Round Table: Health as a Big Data Problem: Genomes to Populations
Kathryn Cooper, Ph.D.
School of Interdisciplinary Informatics
University of Nebraska at Omaha
In October 2014, the NIH Associate Director for Data Science Dr. Philip Bourne spoke about the expanding challenge of biomedical research analyses, saying "the health of each one of us is a big data problem." Indeed, the limits of health data collection on any one individual have no bounds, from standard data collection (routine lab checks, diagnostic tests, electronic health records) to newer and less common technologies (personal genotypic profiling, lifestyle monitors, social networking). As the scale and dimensionality of this data grows, so grows the need for robust, informative and reproducible methods for analysis of health data. One emerging approach to analysis of biomedical and clinical data has been the use of network models to display relationships between elements of a system. The network model is flexible in creation and type – any data from which relationships can be inferred can be represented as a network – but the network model also can also be instructive. When built with integrity, a network's structure can reveal important elements about a system. For example, in a protein-protein interaction network, nodes that have many physical interactors tend to represent proteins that are essential for organism viability. In a network representing airports and the flight paths between them, the radius of the network can tell the user how many flights minimum would be required to traverse an entire country. In this way, the network model offers a host of advantages that can allow it to serve as a main or supplementary approach to assimilating knowledge from biomedical data. In this talk, I will discuss the research I have performed to (1) apply and develop network modeling approaches to genomic and public health data and (2) to ensure the integrity of the network model. In addition I will (3) discuss the future of biomedical big data research, the "quantified self", and my interest in the design improvement of biomedical big data research experiments that will lend itself toward more informative and employable results.
Kate Cooper is a Post-Doctoral Research Associate in the School of Interdisciplinary Informatics at UNO. Kate’s background lies firmly in bioinformatics and she specializes in identification of actionable hypotheses using network-based analysis of genomic and proteomic data, with interests in expanding these approaches to studies in public health, disaster preparedness, and strategic biodefense. Kate has been a member of the UNO Bioinformatics research group since 2005 and has found success collaborating on a number of disease-related biomedical research projects including but not limited to hearing, arenaviruses, HIV/AIDs, Chikungunya virus, Burkitt’s lymphoma, and aging.