FootPrinter3: phylogenetic footprinting in partially alignable sequences

Abstract

FootPrinter3 is a web server for predicting transcription factor binding sites by using phylogenetic footprinting. Until now, phylogenetic footprinting approaches have been based either on multiple alignment analysis (e.g. PhyloVista, PhastCons), or on motif-discovery algorithms (e.g. FootPrinter2). FootPrinter3 integrates these two approaches, making use of local multiple sequence alignment blocks when those are available and reliable, but also allowing finding motifs in unalignable regions. The result is a set of predictions that joins the advantages of alignment-based methods (good specificity) to those of motif-based methods (good sensitivity, even in the presence of highly diverged species). FootPrinter3 is thus a tool of choice to exploit the wealth of vertebrate genomes being sequenced, as it allows taking full advantage of the sequences of highly diverged species (e.g. chicken, zebrafish), as well as those of more closely related species (e.g. mammals). The FootPrinter3 web server is available at: http://www.mcb.mcgill.ca/~blanchem/FootPrinter3.

Figure 1

Output of the FootPrinter3 web server when run on a set of vertebrate promoter regions of the FOS gene. Each orthologous sequence extends 1 kb upstream of the start codon (in most species, the 5′-untranslated region 5′-UTR extends over ∼150 bp). Full length sequence for chicken was not available. Alignment blocks are shown in thin colored lines, while the motifs found appear as colored bars. Arrows indicate the position of experimentally verified transcription factor binding sites from Transfac. Below the graphical output are part of the set of sequences used as input, with the motifs highlighted in the same color as in the graphical display. The list of all motifs found, with the position of each of their substrings and the score statistics are shown on the right, again with a consistent color scheme. The parameters used for the run were: motif size: 8 bp; Number of mutations allowed: 2; Maximum number of mutations per branch: 1; Use TBA alignment: ‘Yes’; Allow regulatory element losses: ‘Yes’; Spanned tree significance level: ‘Significant’; Motif loss cost: 1; The running time of the query was less than 2 min. The set of input sequences and the output files are available as example from the web server.