Since the completion of the Human Genome Project it has become increasingly clear that most phenotypes — observable traits in an organism — do not result from single genes or isolated events. Rather, they emerge from the activities of thousands of molecular components as they interact within complex regulatory networks. These components, including nucleic acids, proteins, metabolites, and small molecules, all work together to produce specific traits.

If the careful balance within a regulatory network is dysregulated — for example, through a gene mutation, overexpression of a particular protein, infection, or environmental exposures that affect an organism’s physiology — a cascade of events may ensue that results in an altered phenotype. In many cases, such events can lead to phenotypes associated with disease.

From this perspective, the challenge now facing biologists is to understand how interactions among proteins, the cell’s regulatory logic, evolution at the genetic level, and epigenetic influences (heritable changes in gene expression caused by factors other than DNA sequence) all work together to produce physiologic and pathologic traits. Addressing this challenge requires the in-depth, system-wide characterization of all of the molecular interactions involved in specific behaviors of normal cells, tissues, and ultimately organisms, and identifying how these interactions differ from those in diseased states.

Joining quantitative analysis and experimentation

Researchers in the Columbia University Department of Systems Biology seek to understand biology at this high level of complexity. Operating in collaborative, multidisciplinary teams, we work to:

• map and reconstruct cell regulatory networks at the molecular level
• use computational models to predict how genomic and epigenomic diversity processed by these networks leads to physiologic or pathologic phenotypes
• experimentally validate the computational predictions that are derived from these models
• develop new technologies for studying biology at the systems level

This new, systems-level approach to biology requires the production and analysis of massive amounts of data. It has become possible only over the last decade with the arrival of several revolutionary new technologies, including 1) high-throughput screening devices for observing and comparing large numbers of network perturbations, and 2) next-generation gene sequencing platforms that allow us to compare genomes of many individuals within a population or gene expression under different conditions. At the Columbia Genome Center, these techniques help us to identify differences at the systems level that are causally, not just statistically, associated with distinct phenotypes. In addition, new high-throughput technologies for elucidating macromolecular structures and for generating metabolomic, phosphoproteomic, and epigenomic profiles are already playing important roles in identifying factors that affect the emergent behavior of cellular networks.

The challenge now facing biologists is to understand how phenotypes emerge from complex molecular networks.

Also critical to our work is the ability to integrate and analyze various types of data using computational algorithms running on large supercomputers. Our Center for Computational Biology and Bioinformatics (C2B2) has played a particularly important role in the evolution of the the Department of Systems Biology. C2B2 researchers have developed dozens of algorithms for conducting research in computational biology, and have made these available through a publicly accessible platform called geWorkbench. These software tools can be used to model a wide range of biological phenomena, including gene regulatory networks, molecular structure, and factors that modulate gene expression. We have also built a computing cluster that is among the largest platforms for conducting research in systems and molecular biology. These resources enable us to develop models of regulatory networks that can then be tested in the laboratory.

Research areas and goals

The integration of quantitative analysis, high-throughput experimentation, and technology development is the hallmark of systems biology at Columbia. In everything we do, the Department of Systems Biology stresses the importance of combining quantitative approaches with experimentation, and of bringing together theory and practical applications. This approach enables us to pursue research in a wide range of areas, including:

• Prediction of protein structure, function, and localization
• Study of protein-protein and protein-DNA interactions
• Gene expression analysis and prediction of regulatory network structure
• Study of complex inherited traits
• Reconstruction and analysis of metabolic networks
• Dynamic simulations of cellular networks
• Image analysis and interpretation

Our goal is not just to generate new biological insights, but also to develop innovative algorithms, technologies, and methodologies that will lead to future discoveries. In addition, much of our work is organized around driving biological problems related to fields such as cancer, immunology, stem cells, developmental biology, neurodegenerative diseases, and metabolic syndromes such as diabetes. Ultimately, our hope is that isolating critical points within biological networks could help to identify definitive biomarkers of disease or lead to new strategies for improving human health.