Another part of my research focuses on data visualization as a powerful tool for scientific discovery. Modern scientific studies generate complex, high-dimensional datasets that are often difficult to interpret using traditional analysis alone. By designing visualizations that reveal structure, highlight patterns, and make uncertainty explicit, we can transform raw data into insights that are both accessible and actionable. These techniques not only support researchers in exploring their data more effectively, but also help communicate scientific findings clearly to broader audiences.
FluxE
FluxE is an astrophysical visualization tool, rendering the computed flux density of astrophysical simulations, with a focus on stellar merger simulations. This system is an implementation of a framework designed to combine the approaches of observational and theoretical astronomers in studying distant astrophysical phenomena through the combination of a simulation volume and synthetic light curves. FluxE gives computational astrophysicists the ability to view simulated stellar mergers in the same way that observational astronomers view stellar mergers, with the added ability to view the simulated merger with spatial detail at multiple wavelengths. The researcher is provided with a variety of powerful features for exploring the data, including the abilities to superpose images from different wavelengths, to inspect precise wavelength intensities, to navigate through and within frames, and to animate sequences of images.
Related Publications:
- John Wenskovitch, James C. Lombardi, and Roger W. M. Hatfull. “FluxE: Exploring Flux in Astrophysical Simulations,” in SIGGRAPH ASIA 2016 Symposium on Visualization. SA ’16. Macau: ACM, 2016, 15:1–15:8. DOI: 10.1145/3002151.3002154.
Fixing TIM
Fixing TIM is a visual mining and analysis tool implemented as a solution to the 2013 BioVis Data Contest. Through a collection of linked views, experts can inspect the sequences and structures of proteins, in order to identify protein mutations and to help discover the effect of these mutations on protein function. We follow a client-server approach in which distributed data sources for 3D structure and non-spatial sequence information are seamlessly integrated into a common visual interface. Multiple 3D rendered views and a computational backbone allow comparison at the molecular and atomic levels, while a trend-image visual abstraction allows for the sorting and mining of large collections of sequences and of their residues. We evaluate our tool on the triosephosphate isomerase (TIM) family of structural models and sequence data, and show that our tool provides an effective, scalable way to navigate a family of proteins, as well as a means to inspect the structure and sequence of individual proteins. More information about the project, as well as a download link, can be found at the official website.
Related Publications:
- Timothy Luciani, John Wenskovitch, Koonwah Chen, David Koes, Timothy Travers, and G. Elisabeta Marai. “FixingTIM: Interactive Exploration of Sequence and Structural Data to Identify Functional Mutations in Protein Families,” BMC Proceedings, 8(2) (2014), S3. DOI: 10.1186/1753-6561-8-S2-S3. Journal Impact Factor: 1.46.
- John Wenskovitch, Tim Luciani, Koonwah Chen, and G. Elisabeta Marai. “Fixing TIM: Identifying Functional Mutations in Protein Families through the Interactive Exploration of Sequence and Structural Data,” in BioVis 2013 Data Contest. BioVis ’13. Atlanta, GA, 2013. [Vis Experts’ Pick for Best Contest Submission].
MOSBIE
Rule-based modeling allows for the construction and simulation of models representing chemical interactions within and between cells. The iterative development of these models presents several challenges, including keeping track of the features encoded in each model and understanding the development history of the family of models. In this work, we present the development and features of a journaling system to meet these challenges. A small multiples view displays the contact map rendering of each model in the development family. Users can interactively compare the similarities and differences between pairs of models, or search for models which contain an individual structure. A history tree view shows the development of the model family with respect to time. More information about the project, as well as a download link, can be found at the official website.
Related Publications:
- John Wenskovitch, Leonard A. Harris, Jose-Juan Tapia, James R. Faeder, and G. Elisabeta Marai. “MOSBIE: A Tool for Comparison and Analysis of Rule-Based Biochemical Models,” BMC Bioinformatics, 15(1) (2014), pp. 316–331. DOI: 10.1186/1471-2105-15-316. Journal Impact Factor: 2.213.
- John Wenskovitch, Leonard Harris, James Faeder, and G. Elisabeta Marai. “A Journaling System for Rule-Based Biochemical Models,” in IEEE BioVis Poster Abstracts with System Demonstration. Atlanta, GA, 2013.
RuleBender
Rule-based modeling allows for the construction and simulation of models representing chemical interactions within and between cells. However, rule-based modeling has several inherent disadvantages. For one, systems biologists are not software developers – spotting syntax and semantic errors in the model file may be extraordinarily difficult. Additionally, the numerical output of model simulations may be difficult to interpret. RuleBender is essentially a development environment for the BioNetGen language, containing a visual debugger, model simulator, and results viewer. The goal of this project is to facilitate rule-based modeling construction, simulation, and analysis in an integrated system. More information about the project, as well as a download link, can be found at the official RuleBender website.
