John and Marcia Price College of Engineering
18 Interactive Visualization of Large Data for Cancer Cell Microscopy
Luke Schreiber and Alexander Lex
Faculty Mentor: Alexander Lex (School of Computing, University of Utah)
A key focus of cancer research is understanding how a patient’s cells evolve in response to different drug treatments. Our collaborators at the Huntsman Cancer Institute use this research to develop personalized medicine plans for cancer patients. Cancer’s complexity and the vast amount of data generated by modern microscopy methods, such as Quantitative Phase Imaging, pose significant challenges for preprocessing and analysis [1]. To address this, the Visualization Design Lab has been developing interactive visualization systems that provide both high-level overviews and detailed contextual information from microscopy images. Previous work by our lab has been published and received an honorable mention at IEEE VIS. However, our lab’s new revised tool named Aardvark shows even greater utility [2]. The focus now is on creating cohesive visualizations that integrate cell images, time-series data, and tree data, moving beyond traditionally separating views of line charts, tree diagrams, and images. This approach aims to reduce the cognitive load on researchers by eliminating the need to mentally combine disparate data types. These design principles prioritize a primary data type as a layout while seamlessly incorporating other data types.
One important factor in the development of this tool has been redesigning the cross-filtering tool to handle large amounts of data. To do this, we began working with a new framework for producing scalable data visualizations called Mosaic [3]. This tool allows us the potential to create smooth visualizations of large amounts of data. We stay in constant communication with our collaborators to learn new ways to help contribute to their workflows. Case studies have shown Aardvark’s effectiveness in analysis, communication, and data cleansing for cancer cell microscopy data [4]. By employing these new visualization design techniques, we aim to help researchers more effectively visualize, analyze, and handle the data related to cancer cell behavior in response to drugs.
Footnotes
[1] Lange et al., Loon: Using Exemplars to Visualize Large-Scale Microscopy Data, IEEE VIS, 2021.
[2] Lange et al., Aardvark: Composite Visualizations of Trees, Time-Series, and Images, IEEE VIS, 2024.
[3] Nguyen TL, Pradeep S, Judson-Torres RL, Reed J, Teitell MA, Zangle TA. Quantitative Phase Imaging: Recent Advances and Expanding Potential in Biomedicine. ACS Nano. 2022 Aug 23;16(8):11516-11544. doi: 10.1021/acsnano.1c11507. Epub 2022 Aug 2. PMID: 35916417; PMCID: PMC10112851.
[4] Heer et al., Mosaic: An Architecture for Scalable & Interoperable Data Views, IEEE VIS, 2023.