Classifying transcription factor targets and discovering relevant biological features

Dustin T Holloway¹ , Mark Kon² and Charles DeLisi³

¹Molecular Biology Cell Biology and Biochemistry Department, Boston University, 5 Cummington Street, Boston, USA

²Department of Mathematics and Statistics, Boston University, 111 Cummington Street, Boston, USA

³Bioinformatics and Systems Biology, Boston University, 44 Cummington Street, Boston, USA

author email corresponding author email

Biology Direct 2008, 3:22doi:10.1186/1745-6150-3-22


Published:	30 May 2008

Additional files

Additional File 1:

Significant Functions of Highly Regulated Genes. This file is the output from the GO Term Finder at the Saccharomyces Genome database. Using only classifiers which had an accuracy on the training sets of ≥ 0.6 and targets identified with a Platt score P ≥ 0.95, the genes regulated by 12 or more TFs were input into the GO Term Finder. The results show statistically enriched GO terms, p-values, and provide the genes annotated to those terms.

Format: PDF Size: 20KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 2:

Significant Function of Master Regulators. This file is the output from the GO Term Finder at the Saccharomyces Genome database. Using only classifiers which had an accuracy of 0.6 and targets identified at P ≥ 0.95, the regulators which are predicted to bind to 300 or more genes were used as input to the GO Term Finder.

Format: PDF Size: 30KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 3:

Accuracy and Number of Positives. This figure plots the classifier accuracy on the left y-axis(blue), and the number of positives (targets) on the right y-axis(green). Classifiers are numbered on the x-axis and sorted according to increasing accuracy. A loose trend is present, showing that increasing the number of positives increases classifier accuracy. This is mainly seen when 50 or fewer positives exist. Classifiers with 20 or fewer examples tend to do poorly.

Format: PDF Size: 11KB Download file

This file can be viewed with: Adobe Acrobat Reader