Figure 3
From: Estimation and correction of non-specific binding in a large-scale spike-in experiment
Histograms of P values for all null probesets. (a) The expected distribution of P values for null probesets is a uniform distribution between 0 and 1, generated at random. The observed P value distribution after normalization using all (b) null probesets as a subset and (c) only FC = 1 probesets as a subset are shown. MAS5 PM-MM was used for NSB correction, probes were normalized with the loess method, probes were summarized into probesets with medianpolish, and P values were generated with Cyber-T.