Omnibus Risk Assessment via Accelerated Failure Time Kernel Machine Modeling

• Activity Number: 168
• Type: Topic Contributed
• Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Jennifer Sinnott*+ and Tianxi Cai
• Companies: Harvard University and Harvard University
• Keywords: risk prediction ; genomic data ; kernel machines ; accelerated failure time model ; omnibus test ; pathways

Abstract

Integrating genomic information with traditional clinical risk factors to improve the prediction of disease outcomes could profoundly change the practice of medicine. However, the large number of markers and the complexity of the relationship make it difficult to construct accurate risk prediction models. Standard approaches often rely on marginal associations and may not capture non-linear or interactive effects. At the same time, much work has been done to group genes into pathways. Integrating such biological knowledge into statistical learning could potentially improve model interpretability and reliability. One effective approach is to employ a kernel machine (KM) framework, which has been recently extended to analyzing survival outcomes under the Cox model. In this paper, we propose KM regression under an accelerated failure time model. We derive a pseudo score statistic for testing and a risk score for prediction of survival. To approximate the null distribution of our test statistic, we propose resampling procedures which also enable us to develop alternative robust testing procedures that combine information across models and kernels.