Statistical Strategies for Identification of the RNA-Protein Binding Site in CLIP-Seq

• Activity Number: 58
• Type: Topic Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Biometrics Section
• Author(s): Jonghyun Yun*+ and Xinlei Wang and Tao Wang and Guanghua Xiao
• Companies: University of Texas Southwestern Medical Center and Southern Methodist University and University of Texas Southwestern Medical Center and University of Texas Southwestern Medical Center
• Keywords: CLIP-seq ; Next generation sequencing data ; Non-homogeneous hidden Markov model ; RNA binding protein

Abstract

The crosslinking immunoprecipitation coupled with high-throughput sequencing (CLIP-seq) technique has been increasingly used for the global mapping of RNA-protein binding sites. There are three key features of the CLIP-seq experiments: The sequence read tags are likely to form an enriched peak around each RNA-protein binding site; the crosslinking procedure is likely to introduce a specific mutation in each sequence read tag at the binding site; and CLIP-seq data are strand-specific, and they can reach near-single base pair resolution. The photoactivatable ribonucleoside enhanced crosslinking immunoprecipitation (PAR-CLIP) is one of CLIP-seq variants, whose crosslinking method induces nucleotide-specific mutations (T to C or G to A) at the site of contact. To pinpoint the binding sites at the single base-pair resolution, we developed a novel approach that adopts non-homogeneous hidden Markov models to incorporate the read coverage, mutation and nucleotide sequence at each genomic location into an integrative model. Both simulation studies and data applications have shown that our methods provide reliable inferences for detection of RNA-protein binding sites from PAR-CLIP data.