Integrating Multiple Glycans and Genetic Data Using Joint Modeling Techniques: An Application in Leiden Longevity Study

• Activity Number: 183
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Jeanine Houwing*+ and Roula Tsonaka and Hae-Won Uh and Manfred Wuhrer and Eline Slagboom
• Companies: Leiden University and Leiden University Medical Center and Leiden University Medical Center and Leiden University Medical Center and Leiden University Medical Center
• Keywords: IgG glycosylation data ; Omics data ; gene-set testing ; families

Abstract

Identification of markers that reflect the biological age of individuals is of paramount importance in aging research. This is also the case for the Leiden Longevity Study (LLS) that motivated this work. In LLS several biomarkers and genetic data have been collected on 1671 offspring of nonagenarian sibling pairs and their partners, who are treated as controls. From the recorded biomarkers, we focus on IgG glycosylation, and investigate their joint association with familial longevity. So far association testing between glycosylation and healthy aging is done separately for each glycan. However, joint analysis of multiple glycans and genetic data could enhance our understanding of the biological mechanisms and improve predictions. Therefore we have developed a novel joint modelling framework to test associations between multiple omics data, genetic data and healthy aging. In particular, we use hierarchical random-effects models to model jointly the multiple glycosylation data and build associations with gene-sets and phenotypes. The advantages of this approach are that it can accommodate familial relationships, the sampling design and the between glycans and SNPs correlations.