Skip to main content
. 2011 Jul 23;39(18):e126. doi: 10.1093/nar/gkr574

Figure 3.

Figure 3.

An illustration of how the STEME branch-and-bound algorithm works. Top: the current estimate of the motif in the EM algorithm. This is actually the motif for Stat5 from the TRANSFAC database (M00223). Bottom: part of the suffix tree representing the sequences. We can see that if we have descended the tree to the node that represents the prefix, GCAT, our match to the motif is poor. If the bound for the 〈Zn〉 of all the nodes below this is small enough, we can stop our descent here.