Skip to main content
Figure 1 | BMC Bioinformatics

Figure 1

From: Improved ChIP-chip analysis by a mixture model approach

Figure 1

GC-dependent normalization of one sample. Scatter plots of log intensities of Cy3 and Cy5 signals (from array GSM254806) based on the number GC base pairs of each 50-mer probe: 15 (a), 20 (b) or 30 (c). Density plots of raw data (d), MA2C (robust, C = 2) normalized data (e) and Lowess normalized data (f). Three curves are overlaid on figures (a)–(c). The blue line depicts the baseline model of MA2C normalization. The red line is fitted by median regression and the yellow line is the Lowess fit. In figures (d)–(f), vertical lines indicate mode and median of all probes. In raw and MA2C normalized data, the mode is bigger than median (d, e), indicating a heavier tail on the left. This unexpected feature usually indicates a problematic array or insufficient normalization.

Back to article page