September 9 – 14, 2012, Dagstuhl Seminar 12372
Biological Data Visualization
Carsten Goerg (University of Colorado, US)
Lawrence Hunter (University of Colorado, US)
Jessie Kennedy (Edinburgh Napier University, GB)
Sean O'Donoghue (CSIRO – North Ryde, AU)
Jarke J. van Wijk (TU Eindhoven, NL)
For support, please contact
Biology is rapidly evolving into a 'big data' science, and as a consequence there is an urgent and growingneed to improve the methods and tools used for gaining insight and understanding from biological data. Over the last two decades, the emerging fields of computational biology and bioinformatics have led to significant advances primarily in automated data analysis. Today, however, biologists increasingly deal with large, complex datasets (e.g., 'omics' data) where it is not known in advance what they are looking for and thus, automated analyses alone cannot solve their problems. Interactive visualizations that can facilitate exploratory data analysis and support biologists in creating new hypotheses lend themselves to complement automated analyses. Bioinformaticians already have built a variety of tools for visualizing different types of biological data and those tools are widely used in the community. So far, most bio-related visualization research has been conducted by people outside of the visualization community, people who have learned about visualization but are often not aware of research in the visualization community. Consequently, the current tools do not embody the latest advancements in design, usability, visualization principles, and evaluation.
One main goal of this first Dagstuhl Seminar on Biological Data Visualization was to bring together the users (biologists), current visualization tool builders (bioinformaticians), and visualization researchers to survey the state-of-the-art of the current tools and define a research agenda for systematically developing the next generation of tools for visualizing biological data. Only a close collaboration of the researchers from all three communities can create the synergies necessary to address the challenges in analyzing and visualizing large and complex biological datasets.
Topics discussed during the seminar included:
- Challenges in visualizing biological data. Biological data is very heterogeneous. It contains spatial data, graphs, tabular data, and textual data. Challenges are wide spread: open-ended data quantity, open-ended exploratory tasks, long-term analyses, rich analytics, heterogeneous data, usability and evaluation of tools.
- Design and visualization principles, research in human-centered design, usability, and evaluation of interactive data-analysis.
- Creating a common research agenda and a common understanding of the problem field of biological data visualization.
- Integration of multiple visualizations for different data types and tasks into one tool to support more complex analysis scenarios.
- Designing an infrastructure for next generation visualization tools.
- Establishing collaborations between computer scientists and biologists.
41 researchers from 9 countries participated in this seminar. Many participants came from the US and from Germany, others came from Canada, Australia, and a number of other European countries. There was a good mix of researchers from the visualization, bioinformatics, and biology communities.
- Computer Graphics/Computer Vision
- Information Visualization
- Data Visualization
- User Interfaces
- Visual Analytics