Sequence features of DNA binding sites reveal structural class of associated transcription factor

• DOI: 10.1093/bioinformatics/bti731
• Publication Date: 2005-11-03T01:13:48Z
• Authors: Leelavati Narlikar; Alexander J. Hart...
• Keywords: DNA binding site

Abstract Motivation: A key goal in molecular biology is to understand the mechanisms by which a cell regulates transcription of its genes. One important aspect this transcriptional regulation binding factors (TFs) their specific cis-regulatory counterparts on DNA. TFs recognize and bind DNA according structure DNA-binding domains (e.g. zinc finger, leucine zipper, homeodomain). The these can be used as basis for grouping into classes. Although varies widely across generally, within particular class similar fashion, suggesting existence class-specific features sequences bound each TFs. Results: In paper, we apply sparse Bayesian learning algorithm identify small set different classes TFs; simultaneously learns true multi-class classifier that uses predict domain TF recognizes sequences. We train our six largest TRANSFAC, comprising total 587 learn six-class training achieves 87% leave-one-out cross-validation accuracy. also are highly TF, has significant implications how sites should modeled purpose motif discovery. Contact: [e-mail]; [e-mail]