Dokument: Longevity and Genetic Data in Twin and Family Studies
| Titel: | Longevity and Genetic Data in Twin and Family Studies | |||||||
| URL für Lesezeichen: | https://docserv.uni-duesseldorf.de/servlets/DocumentServlet?id=33203 | |||||||
| URN (NBN): | urn:nbn:de:hbz:061-20150122-102650-9 | |||||||
| Kollektion: | Dissertationen | |||||||
| Sprache: | Englisch | |||||||
| Dokumententyp: | Wissenschaftliche Abschlussarbeiten » Dissertation | |||||||
| Medientyp: | Text | |||||||
| Autor: | Dr. Begun, Alexander [Autor] | |||||||
| Dateien: |
| |||||||
| Beitragende: | Prof. Dr. Giani, Guido [Gutachter] Prof. Dr. Janssen, Arnold [Gutachter] | |||||||
| Dewey Dezimal-Klassifikation: | 500 Naturwissenschaften und Mathematik » 510 Mathematik | |||||||
| Beschreibung: | Aging and survival are caused by a complex and non-observable interaction between genetic
and environmental factors. To reveal regularities of this interaction the traditional methods of survival analysis combined with ones of quantitative genetics data are needed. The data used by statistical analysis of longevity usually have a number of peculiarities and drawbacks such as selective sampling and incompleteness caused by censoring and truncation. Genetic data can include information on genes with a known location (genetic markers) for related individuals (e.g. twins, sibs or members of a family). Finding genes that are differentially expressed under two or more conditions is a main object in experiments with microarrays. Searching for such genes is usually based on statistical methods involving t-statistics and multiple testing and uses datasets with information about thousands of genes, but a relatively small number of individuals. Correlations between individuals are usually not taken into account in these studies. Phenotypic traits such as length of life and gene expressions can correlate for related individuals because such individuals share genetic and environmental factors. If we do not take into account these correlations the estimates obtained in the studies can be biased and conclusions are wrong. In this work we develop statistical models that combine the strength of the methods of the bi- and multivariate (survival) analysis with methods of genetic analysis and analysis of gene expression data. In the analysis of survival data we use the concept of frailty assuming that nonobservable susceptibility to death can contain both genetic and environmental components. Additional randomness in death process is caused by underlying hazard. Observed covariates in the form of a Cox-like regression are also included in the survival models. We discuss the methods and the problem of identifiability of such models. We show how genetic markers data can be used to locate the position of longevity or frailty genes. We also discuss how the mixed model method for detecting genes with differential gene expression can be adapted for twin data. All models are illustrated with examples based on analysis of real or simulated data. | |||||||
| Lizenz: |  Urheberrechtsschutz | |||||||
| Fachbereich / Einrichtung: | Mathematisch- Naturwissenschaftliche Fakultät | |||||||
| Dokument erstellt am: | 22.01.2015 | |||||||
| Dateien geändert am: | 22.01.2015 | |||||||
| Promotionsantrag am: | 04.07.2014 | |||||||
| Datum der Promotion: | 11.12.2014 |
