autoPACK Visualization Challenge: Present HIV in Blood Plasma
Graham Johnson, PhD, CMI
For years I've studied the artwork posted on www.cgsociety.org (a popular website and discussion forum for digital artists who work primarily in the 3D, film and game industries). I dreamed of what molecular worlds might look like when visualized with the diverse skill sets represented across both the AMI and CG Society's ~200,000 members. By late 2012, the autoPACK software (originally developed as part of my PhD thesis) could generate reliable cell-scale models assembled with molecular details, and we had refined its Graphic User Interface (GUI) enough to make autoPACK suitable for public use.
With planning and funding support from Chris Andrews and Carlos Olguin of Autodesk, I designed a contest to challenge participants to explore the mysterious world of HIV. Ludovic Autin and I supplied contestants with several choices of 3D models of HIV floating in blood plasma. The more complex models contained hundreds of thousands of molecular components that contestants interacted with via a standardized autoPACK GUI running in Cinema 4D, Maya, and Blender. To broaden the user-base, in the 11th hour Autin worked day and night with Christopher Diggins of Autodesk to provide a GUI for 3D Studio Max, followed by Softimage and PMV. In February, Tom Goddard added support for Chimera.
Ultimately, in an official CG Society art contest (autopack.cgsociety.org) run by Ballistic media, I challenged the participants "to convey humanity's complex relationships with HIV, be they emotional, political, or intellectual." I further asked them "to excite general audiences with visuals that will help our labs spread interest in the search for a cure."
The contest accomplished several major goals for the autoPACK project and fell short on others. Biologists improved the accuracy of the models via a dynamic database to help merge cutting edge science with top quality art. Many entries demonstrated how artists and illustrators could spend more time researching and planning stories and exploring aesthetics with the obstacles of modeling reduced. To truly improve the quality of the models, however, we must lure more researchers to critique and adjust the model input parameters. We hope that providing a portal to "see" their improvements in real time or through the eyes of artists will excite more researchers to participate in the future. While the general packing solutions provided by autoPACK should interest other industries (e.g. math, engineering, and physics) enough to improve the core algorithms, we only received one script improvement. This suggests that future contests should incorporate specific tasks we want to accomplish as part of the challenge itself, e.g., to have a prize for "most improved lipid packing algorithm."
On a small-scale outreach level, the contest introduced many artists who'd never heard of a "lipid" to the fascinating realm of cell biology. Exposing this group to tools like autoPACK, ePMV, mMaya, and BioBlender should have a broader impact by improving the accuracy of molecular depictions in film and TV. In turn, we in the AMI can learn techniques and approaches for molecular visualization and storytelling from the "Hollywood" style artists or find new collaborators for big production projects. Perhaps most exciting, the open aspect of the challenge inspired extreme artistic license, resulting in many metaphorical explorations that can entice general audiences to learn more about the molecular world.
1st Prize (still image excerpt). Alexey Kashpersky from Poltava, Ukraine. kashpersky.comBiology inspires Art. Alexey writes, "I have expressed in this work the pain, suffering and fear of the unknown, which with inconceivable paradox, goes hand in hand with physical beauty, light, and feelings of love and passion." Alexey has recently started working at Thomas Direct Studios.
In a very tight race for 1st place, biomolecular chemist Jiri Klusak tamed and expanded the autoPACK models with his own optimizations and packings to submit an epic physiological poster describing a portion of the HIV life cycle. Jiri writes, "The goal was to depict my inner vision of a dynamic and crowded molecular landscape and AutoPACK really convinced me [that it is] an awesome tool for this purpose."
Please visit autopack.cgsociety.org to enjoy, discuss, and study all of the entries.
Watch for future challenges, learn how the software can help researchers, or to learn more about the structure of HIV.
autoPACK developed by: Graham Johnson1,2 & Ludovic Autin2, Mostafa Al-Alusi2, David Goodsell2, Michel Sanner2 & Art Olson2
1Mesoscope Lab in qb3@UCSF (California Institute for Quantitative Biosciences, University of California, San Francisco)
2Molecular Graphics Lab at The Scripps Research Institute, La Jolla
*Thanks to Thomas Brown for early alpha testing and images used in the contest announcement poster.