Figure 5.6: Smooth density estimates for the M values of the six arrays in the CCl4sn data. For each array, the distribution is centered at 0, but two of the arrays stand out for having a much wider distribution than the other four. Because all six arrays have been hybridized with the same pair of biological samples, this difference in width of the M distributions may indicate variability in the data quality. Further investigation is necessary to make a more definitive conclusion.