Detecting Expression Patterns of the Distributions of Transcription Start Sites Using Marked Point Process
Recent technologies, such as the next-generation sequencers, 5SAGE, and CAGE, allow us to accurately and comprehensively determine exact transcription start sites (TSS) and reveal that the TSS are distributed over a relatively wide region of genes. Although analyzing the TSS distributions is important to develop our understanding of promoters and transcriptional regulations, there is no reliable method to analyze the distributions. We therefore propose a novel method to detect expression patterns of the TSS distributions by using a marked point process, in which points correspond to nucleotides and marks are their attributes, Gauss functions in this study. Our method considers both consistence of Gauss functions to the TSS data and prior knowledge of the data structure. We define a Gibbs energy function and minimize it by using a Monte Carlo simulation to find the optimum pattern of the TSS distribution. The experimental results show the effectiveness of our method.
transcription start sites expression pattern marked point process Monte Carlo simulation
Hiroshi Hatsuda
Department of Computational Biology The University of Tokyo Chiba, Japan
国际会议
海口
英文
53-57
2011-02-22（万方平台首次上网日期，不代表论文的发表时间）