Skip to main content
Figure 4 | BMC Bioinformatics

Figure 4

From: Quantized correlation coefficient for measuring reproducibility of ChIP-chip data

Figure 4

Scatterplots before and after quantization. Quantization and merging identifies background signals and suppresses their effect on the reproducibility measure, as shown in the case of MSL complex binding in Drosophila. (A) A scatterplot between replicates shows that the data quality of the replicates is fairly good, as almost all signal probes are from the X chromosome (red) rather than an autosome (blue) for the complex that is known to have X-specific binding [13]. The standard PCC at the probe level is 0.38, as the background signal makes a substantial contribution to the statistic. (B) The final configuration (with jittering to view data points that would otherwise be overlapping) of the quantization and merging procedure with 100 as the initial number of bins. The cut-offs are the quantiles to which the iteration converges. Almost all the probes from the autosome are grouped into the background as expected, and the correlation of the ranked data at the probe level increases to 0.65.

Back to article page