Computational and biophysical studies of epigenomes

My dissertation research focuses on developing novel computational and biophysical methods to study epigenomes and on application of these methods to unravel molecular mechanisms of epigenetic regulation of gene expression.;In multicellular organisms, DNA is not sufficient to define cell identity. A eukaryotic cell's gene expression pattern is regulated by a series of epigenetic mechanisms. The complete set of epigenetic components in a cell is an epigenome. Chromatin is one of the major carriers of epigenomic information. Studies of epigenetic phenomena accumulated over the years have shown their critical roles in important biological processes including gene regulation, development and disease. However, little was known about the landscape of epigenomes and how and to what extent chromatin states influence the gene expression program. With the advent of Chromatin immunoprecipitation combined with high-throughput sequencing (ChIP-Seq), the new and revolutionary experimental technique has brought about a new era in which fundamental questions can be addressed. Novel statistical methods are called for to analyze the experimental epigenomic data, and comprehensive approaches need to be developed to study epigenomes from the perspectives of computational biology, and computational and statistical biophysics.;In the research projects leading to this dissertation, we performed systematic analyses and developed novel computational methods to analyze ChIP-Seq data, including SICER (Statistical Identification of ChIP-Enriched Domains) method to identify epigenetic marks in the genome and coarse-graining approach to identify large-scale chromatin domains in the genome.;With these novel computational methods, we analyzed both the equilibrium and dynamic epigenomic structure in several types of human cells. In the study of equilibrium epigenomes, we focused on the functional modules of chromatin modifications. We analyzed nearly 40 histone modifications in human CD4 + T-cells in a combinatorial approach. We discovered a group of histone modifications that are highly correlated with each other in the human genome, indicating their cooperative function. We also studied two important histone variants, H3.3 and H2A.Z in the human genome. We found that the nucleosomes containing both of these variants are highly unstable and are functionally important to the regulation of gene transcription.;In the dynamic aspect of epigenomes, we investigated histone modification landscapes and their change during the differentiation of human CD133 + cells to CD36+ cells. The results revealed the chromatin signature in hematopoietic stem cells. We also studied the genomic distribution of a few histone acetyltransferases (HATs) and histone deacetylases (HDACs), enzymes that catalyze histone acetylations, and found their exceptional distinct distributions in active and inactive genes.;This dissertation covers the computational and biophysical contents from a variety of research projects. This work contributes to unraveling the fundamental mechanisms of gene regulation and ultimately to new diagnosis and treatment of various human diseases in the future.