Genome-scale experimental platforms provide vast amounts of data with the potential to dramatically improve our understanding of biology and pave the way to new therapies for human diseases. The challenge now is developing the right conceptual and mathematical frameworks to process and interpret these data, and ultimately transform it into knowledge about biological systems.
The primary goal of my research is to provide actionable data-driven models for use in cancer genomics and drug development. The main interest lies in network-based methods that extend across multiple data dimensions: genome, transcriptome, proteome, metabolome; and multiple biological scales: from single genes and proteins, to complexes and pathways, to entire cells, tissues and organs. My recent work includes algorithms for mapping dynamic changes in networks during development and in response to disease and stress conditions, robust methods for biomarker identification, and probabilistic methods for comparing biological networks across species.
Most recently I developed a computational method for constructing network-based ontologies of gene function. Ontologies are essential in many domains, capturing knowledge as a hierarchy of concepts and their interrelationships. Most ontologies, including those in biology, are typically developed through manual expert curation. In our work, we found that the vast amounts of genomic data now available can be transformed to automatically assemble on ontology of gene function that rivals manually constructed ontologies in coverage and power. Our network-extracted ontology (NeXO) provides a new resource for systems and synthetic biology - i.e. a data-driven multiscale catalog of cellular machinery, from genes to complexes, to pathways and higher-order processes. It also provides a way to systematically update manually constructed ontologies with new terms and relationships derived from high-throughput data (see article and commentary in Nature Biotechnology). You can access the NeXO ontology using Cytoscape or an OBO file browser such as OBO-Edit (see here). NeXO is now also available online:www.nexontology.org.