Zhihua Zhang's HomePage
The way a cell regulates gene expression is crucial to life. To understand the gene expression and its regulation was at the center of my research interests. In human, there are about 200 different cell types, carrying virtually identical genomes, whereas the gene expression patterns differ substantially among the cell types . How do cells specifically express a subset of genes? I have addressed this question from the prospective of epigenetics. Histones are subject to enzymatic post-translational modifications (HMs), primarily of their N-terminal tails , and such modifications provide an additional layer of regulative information to guide gene expression. By building a quantitative model based on associated HMs and the tissue specificity of gene expression, I have systematically illustrated the relationship between the two in human cells (Zhang et al 2010 Submitted). I identified the HM types that were most predictive of a gene’s tissue or cell type specific regulation, and found that the predictive information encoded in HM profiles is redundant. This information redundancy points to the existence of multiple regulation mechanisms defining/deciding tissue or cell type specificity.
Another key factor in cell development is cell to cell variation in gene expression (gene expression noise). Due to stochasticity in gene transcription, the expression level of the same gene may vary from cell-to-cell in an isogenic cell population even when the cells are all in the same developmental and/or cell-cycle status. As a trait, gene expression noise may even be subject to selection. It is intuitive that gene expression noise is under negative selection in essential genes. What interested me was whether there are genes in which noise is subject to positive selection, and was the first to report plasma-membrane transporters as a group of yeast genes which have been selected for elevated expressional noise (Zhang et al 2009. Mol Syst Biol). Gene expression is regulated by proteins called transcription factors (TF), and regulative interactions form a complex web called transcription regulatory network (TRN). TRN has been studied at multiple levels during the past decades. At the sequences level, the primary question is how to associate the binding of a TF in a promoter region with the expression level outcome of the downstream gene. By analysing the accuracy and usage of the motif in transcriptional regulatory networks, I found that it is more important to identify most motifs for every TF than to identify all motifs for some TFs (Zhang et al. 2008. Nucl Acids Res). At the network topological level, I studied the dynamic behavior of crucial TFs in yeast under different conditions (Zhang et al. 2006. PLoS CB). In this study, I developed of the subgraph preference profile, which is a powerful tool to describe the dynamic behavior of central players in the network. This work was one of the very first studies on the cellular network dynamic behaviors, and the model could also be applied to other complex directed network models, e.g metabolic network. Global properties of networks normally represent profound design principles of complex systems, such as gene expression regulation(Zhang et al. 2009 PLoS ONE), which suggests that network analysis is a powerful approach to biological questions. Early in my Ph.D study, I did a comparative genomics project in mammals. By comparing the structure of conserved genome sequence segments between human, mouse, and rat, I detected gene fusion/fission events during the evolution of the three mammals (Zhang et al. 2006. JBT ). Such events are important because gene fusion are normally associated with disease, protein-protein interaction, or new genes. For example, gene fusion has been widely observed in a variety of cancer types (e.g. TMPRSS2– ERG fusion in prostate cancer). Long term goals. My future research interests will focus on three interconnected fields To understand of gene regulation. Specially, to understand the dynamic changes in the epigenetic networks of human cancer cells and stem cells. Epigenetic marks, like histone modifications and DNA methylation, play important roles in gene expression regulation. The patterns of epigenetic marks are highly dynamic and responsive to external and internal stimuli. Thus, such epigenetic patterns could serve as biomarkers of the cell status.We will integrate massive data sets in the public domain, e.g. the ENCODE project and the NIH epigenomic roadmap project, and generate our own data as well. By comparing data between different cancer stages and between cancers and stem cells, we aim to identify epigenetic signals which indicate stage changes in cancer cells, as well as signals which distinguish cancer cells from stem cells. With this specific aim, a sophisticated computational model will be built to sort dynamic changes in epigenetic networks. To understand the genetic interaction between the human genome and the human living environment is my long term goal. One promising approach to probe this question is metagenomics. The vast majority of microbial species live in complex, highly diversified and largely uncharted communities. These communities play central roles in human health, ecosystem dynamics, agriculture, and environmental stewardship. Research focused on the human microbiome, particularly in the gut, is of great interest. New and exciting advances in metagenomics have produced novel data and provided insight into the composition of the human microbiota. Yet to date only basic comparative genomic analyses have been applied to these data, and relatively little effort has been directed at conducting a largescale, systemlevel study of these complex communities, let alone the interaction between these communities and human host. By integrating multiple data sources, many interesting question could be asked, e.g. how different food sources would change the bacteria profile in the gut and how the human body react to such changes? To what extent occur gene transfer between genetically modified food and bacteria in the gut and the human host? Can we model the process mathematically? This is important not only because of the obvious public interest, but it also could be used to develop novel methods to study gene flow between species and therefore contribute to the study of more basic evolutionary questions . The evolution of non-protein-coding RNA (ncRNA). As the main component of the“dark matter” in the human genome, ncRNA has showed increasing importance for gene regulation. For example, microRNAs are widely accepted as a major player in the gene regulation network, and piRNAs have been showed to play important roles in germ line development. The list of non-protein-coding RNA has increased in recent years. However, except for microRNAs, the evolution of non-protein-coding RNA has barely been studied. Drosophila species provide a very good model to study the evolution of non-protein-coding RNAs. With the help of the next generation sequencing, we will detect the early events on the origin of non-protein-coding RNAs in several Drosophila species. Together with computational simulation studies, we want build an evolutionary model of how non-protein-coding RNAs originated and were integrated into the gene regulation network,thereby obtaining insights into the evolution of the gene regulation network. Because so many types of non-protein-coding RNAs have been detected, and they are not only different in size but also in genesis and biological functions, we will first focus on specific types of non-protein-coding RNAs. One candidate are the piRNAs which have a relatively well characterized biogenesis and biological function. Another candidate are long noncoding RNAs (lncRNAs), given that lncRNAs have a broad functional spectrum and have potential interactions with other epigenetic regulatory players. Small tools I have developed Peano3D.m (Matlab), FFTPF1D (Matlab)
|
