Efficient Bayesian Model-Based Clustering Method for Identification of Transcription Factor--Binding Sites for ChIA-PET Data

• Activity Number: 501
• Type: Contributed
• Date/Time: Wednesday, August 3, 2016 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Ioannis Vardaxis* and Bo Henry Lindqvist and Finn Drabløs and Morten Beck Rye
• Companies: Norwegian University of Science and Technology and Norwegian University of Science and Technology and Norwegian University of Science and Technology and Norwegian University of Science and Technology
• Keywords: Model-based ; ChIA-PET ; Protein Binding Sites ; MACS ; Bayesian ; Clustering

Abstract

It is known that genomes are organized as three-dimensional rather than linear structures in the nucleus of the cells. Those structures play an important role in chromosomal activities such as transcription. Transcription factors (TFs) are proteins that bind on specific DNA regions, called transcription factor binding sites (TFBSs), and regulate the transcription of DNA to RNA. Many TFs bind on DNA regions that are far from the transcription start sites (TSSs), suggesting that there might be a mechanism that brings together the TFs and the TSSs in order for the transcription to be regulated. This mechanism results in the interaction of two, or more, linearly distal DNA regions by making them spatially close through loops of the DNA. The location of the TFBSs and their distances to other TFBSs is very important. In this paper we propose a Bayesian model-based clustering and inference method for finding the TFBSs in a ChIA-PET data. We show that our method distinguishes close TFBSs more accurately than the MACS algorithm and we also find more TFBSs that MACS was unable to find.