Improved Understanding of Huntington's Disease Onset Through a Genetic Mixture Model with Genetic Biomarkers and Distribution-Free Random Effects

• Activity Number: 62
• Type: Topic Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Mental Health Statistics Section
• Author(s): Tanya Garcia*+ and Yuanjia Wang and Yanyuan Ma and Karen Marder
• Companies: Texas A&M and Columbia University and Texas A&M and Columbia University
• Keywords: Huntington's disease ; biomarkers ; semiparametric ; censoring

Abstract

Huntington's disease (HD) is an autosomal dominant, neurodegenerative disease, for which biomarkers (i.e., CAG repeats) are currently being investigated. A key interest is inferring the age of disease-onset (AAO) from these biomarkers as such knowledge is essential for understanding disease development and assisting patients to make important medical decisions (e.g., treatments). We present a new approach for estimating the AAO distribution of HD based on a varying-coefficient, logistic mixed effects model. We use semiparametric techniques to incorporate the effect of different biomarkers and accommodate family-induced correlation without the need to postulate or estimate any latent variable distributions. Finally, we handle censored observations using a jack-knife estimator which yields consistent and efficient model parameter estimates. Our method addresses several key issues in HD research: incorporating patient biomarkers using a data-driven, nonparametric approach, accounting for family-induced correlation via distribution-free random effects, and handling censoring from patient-dropout.