Biomedical animator Drew Berry from the Walter and Eliza Hall Institute of Medical Research presents his methodology for creating scientifically accurate and visually engaging animations of complex cellular processes. He details a production pipeline that leverages video game technology, combining data from scientific literature, the Protein Data Bank (PDB), cryo-electron microscopy, and molecular dynamics simulations. Using Autodesk Maya for modeling and the Unity game engine for real-time rendering, his team creates vast, interactive molecular landscapes.
The talk provides detailed case studies on animating components of cellular respiration, including ATP synthase, Enzyme Complex II, and the pyruvate dehydrogenase complex. Berry explains his techniques for balancing realistic Brownian motion with narrative clarity, using procedural animation rigs to control molecule behavior. He concludes by highlighting the transformative impact of new AI-powered tools like AlphaFold, which can now generate complex protein models in minutes, a task that previously required weeks of manual reconstruction, heralding a new era for molecular visualization.
Visualizing the Invisible: Animating Molecular Biology
00:00:01 - Introduction: Drew Berry, a biomedical animator at the Walter and Eliza Hall Institute, introduces his work creating animations of cellular processes like cancer and malaria to make them accessible and understandable.
00:00:39 - Dual Inspirations: His career path was shaped by a background in cell biology and microscopy, combined with a passion for computer graphics that began with the Amiga computer and video games in the 1980s. He specifically credits early games for introducing him to the concept of seamless animation loops, a technique he still uses.
00:04:36 - The Game Engine Pipeline: Berry outlines his animation workflow. It begins with intensive research, gathering data from scientific sources like the Protein Data Bank (PDB) and molecular dynamics simulations. This data is integrated and animated in Autodesk Maya and then exported to the Unity game engine for real-time rendering, lighting, and final capture.
00:08:15 - Capturing Realistic Motion: To accurately portray the natural movement and vibration of molecules, he uses molecular dynamic simulations. These simulations provide the basis for the behavior of molecules like ATP and coenzyme Q in his animations.
00:10:34 - Case Study: ATP Synthase: He demonstrates how he built the animation for ATP synthase by combining multiple layers of data, including its 3D structure, conformational changes, and electron microscopy data showing how the enzymes organize into pairs and rows within the mitochondrial membrane.
00:18:31 - The Challenge of Brownian Motion: Berry discusses the artistic and technical challenge of representing the random, chaotic movement of molecules. He explains the need to balance true randomness with controlled, repetitive patterns to ensure the animation is understandable to a human audience.
00:26:27 - Technique for Controlled Randomness: He reveals a key technique for animating molecules: using procedurally generated curves (like Maya's Paint Effects) as motion paths. This allows molecules to appear as if they are moving randomly while giving the animator full control over their timing and destination for key reaction events.
00:27:49 - Case Study: Pyruvate Dehydrogenase: This animation was particularly challenging due to the enzyme's massive size and flexible tethers, which are not captured in static PDB data. The reconstruction required extensive research, crediting scientific illustrator David Goodsell's work as a crucial "roadmap."
00:34:43 - The Future is AI: Berry highlights the revolutionary impact of AI-driven protein prediction models like AlphaFold. He demonstrates how AlphaFold generated a superior and more detailed model of a complex enzyme in minutes, a task that had previously taken him weeks to reconstruct by hand. He sees this as a game-changer for the future of molecular animation.