Figure 6.

Examples of applications of the theoretical results. Two real motifs of width W = 15 from the JASPAR database (ZFP423 and TATA-box, TBP) are used to illustrate the application of the theoretical predictions. For a strong motif like Zfp423, increasing the number of sequences, n, from 20 to 30 (arrow 1) or reducing the sequence length from 1000 to 100 (arrow 2) could sufficiently reduce the bound so that the real motif is found in a region where the expected number of false positives is less than 1. For a weak motif like TBP, detection in sequences of length 100 might still be prone to false positives, so instead a large increase in the number of sequences, n, (arrow 3) is needed.

Zia and Moses BMC Bioinformatics 2012 13:151   doi:10.1186/1471-2105-13-151
Download authors' original image