. However, like all other microarray technologies, the obtained measurement data is noisy. In addition to the measurement noise, the heterogeneity of biological samples and cancer cells add biological noise and it also needs to be taken into account. To infer reliable results, analysis of aCGH data requires that different sources of uncertainty are carefully considered. We present an analysis framework that can be used to find reliable estimates of genomic aberrations that are frequent throughout a set of aCGH data. These commonly aberrant segments of DNA and the genes that reside in them, are key factors in understanding cancer. We demonstrate our framework by applying it to a set of aCGH data obtained from two different types of cancer. We also investigate what biological processes are affected by the mutations uncovered by our analysis of these cancer types using gene ontology enrichment. Based on this enrichment analysis, our framework reliably finds common aberrations in aCGH data.