Statistical Inference Problems Arising from the Availability of Multiple RNA-Seq Data Sets

• Activity Number: 19
• Type: Topic Contributed
• Date/Time: Sunday, August 9, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Jingyi (Jessica) Li*
• Companies: UCLA
• Keywords: RNA-seq ; multi-sample ; inference ; transcript ; isoform ; alternative splicing

Abstract

Since the inception of next-generation mRNA sequencing (RNA-seq) technology, various attempts have been made to utilize RNA-Seq data in assembling full-length mRNA isoforms, estimating isoform abundance, and finding alternative splicing sites. These problems are challenging and often involve identifiability issues in statistical modeling. Thanks to the decreasing cost of RNA-seq assays, an increasing number of RNA-seq data sets have been produced. The availability of multiple RNA-seq data sets provides more information than a single RNA-seq data set of a biological sample, therefore shedding new light on addressing these problems with improved accuracy. In this talk, I will present my two recent works about how to jointly use multiple RNA-seq data for statistical modeling and inference. The first part will address the mRNA isoform discovery and quantification problem, and the second part will be related to the discovery of novel alternative splicing sites.