Heat maps showing two measures for factor binding specificity in blastoderm embryos. (a, b) Each row shows the GSC z score of the likelihood that the overlap between a cohort of ChIP/chip peaks for one factor and regions bound each factor in turn occurs by chance. (c, d) Each row shows the Pearson correlation coefficients between the intensity scores of a cohort of 300 peaks for a factor and the intensity scores of the equivalent 500-bp windows at the same genomic locations for each of the other factors in turn. (a, c) Results for the most highly bound 300 peaks (1-300). (b, d) Results for the second most highly bound cohort of 300 peaks (301-600). Note that the 1% FDR threshold does not lie within ranks 1 to 600 for 17 of the 21 factors (Table 2), and, thus, for these proteins the differences observed between the 1-300 and the 301-600 cohorts are not attributable to false positives.