Skip to main content


Figure 1 | BMC Genomics

Figure 1

From: Autocorrelation analysis reveals widespread spatial biases in microarray experiments

Figure 1

Periodic autocorrelation dependent on microarray design. (A) Comparison of the periodic autocorrelation in 4 related experiments. The SMD mciroarray is from Spellman et al., 1998 [27], and the Affymetrix YE6100 is from Cho et al., 1998 [24]. Labeled cDNA prepared form the same RNA source of cells traversing the cell cycle was hybridized onto UHN and UMC microarrays. Despite the similar, or identical, underlying biology, the autocorrelation periods are different and reminiscent of each microarray design. Thus, the characteristic period of each microarray directly corresponds to the distances of the probes on the microarray as a function of the distance of the genes in the genome. The latter is shown for the UHN (B), UMC (C), and SMD (D) microarrays alongside the autocorrelation levels for the different gene distances. Probe placement data for the Affymetrix YE6100 microarray was unavailable. In the UMC microarray, the autocorrelation period is 1/4 of the probe placement period. This is due to the precise nature of the spatial bias associated with the specific microarray analyzed (see text for details). Complete correspondence between the two parameters could be observed in other experiments (not shown).

Back to article page