Figure 4.

Proportion of functional gene classes in which the test statistic falls within the nominal p < 0.05 tail. Nine publically available gene functional annotation schemas (listed in Table 2), each composed of multiple functional gene classes, were used to determine if specific functional subsets of expressed genes were more likely to show significant change in RNA detection between tissue treatments. Percentage of individual gene classes with test statistics above the nominal p = 0.05 level (F Statistic >3.51 in the dataset of 7853 expressed genes) are plotted on the Y axis for each of 9 classification schema (X axis). Results for functional classes of genes with a minimum of 50 available expressed genes are shown as individual data points (circles) the distribution of which is summarized for each schema by the superimposed notch plot. There are only two high outliers (arrows) amongst the 113 gene classes shown. These are a messenger RNA splicing factors in the Biological Process (GO) schema (Schema 1), and translation factor in the Pathways (GenMAPP) schema (Schema 9). This frequency of 2/113 outliers is no greater than expected by chance, (>5).

Mutter et al. BMC Genomics 2004 5:88   doi:10.1186/1471-2164-5-88
Download authors' original image