Figure 1
From: A novel ensemble learning method for de novo computational identification of DNA binding sites
Flow diagram for SCOPE. BEAM and SPACER are run independently; PRISM runs on the top 100 motifs output by BEAM. For yeast (whose upstream regions are standardized to 800 bp), BEAM and PRISM use the overrepresentation-KS objective function (so/ks), while SPACER's slower running time requires the simpler overrepresentation objective function (so). The top 5 motifs from SPACER are rescored using the combined objective function. For bacteria and Drosophila, upstream regions are defined to be the intergenic region upstream of each gene; thus, the KS objective function is not used. The results of each program are sorted by Sig and lower scoring motifs that substantially overlap higher scoring motifs are removed. The filtered lists of motifs from the three programs are finally merged by Sig score. Repetitive motifs are identified and removed during all stages.