Directly to content
  1. Publishing |
  2. Search |
  3. Browse |
  4. Recent items rss |
  5. Open Access |
  6. Jur. Issues |
  7. DeutschClear Cookie - decide language by browser settings

Application of motif scoring algorithms for enhancer prediction in distantly related species

Dolle, Dirk-Dominik

PDF, English (Dissertation) - main document
Download (34MB) | Terms of use

Citation of documents: Please do not cite the URL that is displayed in your browser location input, instead use the DOI, URN or the persistent URL below, as we can guarantee their long-time accessibility.


Although many studies proposed methods for the identification of enhancers, reliable prediction on a genome-wide scale is still an unsolved problem. One of the reasons for this is the highly flexible regulatory logic underlying a detectable enhancer activity. In each cell type or tissue and at any given time, a mostly unknown set of transcription factors activates specific regulatory elements by coordinated binding to the corresponding genomic region. Position, spacing, and orientation of the individual bound factors can thereby vary between different enhancers yet result in a highly similar spatio-temporal activity. Due to this inner flexibility, so-called “alignment-free” methods have been proposed for enhancer prediction, as they are able to cope with rearrangements by comparison of word profiles rather than linear sequence. However, the problems caused by allowing for permutation in sequence comparison have not been investigated so far. In this study I implemented several published alignment-free metrics and analysed, which parameters affect their ability to successfully predict regulatory regions. As results show, single point mutations and the increasing amount of spurious matches with decreasing word size pose the biggest challenge to alignment-free techniques, especially when applied on a genome-wide scale. Alignment algorithms usually solve these problems quite efficiently but cannot handle permutation. I therefore implemented a new technique for enhancer prediction that combines the advantages of both algorithm types and used it for the identification of regulatory regions in the teleost fish Oryzias latipes (Medaka) based on a set of known and validated human enhancers. Predicted medaka regions and human enhancers were subsequently used in an in vivo enhancer assay and analysed for their activity. In total, 12 predicted regions corresponding to 9 human enhancers showed clear enhancing activity in the fish. This shows that the principle implemented here is able to predict active enhancers at a high rate on a genome-wide scale even in species as diverged as human and fish. Furthermore, evidence for motif-level conservation between some of the human and medaka enhancers could be found that was invisible for most of the alignment-algorithms used for comparison.

Item Type: Dissertation
Supervisor: Wittbrodt, Prof. Dr. Jochen
Date of thesis defense: 9 November 2012
Date Deposited: 19 Nov 2012 07:55
Date: 13 November 2012
Faculties / Institutes: Service facilities > Centre for Organismal Studies Heidelberg (COS)
Subjects: 004 Data processing Computer science
500 Natural sciences and mathematics
570 Life sciences
590 Zoological sciences
Uncontrolled Keywords: enhancer, prediction, gene regulation
About | FAQ | Contact | Imprint |
OA-LogoDINI certificate 2013Logo der Open-Archives-Initiative