HomeArticlesVisual Communication in Biology 2: Animating Molecular Biology, Part I – Janet Iwasa (U. Utah)
Visual Communication in Biology 2: Animating Molecular Biology, Part I – Janet Iwasa (U. Utah)
November 4, 2019
Hi. My name is Janet Iwasa. I am in the Biochemistry Department at the University of Utah. And I’m going to be talking to you today about animation. I’ve worked with a lot of different researchers over the years to depict the kind of molecular and cellular processes they study, and I’ll be providing you with a bit of an overview of what that process looks like. So, first I wanted to show you just a couple of animations, to show you kind of a range of different ways animations can be made. So, this is an animation I made very, very early on in my career that depicts a bacterial actin-like filament that’s polymerizing. This was made with in collaboration with Ethan Garner and Dyche Mullins. And so this is a pretty basic animation, and more recently I’ve been creating more complex animations, like this one. This one shows a HIV that’s budding from a cell. So, both of these animations… basically, despite there being pretty major differences in the complexity of the animation, how many proteins are involved here… basically, the steps to create these animations are the same. So, what are these steps? So, I’m going to walk you through the general animation process. So, the first step — and this is sort of similar to when you’re creating any sort of model figure — is really describing the process. What’s the story? In animation, you really have to do this in much more detail than you would if you were just drawing a figure. So, you need to ask the questions of, you know, what’s happening?, where is it in the cell?, how many proteins are there?, and in particular, what structures or other data may be available to tell that story? So, one thing to note is that not all processes are great for 3D animation. So, in general, the things that you think about that are really good for animation are processes where we really have a very deep mechanistic understanding of what might be happening, or at least an idea that we wantto put forward that’s very, very detailed. It’s also great when we have… most of the structures are available, so that we can use those directly into the animation. And where we’re also showing something that changes significantly over time. Some of the things that may be less great for 3D molecular animations are processes where there’s a lot of uncertainty, where we’re not quite sure where things are happening at different… at multiple different stages of the process. And also, where we’re providing maybe a high-level overview of a process. Because the nature of the molecular animation, of 3D molecular animation, is to show a great amount of detail. So, it’s kind of harder to show a general overview when we’re really kind of digging pretty deep like that. And so for those types of processes, it might be better to consider creating a 2D illustration, or even a 2D animation, to animate those types of processes. Alright. So, once you’ve described that story, next you have to think about your audience. Who is this animation for? In general, when I work with other people, with my collaborators, there are two major types of animations that we create in sort of the research sphere. One is animations that are created for supplemental figures. And a second is animations that are made for kind of a more general audience, for use both in talks as well as online. So, when you’re thinking about an animation for a supplemental figure, you may have different considerations than kind of a broader audience. So, in supplemental figures, you usually have a pretty targeted audience, that has probably a good knowledge of the background that’s needed to understand that figure. So, as a result… and you want to really convey ideas as quickly and efficiently as possible. So, in general, for supplemental figures, I create pretty minimal animations. Not a lot of background, not a lot of context is generally needed. Another thing to think about is that the color scheme that you use in these animations should match the other figures. So, if you have a Figure 2 somewhere that shows, like, a model figure, the animation should really match. It should be the same color scheme. And other kinds of elements should be really passed from a figure or a description in the paper to that animation. And in general, these kind of supplemental animations… figures for… anim… animations used as supplemental figures may or may not require narration. So, this is an example of an animation that was created for a supplemental figure. This was made for a review on chromatin remodeling. And what we have is this animation where, again, I’ve used the same kind of color scheme that’s used in the rest of the paper. So, really, the viewer can look at those figures and then watch the animation, and everything is relatively seamless. We also use some kinds of vision… some of the visual representations that were used in other parts of the figure. For example, we had this gray axis that marks the central point of this nucleosome. And so that’s used in other parts of the figures in the review as well. So, these are all things to consider. And again, here, the background is pretty minimal. There’s not a lot of storytelling that’s going on behind the scenes there. So, when you’re thinking about an animation that’s really for broader audiences, this may be both… either for a talk or for posting online, different considerations have to… have to play a role here. So, one of the things you have to think about is that audience. So, if you’re trying to appeal to a high school student, a member of the general public, you really need to tell a story well, in