Dokument: Detecting binding sites in PAR-CLIP data using a Bayesian hierarchical model

Titel:Detecting binding sites in PAR-CLIP data using a Bayesian hierarchical model
URL für Lesezeichen:https://docserv.uni-duesseldorf.de/servlets/DocumentServlet?id=57590
URN (NBN):urn:nbn:de:hbz:061-20211004-130723-7
Kollektion:Dissertationen
Sprache:Englisch
Dokumententyp:Wissenschaftliche Abschlussarbeiten » Dissertation
Medientyp:Text
Autor: Hüßler, Eva-Maria [Autor]
Dateien:
[Dateien anzeigen]Adobe PDF
[Details]5,57 MB in einer Datei
[ZIP-Datei erzeugen]
Dateien vom 30.09.2021 / geändert 30.09.2021
Beitragende:Prof. Dr. Schwender, Holger [Gutachter]
Prof. Dr. Rahnenführer, Jörg [Gutachter]
Dewey Dezimal-Klassifikation:500 Naturwissenschaften und Mathematik » 510 Mathematik
Beschreibung:MicroRNAs (miRNAs) play an important role in gene regulation by interacting with messenger RNA (mRNA) sites. To detect these binding sites on the mRNA, biochemical methods such as Photactivatable-Ribonucleoside-Enhanced Crosslinking and Immunoprecipitation (PAR-CLIP) can be conducted. The PAR-CLIP method introduces T-to-C substitutions at sequenced cDNA that help to distinguish between binding site positions in the PAR-CLIP data and noise. T-to-C substitutions could, however, also occur due to other reasons, such as SNPs or mismatches. Most of the few existing statistical procedures for detecting binding sites in PAR-CLIP data do not account for types of substitutions other than PAR-CLIP induced substitutions. None of the existing methods enable the inclusion of additional information that are relevant for the biology of miRNA binding sites, such as the type of mRNA region that can help to detect binding sites. Moreover, the focus of existing models lies in detecting binding sites from only one experiment.

To detect binding sites, BayMAP, a Bayesian hierarchical mixture model taking other sources of substitutions into account, will be presented in this thesis. This allows the incorporation of additional information as well as a structure for reflecting dependencies of substitution positions very close to each other. The incorporation of additional information and neighborhood dependencies allows a better detection of miRNA-binding sites. Additionally, it also offers a better understanding of the biology of binding sites. Moreover, a method to combine the results of several PAR-CLIP experiments is developed to state the prediction of binding sites more precisely. Finally, BayMAP is compared to existing models in applications to real PAR-CLIP data sets as well as simulated data sets.
Fachbereich / Einrichtung:Mathematisch- Naturwissenschaftliche Fakultät » WE Mathematik » Mathematische Optimierung
Dokument erstellt am:04.10.2021
Dateien geändert am:04.10.2021
Promotionsantrag am:15.07.2021
Datum der Promotion:22.09.2021
english
Benutzer
Status: Gast
Aktionen