Data Visualization Tools for Large Biological Data Sets

Researchers have access to an ever-growing volume of data available at multiple levels of biological analysis. Many visual analytic tools have been developed to display a variety of biological data types but many of these tools are challenging to use and only examine one biological level of analysis at a time. The development and testing of hypotheses is difficult when the information is hard to integrate and laborious to interpret. The application of data visualization principles and user experience design best practices could improve systems biology research workflows by providing visual analytic tools with what is known in the information visualization community as a “transparent” user interface. This thesis consists of four papers that explore two central questions: 1) What is the best way to represent biological information at different levels of analysis? and 2) How do we enable researchers to explore and interact with their data as naturally and intuitively as possible? The first paper describes, ePlant, a tool for visualizing multiple levels of data that was developed using an agile process that included several rounds of user testing. The second paper presents Gene Slider, a tool for visualizing the conservation and entropy of orthologous DNA and protein sequences using a data visualization paradigm that takes better advantage of preattentive visual processing than current methods. The third paper describes Topo-phylogeny, a tool for visualizing phylogenetic relationships using a topographic map visualization paradigm that requires less cognitive processing to interpret than traditional tree diagrams. The final paper demonstrates the importance of user testing when developing a “rapid serial visual presentation” interface for identifying genes of interest using electronic fluorescent pictographs. Together these papers illustrate the complexities and benefits of applying data visualization principles and user experience design best practices to building data visualization tools for the analysis of large biological data sets. Given that hypothesis generation is fundamentally a creative process, any tools or techniques that can help researchers consider their data at a deeper level should be valuable to the scientific community.Ph.D.2019-02-28 00:00:0