Mercury betrayed by Venus
Should we compromise accuracy in scientific animations for the sake of beauty?
A particular group of people including bioinformaticians, data visualizers, graphic designers, medical illustrators, artists, biologists, and engineers, from both academia and industry gather for an unconventional conference. We are sitting at the Klaus Tschira Stiftung (KTS) Auditorium at the heart of the EMBL Advanced Training Centre, in Heidelberg, for the 3rd EMBO Conference on Visualizing Biological Data (VIZBI 2016). We are about to start our discussion on scientific animation and the level of accuracy that they should have, while achieving their main objective – transmitting a clear message to their target audience.
Our discussion starts.
“Natural beauty is universal” says Monica Zoppè, from the Scivis group at the CNR in Pisa (Italy).
Perhaps we should start by defining beauty? Beauty can be defined as a quality of an object, place or concept that should provide us with an experience of pleasure or satisfaction (at least according to Wikipedia). We do not discuss whether a flower is beautiful or not. It just is. Even those that are carnivorous... Well, beauty in different contexts may be a question of taste.
When we look at the work of David Goodsell, Professor of Molecular Biology at the Scripps Research Institute, his watercolours seem to represent entire compendia of biological information on bacteria, cellular walls, and structures of proteins and so on. These beautiful compendia are, by definition, the result of a great effort in their compilation and translation into something that aims at pleasing the eye.
But, how accurate should scientific visualization be? “There are always compromises in the process of scientific visualization”, said David in a recent interview.
How we obtain the information used to build these visualizations has to be taken into consideration. For instance, protein structures used in scientific animations are usually taken from the Protein Data Bank (PDB). In PDB, more than 85% of the structures were collected using techniques such as X-ray crystallography. However, this method gives us very rigid representations of proteins, taken out of their natural environments. According to McGill (2008), proteins are quite dynamic. Among all possible dimensions, including time, proteins may show several “personalities” in their path to explore the thermodynamic world (Henzler-Wildman and Kern 2007). They could jump, run, or even “scream”, depending on how excited they might feel in their natural environment. Or, they could stay sitting “in a corner” (folded state). Thus, the dynamic side of proteins has been frequently left out of the animation. Often, only a static protein is represented.
“I hope I wasn’t the only one reading my recent paper,” sighs Stuart Jantzen, amongst the audience. Stuart, from ScienceVis Jenkinson Research Lab at the University of Toronto (Canada), was referring to the Transparency in film featured on Nature Methods (Jantzen et al. 2015). There is a lack of a reference system informing the audience about the context in which an animation is produced and whether it truly reflects scientific knowledge.
Another question raised was whether we should have minimum standards to be used by the community of scientific animators and artists. Minimum standards, like the ones scientific publishers have when researchers submit a manuscript to be published in Nature or Science, for instance. Instead of font style, size and number of pages permitted, we could define that in scientific animations color would be used in a certain way to distinguish certain structures. Using a common set of rules could, in principle, set a standard of quality that is usually associated to any scientific publication or work.
Meanwhile, Vincent Bos, from Demcon/Nymus 3D was helping to lead the discussion. We cannot forget that scientific animations are usually a product – ordered by clients - who usually have a clear idea of what they want to get. And animations are subject to a budget that dictates the precision and quality of the final product.
The animation usually starts in the storyboard, on a piece of paper. The story of how the animation goes is a process discussed between the company making the animation and the client – in an iterative manner. Well, not too iterative, because if the client changes his, or her, mind too many times, the budget will suffer. However, it is during this iterative process that a clarification of the message also takes place. This simplification of the animation, leaving some details aside, is done for the sake of conveying the right message – the message the client wants to pass to his audience.
David Goodsell has his own style: to some, it resembles comic book style,, very colorful with some black outlines here and there, although he tries to avoid a very pronounced line. His watercolors make me think of Alphonse Mucha, the Czech art nouveau artist. Beautiful – to many, that is. And every artist or studio will have their own interpretation of the information on which they base their animation. In the end, the quality of the final product will be assessed by one simple question: has the audience understood the message?
Henzler-Wildman K., Kern D. (2007) Dynamic personalities of proteins. Nature 450(7172):964-72.
Jantzen S. G., Jenkinson J. McGill G. (2015) Transparency in film: increasing credibility of scientific animation using citation. Nat Methods 12(4):293-7.
McGill G. (2008) Molecular movies... coming to a lecture near you. Cell 133(7):1127-32.