Finding differentially expressed regions of arbitrary length in quantitative genomic data based on marked point process model.

Abstract

High-throughput nucleotide sequencing technologies provide large amounts of quantitative genomic data at nucleotide resolution, which are important for the present and future biomedical researches; for example differential analysis of base-level RNA expression data will improve our understanding of transcriptome, including both coding and non-coding genes. However, most studies of these data have relied on existing genome annotations and thus are limited to the analysis of known transcripts. In this article, we propose a novel method based on a marked point process model to find differentially expressed genomic regions of arbitrary length without using genome annotations. The presented method conducts a statistical test for differential analysis in regions of various lengths at each nucleotide and searches the optimal configuration of the regions by using a Monte Carlo simulation. We applied the proposed method to both synthetic and real genomic data, and their results demonstrate the effectiveness of our method. The program used in this study is available at https://sites.google.com/site/hiroshihatsuda/. H.Hatsuda@warwick.ac.uk.