A collaborative visual analytics suite for protein folding research.

Area of Science:

• Computational Biology
• Biophysics
• Data Visualization

Background:

• Molecular dynamics (MD) simulations generate vast, high-dimensional datasets crucial for understanding biochemical processes like protein folding.
• Analyzing these massive MD datasets is computationally challenging due to their high dimensionality, making it difficult to explore individual conformations or gain a global data view.

Purpose of the Study:

• To develop an interactive, collaborative visual analytics tool for exploring large, high-dimensional molecular dynamics simulation data.
• To enable researchers to easily navigate, examine, and understand molecular conformations within massive datasets.

Main Methods:

• Developed a visual analytics tool employing a topological landscape metaphor, representing high-dimensional data as a 2D terrain.
• Integrated visualization of biochemical properties (e.g., contact density) onto the 2D terrain.
• Implemented a client-server architecture for real-time collaboration among researchers.

Main Results:

• The tool facilitates effective navigation and exploration of otherwise 'invisible' MD simulation data.
• Researchers can examine molecular conformations at both individual and global levels.
• The topological landscape metaphor aids in understanding the protein energy landscape and associated properties.

Conclusions:

• The visual analytics tool significantly enhances the interpretability of massive, high-dimensional MD simulation data.
• It provides a collaborative platform for researchers to share insights and build 'storyboards' of biochemical events.
• The software, written in Scala and available under GPL, offers a powerful solution for computational biophysics research.