Recently, with the increasing application of high-resolution CNV detection methods, a genome-wide spatial autocorrelation or ‘wave’ pattern in signal intensity data was described that interferes with accurate CNVdetection (4). We use the term ‘genomic wave’ to refer to these patterns of signal intensities across all chromosomes, where different samples may show highly variable magnitude of waviness. This phenomenon has been observed before (5), but the first formal description appeared recently for CNV analysis using an array-CGH platform (4). Marioni et al. (4) described the presence of genomic waves in their Whole-Genome Tiling Path arrays used for CNV detection, and demonstrated that the wavy patterns they observed appeared to be a ‘general feature of aCGH data sets’. They developed a method based on Lowess regression to ‘break’ the waves and improve CNV calling. Furthermore, Komura et al. (6) also described the wavy patterns in signal intensities of Affymetrix arrays, and they reduce the signal noise by incorporating probe and target sequence characteristics in the Genomic Imbalance Map (GIM) algorithm. Nannya et al. (7) has also described similar phenomenon in CNV studies