an engaging way. And so this usually means that you have to try to make it as visually appealing as possible. Adding context and background will really help people understand the story better. And oftentimes these are going to be narrated animations, so the timing needs to match the speaker’s narration. So, an example of this type of animation is shown here. So, this is an animation of gene editing using CRISPR. And what you can note is we started with an entire cell. We’re zooming into the nucleus. And we’re seeing all this kind of chromatin in the background to give a little bit more of that story. And there’s also a lot of labeling here, to make sure that the viewer is able to really follow what’s happening. So, here Cas9 is recognizing a piece of that DNA and then cutting it. And then a new piece of DNA is being introduced. And so the labeling helps… this also… this animation also typically is run with narration as well as with music. And so that really helps, I think, convey… make the animation a little bit more memorable for different viewers. So, next, once you’ve kind of gone through the process of thinking about your story — what is it going to tell? — and thinking about your audience, next you want to create a storyboard. And a storyboard is similar to, probably, what you’ve heard that animation studios do, but on a much, much smaller scale. So, storyboards are really… it’s kind of a way of drawing out what happens at each step of the animations. And the ideal storyboard should really walk through every single step. Every single major change in the animation should be drawn out in the storyboards. In general, the storyboards I create have three parts… two to three parts. One is a drawing, the sketch of what’s going on. The second is a description, just written in text, of what’s happening. And then usually, depending on whether the animation has narration, it’ll have that narration as well. So, three parts. They… so, this is an example of a storyboard. This was drawn by Dyche Mullins on a collaborative animation that we’re working on. So, you can see here, he’s drawn in kind of a comic book style every step of this animation. He’s used arrows to depict both, kind of, steps in time as well as movement. He’s done a lot of labeling, and he’s also written this description. So, you know, you can… I’ve made this kind of a template to be able to draw these storyboards, and you can create these pretty easily yourself, or download them online for free. This is another example of a storyboard. So, many of my storyboards are collaborative. The animation projects are collaborative. And you know, one of the major points of doing a storyboard is really to save time on the animation process. By defining every step of the animation, by showing that to your collaborators, with your team members, you’re hoping to really minimize the amount of changes you have to make once you’re in the animation software, because that can get super time-consuming. And so you’re trying to address all of the problems kind of as early as you can. And so, in this particular storyboard, which covers the integration step of HIV infection, I’ve shared this storyboard, which I’ve created with a collaborator who’s an expert in the field, and he has gone ahead and made notes in red about things that he would change or alter in the storyboard. And so I can go back and change that before I go into the animation software. So, it’s really about saving time and being efficient. So, next, after you create the storyboard, the next step is to record a narration. My general recommendation is to write a narration and record it no matter whether you’re planning on using a narration in your animation or not. And this is because usually, you know, if an animation will get used in a talk or something like that, it needs to match the timing of the speaker. You need to know whether, you know, basically, a long time is going to be spent on this part, describing this process, so that the animation, you know… you know basically what the timing is going to be. Another thing to keep in mind is that animations are short. They’re typically… especially molecular animations. They typically are less than a minute in length, and so you have to be pretty concise. There’s no need to describe everything within the narration. Often, you can do that outside of the context of the narration of the animation. And you also want to minimize dead time. You want to minimize time where you have some narration that’s going and nothing is happening on the screen. So, it’s really a good idea to try and really define the narration and the animation hand-in-hand. So, writing out a narration and recording it is pretty much a… should be a very, very early step when creating an animation, before you go into the animation software. The next step is creating the molecular models. There are a couple different major softwares that you… allow you to basically export a molecular model from, for example, a PDB, and import them into various 3D animation software. So, my favorite software to use, the one I’m most familiar with, is UCSF Chimera. That’s free and free to download. And a lot of other people also can you PyMOL. So, those are the two most common PDB viewers that can be used to also generate 3D models. So, I’ll be talking you through my process of exporting molecular models from UCSF Chimera in just… in just a couple of slides. Another thing to note is that you can also import molecular models using a number of different plugins for different 3D software. So, I’ll be talking in the next video about, really, the animation process and different software you can use. But I’m just noting here that you can also import things like PDBs using a number of these different plugins, including