Title: Incorporating a large number of functional annotations in penalized linear regression
Authors: Hu Yang - Central University of Finance and Economics (China) [presenting]
Wei Pan - University of Minnesota (United States)
Abstract: Many approaches based on penalized linear regression (such as lasso, adaptive lasso, elastic net, network penalty, etc.) have been successfully applied to analyze sparse and high-dimensional genetic data, including variable selection to facilitate interpretation of associations between predictors and a response. However, these methods are unable to incorporate a large number of function annotations, which can be collected from various omic studies, offering useful prior information. A new penalized linear regression method is proposed to incorporate such information so that we can better select predictors and predict gene expression.