Figure 2

Quantile-quantile plots and density plots for HGU95 Spike-In data. (a) Quantile-quantile plot with quantiles of the standard normal distribution on the horizontal axis and quantiles of the noise distribution as estimated by the exponential-normal convolution model. If the normality assumption is correct, the plotted values should lie on a straight line. (b) Quantiles of an exponential distribution versus the background corrected probe-level intensities from the exponential-normal model. Again, any departures from a straight line indicate a lack of fit for the exponential distribution. (c) Density estimates of the log base 2 PM intensities for the original (uncorrected) probe-level intensities (solid line) and the estimated background using the exponential-normal model (dashed line). (d) Density estimates of the log base 2 PM intensities from the original data versus a simulated convolution of a normal distribution and an exponential distribution. The parameters for the normal and exponential distributions were obtained using estimates given by the Bioconductor implementation of RMA.