Links

Tools

Export citation

Search in Google Scholar

Anwendung und Entwicklung biostatistischer Methoden zur Identifikation genetischer Risikofaktoren ; Application and development of biostatistical methods for identifying of genetic risk factors

Thesis published in 2015 by Sven Knüppel
This paper is available in a repository.
This paper is available in a repository.

Full text: Download

Question mark in circle
Preprint: policy unknown
Question mark in circle
Postprint: policy unknown
Question mark in circle
Published version: policy unknown

Abstract

Background: Most common chronic diseases are influenced by genetic and environmental factors. Analysis of single genetic markers, such as SNPs (single nucleotide polymorphisms), explain a small part of genetic causes. The application and development of biostatistical methods are necessary to get further insights into the genetic causes of chronic diseases by taking into consideration of single and multiple genetic markers. Objectives: The objective was to develop biostatistical methods to identify genetic risk factors using single SNPs and their combinations and to apply these methods to select systematically risk-related allele combinations, such as haplotypes, from a large number of SNPs. In addition, three candidate gene studies were carried out to evaluate the effect of different promising genes. Methods: The Multi-locus stepwise regression (MSR) method was developed using data of a German genome-wide association study on atopic dermatitis. The MSR combines the advantages of stepwise selection methods with haplotype-based approaches. The MSR extends stepwise SNP-combinations successively if the result of the haplotype-based test is statistically improved until a stop criterion is met. The MSR was subsequently applied to investigate unlinked SNPs and their combinations as part of the EPIC-Potsdam study. The German genome-wide association study of atopic dermatitis consists of a case-control study (939 cases and 975 controls) and a family study (268 families with 529 children). 94 tagSNPs of EDC region on chromosome 1q21 and four known FLG-mutations encoding structural proteins that are expressed during terminal differentiation of the human epidermis were used Within the EPIC-Potsdam study the following associations were investigated in four sub-studies: [1] 41 SNPs for body-mass index (kg/m²) and waist circumference (cross-sectional study, MSR, and permutation test), [2] 2 SNPs (ADH1B, ADH1C) as marker for alcohol intake in relation to incident cardiovascular diseases (case-cohort study, modified Cox proportional hazards regression), [3] 1 SNP from MTTP gene that encodes a lipid transfer protein for cardiovascular diseases under consideration of cholesterol levels (case-cohort study, modified Cox proportional hazards regression), and [4] 7 tagSNPs from SCD1 gene encoding a protein that is involved in lipid metabolism for metabolic risk factors (cross-sectional study, analysis of covariance). Results: The MSR used in the genome-wide association study identified a haplotype pattern in the case-control study and was replicated in the family study. This haplotype pattern of four SNPs reflects the well-known FLG effect and an additional FLG-independent effect on atopic dermatitis. In the EPIC-Potsdam study, the MSR identified SNP-combinations associated with body-mass index and waist circumference, but these were not statistically significant compared to the simulated distribution under the null hypothesis of no genetic effect. The SNPs in ADH1B and ADH1C showed no associations to risk of cardiovascular diseases. An interaction was observed between MTTP-SNP rs1800804 and cholesterol levels on cardiovascular diseases. No evidence for an effect of single SCD1-SNPs and their corresponding haplotypes on metabolic risk factors was found. Conclusion: A stepwise haplotype-based SNP selection method was developed and successfully applied to one candidate gene region. The application of the method to unlinked SNPs requires a careful selection of SNPs. Although the data that were used to exemplify the method did not add additional evidence concerning the genetic etiology of combined SNP effects beyond single SNP analysis. Further studies are needed to assess the real value of the MSR. In the candidate gene studies as well none essential multi-locus marker effects were found.