Challenges in Estimation of Genetic Effects from Multiple Cases Family Studies

• Activity Number: 8
• Type: Invited
• Date/Time: Sunday, August 4, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Roula Tsonaka*+ and Jeanine Houwing-Duistermaat
• Companies: Leiden University Medical Center and Leiden University Medical Center
• Keywords: Ascertainment ; Heritability ; Outcome-dependent sampling ; Penalized likelihood ; Prevalence

Abstract

Estimation of genetic effects in multiple cases family studies is often complicated by the outcome-dependent sampling and the within families correlation. An approach to deal with both of these features is the ascertainment corrected mixed-effects model. However, for small sample sizes and when the disease is rare convergence issues may arise, because the data do not provide sufficient information to estimate all model parameters. To overcome such numerical difficulties, we propose a penalized maximum likelihood estimation procedure which reliably estimates the model parameters in small family studies by using external population-based information. Extensive simulations have shown this to be a promising approach to reliably estimate the quantities of interest in small family studies. Finally, we illustrate our proposal using the Genetics In Familial Thrombosis study, which considers families with at least two affected siblings to contribute to the search for novel genetic risk factors for venous thrombosis.