Analysis of cell type and tissue-specific regulatory networks

Revision as of 12:26, 16 February 2015 by DanielMarbach (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Background: Characterizing the transcriptional regulatory networks that underlie the functioning of human cells in health and disease is a major challenge. Initial drafts of genome-wide regulatory networks are available for unicellular model organisms, but remained largely unchartered for higher organisms. Using novel methods applied to promoter and enhancer activity data from the FANTOM5 project, we have constructed a unqiue collection of 394 cell type and tissue-specific regulatory networks for human. Each of these networks specifies the genome-wide connectivity of TFs, enhancers, promoters and genes. We are currently using them to identify pathways that are perturbed by genetic variants underlying complex traits and diseases. However, we did not yet analyze the architecture of these networks in depth to learn more about cell type and tissue-specific gene regulation

Goal: The aim of this project is to characterize the architecture and "design principles" of cell type and tissue-specific regulatory networks. The students will -- with our guidance -- independently develop interesting questions and pursue them using diverse network analysis approaches. For example, we could ask whether certain cell types have a higher regulatory complexity than others, which TFs are most important in a given cell type, whether certain TFs can be associated with specific groups of cells or tissues, whether certain network components are shared across cell types, etc.

Mathematical tools: This project has a computational flavor and is best suited for students with interest in programming. We will apply standard tools used in many diverse problems in computational biology (e.g., clustering, gene ontology enrichment analysis, etc.), standard network analysis approaches (e.g., degree distributions, network motif analysis, etc.) and other methods depending on the chosen questions (e.g., inference methods to model TF activity). A programming environment such as R or Matlab will be used.

Biological or Medical aspects: The students will get more familiar with the field of network biology and gene regulation.

Supervisor: Daniel Marbach