Figure 7.

Frequency of occurrence of motifs near transcription start sites (TSS's). For each of the five motifs, we compared the frequency of the motif in randomly-selected 9 gene samples (1 gene from each promoter subfamily) occurring within 250 bp of putative TSS's (black) versus random 500-bp windows in gene blocks (grey). A total of one thousand 9-gene samples were taken to produce the distribution (see Methods). The frequency of random occurrence in gene blocks for Motif 1 is higher (and less resolved from the test set) because Motif1 is especially AT-rich, and thus is expected to occur more frequently by chance in gene blocks, which are also AT-rich. The five motifs are present near TSS's in an average of 7.1 genes per 9-gene cross-subfamily sample; background frequency of the five motifs within gene blocks average 2.8 genes per 9-gene cross-subfamily sample.