Figure 3

Results of the proposed spatial segmentation method (seg) on a bladder cancer array. Bladder cancer array with local spatial bias accurately detected by the proposed normalization method. (a) Median-centered log-ratios; (b) spatial trend; (c) spatial segmentation; (d) local spatial bias. The border of areas affected by local spatial bias that have been detected in panel (d) are reported on panels (a), (b) and (c) as a black step-function for easy interpretation; (e) genomic profile without spatial normalization (spots detected as local spatial artifacts are marked in red, and the vertical gray dashed lines indicate the separation between chromosomes).