Figure 1.
A window, in our terminology, is a possible binding site for a TF; in the case of phylogenetically unrelated sequences it is simply a set of m contiguous bases in a sequence, with m the binding site width. This figure shows a configuration C containing a total of eight windows (rectangles) for three different WMs (red, blue, and green). Note that a single sequence of length L has L − m + 1 windows in it.
Figure 2.
An Alignment of Four Sequences Showing Three Legitimate Windows and One Illegitimate Window
Vertically aligned capital letters are phylogenetically related bases, assumed to have evolved from a common ancestor. Thus, any window placed on these bases is extended to cover all related bases. Three legitimate windows are surrounded by solid boxes. The window surrounded by the dotted box is illegitimate because the gap in the top sequence makes the alignment of bases inconsistent. Note that lower case letters are not aligned and that, in order to complete a window with aligned sequences, one may slide lowercase bases “through” adjoining gaps. For example, if the window on the bottom two sequences were to move two steps to the left, the “c” and “a” on the left side of the preceding gaps would slide through the gaps to the right to complete the window.
Figure 3.
Performance of PhyloGibbs and Non-Phylo Motif-Finding Algorithms on Alignments of Orthologous Intergenic Regions as a Function of the Evolutionary Proximity of the Orthologs and the Quality of the WM
PhyloGibbs with phylogeny (red), PhyloGibbs in non-phylo mode (light blue), WGibbs (dark blue), and MEME (pink) were run on alignments of S = 5 intergenic regions of length L = 500, each at a proximity q to the common ancestor and each containing s = 4 binding sites from a single WM of width w = 10. In the upper left panel, WMs had polarization p = 0.6, in the upper right p = 0.75, in the lower left p = 0.9, and in the lower right random WMs (drawn uniformly from the simplex) were used. The solid lines show the average overlaps between the predicted sites and the real sites, and the dotted lines show two standard errors (estimated from 50 different datasets generated with equal parameters for each data point).
Figure 4.
Performance of PhyloGibbs in Recovering a Single Site of a Randomly Chosen WM of Width w = 10 from the Alignment of S Orthologous Intergenic Regions of Proximity q = 0.5 and Length L = 500 as a function of S
The solid line shows the average overlap between the true site and the predicted site and the dotted lines show two standard errors.
Figure 5.
Performance of Several Motif-Finding Algorithms on Synthetic Data Prepared as for Figure 3
A total of 250 alignments of S = 5 orthologous intergenic regions of length L = 750 and proximity q = 0.5 were created with three binding sites sampled from each of three different random WMs. The left panel shows how the fraction of predicted sites that match true sites (specificity) depends on the fraction of true sites that are among the predictions (sensitivity) for PhyloGibbs (red), EMnEM (yellow), PhyME (green), PhyloGibbs without phylogeny (light blue), WGibbs (dark blue), and MEME (pink). Dashed lines correspond to two standard errors. The right panel shows the ability of the different algorithms to assess their own reliability. The true specificity is shown as a function of the specificity that the algorithm predicts for the sites that it reports. The black line y = x corresponds to a perfect assessment of the algorithm's reliability.
Figure 6.
Performance of Several Motif-Finding Algorithms on 200 Alignments of Orthologous Intergenic Regions from Five Saccharomyces Species Containing Documented Binding Sites
The left panel shows how the fraction of predicted sites that match true sites (specificity) depends on the fraction of true sites that are among the predictions (sensitivity) for PhyloGibbs (red), EMnEM (yellow), PhyME (green), PhyloGibbs without phylogeny (light blue), WGibbs (dark blue), and MEME (pink). Dashed lines correspond to one standard error. In order for the specificities, predicted by the various algorithms, to match the true specificities, we have to assume that the known sites are only a fraction of all true sites. The right panel shows what the fraction of known sites among all true sites should be in order for the algorithms' predicted specificities to match the true specificities. The black line shows an independent estimate of the fraction of real sites in these upstream regions that is documented (see text).
Table 1.
Results of PhyloGibbs on Collections of Intergenic Regions for 21 TFs for Which the Motif-Finding Algorithms in [27] Failed to Recover a Significant Motif but for Which a Literature Consensus Motif Is Available
Table 2.
Results of PhyloGibbs on Collections of Intergenic Regions for 24 TFs for Which the Motif-Finding Algorithms in [27] Found a Significant Motif
Table 3.
Results of PhyloGibbs on Multiple Alignments of Upstream Regions Taken from the Literature