ePMV, and these are for specifically… for some specific types of 3D animation software. There’s also MolecularMaya and BioBlender. So, those are things that are worth looking into if you’re using these specific softwares, and you want kind of a shortcut to importing PDBs. Alright. So, let’s jump into Chimera, and I’ll show you the process of importing a molecular structure and exporting that to be animation-ready. UCSF Chimera is a… you can just check out the website online and download a version that works for your computer. And then, we’ll go over to the PDB. So, this is a database of molecular structures. And in this case, we’re looking for a structure of actin, an actin filament, that we’ll be animating later. So, I’ll go ahead and do that search. And you can see there are a number of different hits that we have here. So, we can go ahead and select one, and take a look at this. So, this one is called 6BNO. And that is what we’re going to put into Chimera. So, here’s Chimera. So, what we’re going to do is Fetch by ID, and put in that PDB ID here. And that’ll give us the structure, that it just fetched from the web. And I’ll color this using a preset so that the different chains are different colors. So, you can see we have eight different actin monomers here. And those are also listed on the PDB. So, you can see chains A through H are all actin, actin monomers. And so what we’ll want to do in Chimera is to use a command to create surfaces on this. So, these are by default shown as ribbons. And typically for a molecular animation, you don’t really want to export things as ribbons. It’s just really kind of a lot of geometry. And in general, if you have a lot of molecules, it’s a lot easier to visually distinguish them if you… if you use a surface. And so what I use to create surfaces, typically, is a command called molmap. And so you can… you can find where the command line is by using this Command Line in the Tools. And then the command line appears down here. And so the molmap command… you can learn more about it using the documentation that’s available with Chimera. So, basically, what you have to do is, in the command line, write molmap and then specify a specific atom, and name a… and also create… include a resolution, a number for the resolution you want to use. And so I’ll just show you how that’s done. And so, in this case what we’ll do is we’ll create a surface for chain A, which is one of the monomers, at resolution 8, which from experience I know is a pretty… pretty good resolution for a lot of the things I do, for the animation we’re creating here. So, you can see if we change this number to a lower number, you get sort of a bumpier, more detailed surface. And a higher number gives you a less detailed, kind of smoother surface. So, we’re just gonna go back to 8. And then I’m just going to repeat this command on each of the individual chains so that we have a surface for each of the actin monomers in this filament, A through H. And so now you can see that we have the actin filament, this partial actin filament here, that has surfaces for each monomer. So, now what we want to do is export this as a surface. So, you can Export Scene, and Chimera has options that include things like OBJ, which we can use to directly import into the 3D animation software. So, I’m going to call this actinfilament, and then I’ll also include the PDB for reference, and save that. So, something else you can do using Chimera is to take a look at some larger structures. So, in this case, I’m taking a look at this human adenovirus structure, 6CGV. So, we can import that into UCSF Chimera… 6CGV. And what you might notice is that this is not the entire… the entire capsid that we saw in the structure over here. So, one thing to note about some of these structural files is that you can have some information that’s encoded inside the PDB that you might not see immediately in Chimera. So, in this case, we have some symmetry, information about symmetry, that’s included in the PDB, and that we can access in Chimera. So, in this case, what we want to do is use the molmap command, again using chain A, maybe we’ll do a resolution of 10. And in this case, we’re going to tell molmap to use the symmetry information that’s included in the PDB to create surfaces that include information about the entire capsid. And so, there are… there are actually a large number of chains here, so I’m just going to… you can actually also indicate to create one surface using a number of chains, so, in this case, A through W. We’ll take a look at what that looks like. So, now we have the entire capsid. So, we’ve basically created a surface over every single chain here. And we can export this now. So, we can add a little bit more detail here. And now we can export this entire thing, again using an OBJ. So, now we’ve created the molecular model, and we’ve exported it from Chimera so that it’s animation-ready. And the next step is really going into the animation software and animating everything to show what the process… the process that’s occurring. And so this step and the next one I’ll be showing in the next video. So, the last step after you create this animation is to do compositing. So, the important thing to note is that animation software… typically what you do at the last stage is you render out a series of frames that you can then use compositing software to create a video out of. And at this stage, you’re also typically adding things like audio — so, if you have music, if you have narration. You can also add things like labeling, titling, pauses. Things like that can all be added at the compositing step. And those things… those two last steps, steps 6 and 7 are things that we’ll be talking about in the next video, so please stay tuned for that.