Figure 5

Representation of Z-distribution. The empirical distribution of the absolute value of the log-ratio of logarithms of intensities (Z) of all the signals recorded from the analyzed arrays is represented. On the x-axis the value of the variable Z is plotted; on the y-axis the value of the density function of the variable Z probability is reported. The Z-distribution appears as a bimodal density with two peaks, and it can be decomposed into two normal distributions: one with positive mean (see the peak on the right), representing the distribution of those SNPs with very different intensity values for the two alleles (homozygous genotypes), and the other with mean zero (see the peak on the left), representing the distribution of those SNPs with similar intensity values for both the alleles (heterozygous genotypes). The bold line (with filled triangles) represents the best fitting to a mixture of normal distributions obtained through an E-M algorithm, while the dashed line (with open squares) represents the real data.