< Back to Article

Functional Characterization of Transcription Factor Motifs Using Cross-species Comparison across Large Evolutionary Distances

Figure 1

Computational pipeline for charting a motif function map.

(A) Each motif is scored against each gene's promoter (“motif scanning”). The top scoring target genes of a motif (“motif module”) are analyzed for enrichment for GO gene sets using the Hypergeometric test, and statistically significant motif – GO associations (red cells) from the test constitute a “motif function map”. (B) Different motif scanning methods produce different motif function maps by the process in (A). For each motif, the best motif scanning method (score) is selected by evaluating each motif function map based on the number of associations and a suitable control (see Methods). (C) For each motif, redundant GO associations are identified by using an extended Hypergeometric test (see Methods) and the motif function map is reorganized. This panel shows GO associations of the Fushi tarazu (FTZ) motif, with redundant associations being indented. The “cond-pval” column is the conditional p-value of an association given the stronger association it is redundant with (see Methods). For example, the association with “sensory perception of smell” is highly significant (p-value∼6E-4), but is “statistically explained” by the association with “odorant binding” (conditional p-value∼1); the Venn diagram on the right illustrates why this is the case.

Figure 1