Uncovering mammalian transcriptional regulatory mechanisms by computational approaches

Transcription regulation in eukaryotes is a complicated and highly orchestrated mechanism that involves transcription factor proteins and cis-regulatory elements. Uncovering these regulatory mechanisms in detail may provide a comprehensive understanding of basic cellular functions and insight to pathophysiology of complex human diseases. While experimental methods to study transcription regulations often focus on a small amount of genes or transcription factors, computational methods aim to tackle this problem through a systematic approach. Recently, computational efforts to delineate the transcription regulatory mechanisms have been focused on three aspects: identification of transcription factor binding sites, identification of transcriptional regulators of a set of genes, and identification of the transcriptional regulatory network of all the genes in the genome. With a rapid accumulation of completely sequenced genomes, many studies have developed powerful methods and generated promising results in these three areas. However, most of the methods were designed for simple organisms such as yeast and worms, and there are not many methods that may be applied to mammalian species successfully and reliably. Therefore, this thesis work aims to develop new computational approaches to address these three issues in mammals. The methods presented here may provide a better understanding of mammalian transcription regulation.