Figure 7
From: Autocorrelation analysis reveals widespread spatial biases in microarray experiments
Identification of aneuploidy by autocorrelation. (A) Gene expression comparisons of strains with verified differences in chromosome copy number result in continuous stretches of high autocorrelation values, due to the similarity in expression measurements over long genomic intervals. Data shown is from strains that contain aneuploidies (lanes 1-22) or segmental duplication of 56 or 28 genes (lanes 23-24, respectively) [32, 33]. (B) CGH experiments are frequently associated with similar autocorrelation patterns. The datasets are those described in Dunham et al., 2002 [46], and Dunn et al., 2005 [47]. Note that periodic autocorrelations are also observed in several of these experiments. (C) In specific microarray designs, such as the SMD design, spatial biases of certain shapes can generate "spurious aneuploidies". Note that the autocorrelation is shown for different gene distances than in (A-B).