Abstract:
|
Many complex disorders, such as Alzheimer's disease, breast cancer, and type 1 diabetes, exhibit variable age of onset. Age of onset can provide useful information regarding disease subtypes and locus heterogeneity. Thus, incorporation of age of onset in linkage analysis can potentially increase the statistical power to detect linkage. However, age of onset itself is usually not the primary phenotype. Thus, it is in general not appropriate to treat age of onset as the outcome variable and apply standard survival analysis model in a genetic linkage context. We propose a likelihood approach to incorporating age of onset in the linkage analysis, still treating the disease status as the primary phenotype. We consider two models that underly the proposed likelihood: a pleoitropy model and a tight coincident linkage model. Simulations will be used to illustrate the statistical power comparison when age of onset is included and when it is not.
|