Statistical Strategy for eQTL Mapping Using RNA-Seq Data

• Activity Number: 354
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: International Chinese Statistical Association
• Author(s): Wei Sun*+
• Companies: The University of North Carolina at Chapel Hill
• Keywords: RNA-seq ; allele-specific expression ; isoform ; generalized linear model ; penalized regression ; alternative splicing

Abstract

RNA-seq is going to replace gene expression microarray in the near future for genome-wise assessment of the transcriptome. In addition to be more accurate and more sensitive, RNA-seq also provides new information that is not available from microarray, and novel statistical methods are needed to fully explore these new information. We will introduce two statistical approaches that aim to explore two types of characteristics of RNA-seq data for eQTL (gene expression quantitative trait loci) mapping. First, we develop a likelihood-based approach for eQTL mapping using both total expression and allele-specific expression measurements. Modified generalized linear models are separately developed for total expression and allele-specific expression, and then the two models are joined together by a sharing parameter. Secondly, we will discuss a penalized regression approach for mapping the variation of the isoforms proportions of a gene, or in other words, the genetic basis of alternative splicing.