• Editor's note: The MAGNet center formally closed in July 2016, following the mandatory conclusion of its grant after more than 10 years of activity. The pages in this section constitute an archive of its work.

The mission of the Center for Multiscale Analysis of Genomic and Cellular Networks (MAGNet) is to develop novel structural biology and systems biology methods and tools for the dissection of molecular interactions in the cell and for the interaction-based elucidation of cellular phenotypes. These tools are made freely available to the the members of the research community. They are also validated in the context of the Center's own research program through collaborative projects with experimental biologists. Several of these projects have already led to results of seminal nature, including for instance the elucidation of the role of DNA shape in protein-DNA binding specificity [Joshi et. al. 2007 , Rohs et. al. 2009 , Slattery et. al. 2011], the identification of the Master Regulators of the mesenchymal subtype of Glioblastoma [Carro et. al. 2009], and the discovery of an extensive microRNA-mediated regulatory network of RNA-RNA interactions in brain tumors [Sumazin et. al. 2011].

MAGNet is one of 8 National Centers for Biomedical Computing. These centers, in conjunction with individual investigator awards, are creating a networked effort to build the computational infrastructure for biomedical computing in the nation. The NCBC program is devoted to all facets of biomedical computing, from basic research in computational science to providing the tools and resources that biomedical and behavioral researchers need to do their work. In addition to carrying out fundamental research the NCBCs play a major role in educating and training researchers to engage in biomedical computing.

MAGNet is also one of 12 inter-disciplinary Centers for Cancer Systems Biology (CCSBs), a component of the National Cancer Institute's Integrative Cancer Biology Program. The CCSBs provide a core framework for applying systems biology approaches to cancer research through the development and implementation of computational models of processes relevant to cancer prevention, diagnostics and therapeutics. The CCSBs seek to integrate experimental biology with mathematical modeling to foster new insights in the biology and new approaches to the management of cancer.

Our Research Goals

MAGNet investigators apply a combination of knowledge-based and physics-based methods to study the organizational principles that underlie the operation of cellular networks. A basic tenet of the Center's research program is the notion that by bridging together and integrating molecular-interactions across multiple levels of granularity (from the atomic to the macroscopic) and by using multiple data modalities (from structural coordinates to genetic and epigenetic variability) accelerated progress can be achieved, both within and across research domains.The central theme of the Center is the multiscale analysis of cellular networks. Cellular processes are determined by the concerted activity of thousands of genes, their products, and a variety of other molecules. This activity is coordinated by a complex network of biochemical interactions which control common intra- and inter-cellular functions over a wide range of scales. Understanding this organization is crucial for the elucidation of biological function and for framing health related applications in a quantitative, molecular context.

Driving Biological Projects

The theme of the Center is manifested in a number of driving biological projects (DBPs) that target broad areas of basic research, including:

• tackling the issue of biomolecular interaction directly, at the structural and physiochemical level,
• constructing context-specific maps of cellular interactions, and
• using such maps to dissect complex diseases.

The biological questions posed by the DBPs generate the requirements that drive the computational research carried out by the Center's investigators. This work involves both the development of novel analytical frameworks (basic computational research) as well as the development of tools and models that leverage these frameworks (applied computational research):

Basic Computational Research addresses general computational and algorithmic challenges raised by the DBPs. Research activities include the development of

• machine Learning (ML) algorithms for evidence integration, classification, and inference,
• natural language processing algorithms, and
• software engineering methodologies and frameworks for the assembly of the center repository based software platform.

Applied Computational Research focuses on the development of novel algorithms and tools to support specific biomedical applications. The approaches used are both knowledge-based and physics-based and incorporate the methods yielded by efforts in basic research. Algorithms are combined with existing and new databases to build a modular and extensible bioinformatics platform (geWorkbench) and an associated software toolkit for the analysis of biomolecular interactions.

Software and Data Resources

Research at MAGNet and the Columbia University Department of Systems Biology has led to the development of a variety of methods and data for the study of genomic and cellular networks. Through a platform called geWorkbench, we make these software and data sets available for the wider research community. geWorkbench is an interoperable, grid-enabled, state-of-the-art bioinformatics platform that integrates resources developed at MAGNet with a variety of other existing bioinformatics modules for the analysis, visualization, and management of multiple data modalities. It also enables complex bioinformatics workflows and biomedical applications using a simple yet powerful visual front-end and scripting language.

In addition to MAGNet tools, geWorkbench provides access to a rich collection of components supporting the analysis and visualization of many genomic data types (gene expression, sequence, structure, gene networks, etc). Some of these components have been developed de novo while others wrap popular 3rd party software such as Cytoscape, the Multi Experiment Viewer (MEV) and GenePattern. Over 70 geWorkbench modules are now available. Learn more about our software and data sets.

Research Infrastructure

To support biomedical computation research the MAGNet Center leverages a world-class information technology infrastructure. Additionally, MAGNet's Training Core ensures that the methods developed by the Center are integrated into the educational offerings of Columbia University's Medical School.

DREAM

MAGNet has been a longstanding supporter of the Dialogue for Reverse Engineering Assessment and Methods (DREAM). This annual event invites researchers to participate in a series of “DREAM challenges,” which are designed to identify the best practices for predicting the structure of biological networks. DREAM challenges help to assess the quality of  the algorithms that researchers use to characterize the networks that underlie biological systems. By inviting investigators to test their methods on specific problems and data sets, DREAM aims to achieve a fair comparison of the strengths and weaknesses of these methods and a clear sense of the reliability of the models that researchers create.