Christopher Wilmer and His Visualization of Molecular Self-assembly

The above screenshots are taken from the opening part of an animation named High Density Energy Storage Using Self-Assembled Materials created by Christopher Wilmer and his colleagues. The animation was one of the winners of 2011 NSF International Science & Engineering Visualization Challenge. Starting from a single metal ion (blue), we see organic molecules (white/grey) attached to metal ions, forming sophisticated 3D structures. The camera zooms out, revealing macroscopic blue crystals made of these organic molecules and metal ions.  And finally we see a futuristic car which is powered by natural gas stored inside these blue crystals.  You can watch the full animation below:

Many alternative fuels that may help slow down global warming suffer from technical barriers. Hydrogen and methane gas (also called 'natural gas') are both more environmentally friendly than gasoline, but contain very little energy per unit of volume.

Chris is currently an assistant professor in the Department of Chemical & Petroleum Engineering at the University of Pittsburgh, having earned his Ph.D. in Chemical and Biological Engineering from Northwestern University. He is also the co-founder and advisor of NuMat Technologies, a cleantech company that computationally designs and synthesizes high performing nanomaterials for gas storage and separation applications. An expert in advanced computational modeling, Chris was named to the “30 under 30 in Energy” list at Forbes Magazine. He was additionally honored as a “top solver” by InnoCentive, being amongst the top 16 individuals out of over 200,000 globally in the submission of innovative solutions to real-world challenges.

I am thrilled to show his work on L2Molecule. Also I am grateful he took the time to answer a few questions about scientific visualization. 

---

Q & A with Christopher Wilme

1. What are your favorite tools for creating scientific visualizations?

My favorite tool is Blender, and I try to use it for everything. The movie "High Density Energy Storage Using Self-assembled Materials" was done entirely using Blender. Sometimes I am pressed for time and so I use more specialized tools, like Visual Molecular Dynamics (VMD). When I need to make a compelling diagram in 5 minutes or less, I use Microsoft Powerpoint, which unbeknownst to many actually lets you build surprisingly sophisticated 3D structures. Other programs I use are Mathematica and 3D Coat. Of course I also use Gimp and Inkscape for simple image manipulation.

2. What were your roles in making the above animation and how long did it take to create it?

The three "directors" were Dr. Omar Farha, Patrick Fuller, and myself. At the time, I was the only one with any experience in animation, but Pat took this opportunity to learn and made several contributions to the animation (including the scene corresponding to the attached image). We came up with the idea for the animation shortly before the deadline for the NSF Visualization Challenge, and so the whole thing (animation, music, narration, editing) was done in three weeks. The research depicted in the movie was carried out primarily by Omar and myself, and so we crafted the narrative together from which I made a typical "story board" before animating anything. I wore all of the hats one could think of in the process: doing the underlying scientific research, conceiving the idea and narrative, script writing, modeling, animating, sequencing, organizing the contributions of others, etc. The only thing I played no part in was composing the music for the third scene, which was done by my brother, Alex Wilmer.

Nature Chemistry cover created by Christopher Wilmer.

Nu-100, a new metal-organic framework material. Image created by Christopher Wilmer.

3. What is the most challenging part in making this animation?

The tools for making high quality animations are not very user friendly or intuitive (at least not the ones I use). If you use them frequently, perhaps because animation is your day job, it's not as much of a problem. For someone like myself, who makes animations like this more sporadically, it's frustrating to keep relearning some of techniques. To combat this, starting next year I will be teaching a course called "Scientific Visual Communication" which will codify a lot of the techniques I have used, which will hopefully help my students (and me!) produce animations like this without so much head scratching.

4. There is a strong connection between the first part of your animation and a scene in Stanley Kubrick’s 2001: A Space Odyssey. Can you tell us a little bit about this?

For many years I have dreamed of making an animation depicting molecular self-assembly as an elegant nanoscale dance, much like the satellites and docking space vessels in Kubrick's movie. The catalyst for this animation, the NSF Visualization Challenge, provided the perfect opportunity to make that dream a reality. In many ways, the analogy between docking space ships and self-assembling molecules is a good one. Both processes are awe-inspiring, and as I argue in the animation, they each leverage extreme technological precision. The engineering and physics utilized to launch satellites into precise orbits is mirrored by the chemistry that is needed to grow macroscopic crystals that are (nearly) perfect in their atomic detail. Someday, when science has advanced further and we are able to routinely make synthetic molecular machines, I might make another Kubrick inspired movie.

5. In your opinion, how important is creating engaging visuals for scientific research?

It's more important than people think, even when taking this sentence into account. Very few people will admit to not being interested in- or not paying attention to- a scientific presentation simply because it isn't visually engaging. However, many people (students and professors alike) have come up to me after my presentations to remark that they were especially engaged because of the animations. It falls on the shoulders of scientists to best communicate their results to the outside world; we shouldn't expect others to decipher our cryptic research if it hasn't been communicated well. 

This visualization shows the absorption of methane molecules (orange spheres) inside different MOF (metal-organic framework) materials. Image created by Christopher Wilmer and Patrick Fuller.

Irving Geis and His Paintings of Proteins

“Crystal structure of myoglobin (1961)” from the Irving Geis Collection. Rights owned and administered by the Howard Hughes Medical Institute. Reproduction by permission only.

The image above was a painting of myoglobin, the first protein structure solved by X-ray crystallography. The painting was created by Irving Geis for a Scientific American article “The Three Dimensional Structure of a Protein Molecule” by John Kendrew, published in December l961. John Kendrew and his colleagues solved the myoglobin structure in l958.

Irving Geis (1908-1997). Photo: Sandy Geis.

Nowadays, one can easily create an image of a protein structure with the aid of a computer and molecular visualization software. In 1961, however, everything had to be done by hand. Creating an image of a protein structure required not only outstanding artistic skills of visualizing complicated 3D structures, but also extraordinary patience. Originally trained as an architect at Georgia Institute of Technology and receiving a Bachelor of Fine Arts from University of Pennsylvania, Geis had all the skills and knowledge to visualize the 3D structures of proteins.

Geis created this painting by first photographing the physical models and then by creating voluminous sketches and studies before painting the finished version [1]. A lot of refinements were made during the sketch step based on the feedback from John Kendrew, as shown by the image below. The final painting took 6 months to complete [2].

“A draft image of the 1961 myoglobin” from the Irving Geis Collection. Comments on the image were made by Irving Geis and John Kendrew. Rights owned and administered by the Howard Hughes Medical Institute. Reproduction by permission only.

I tried to mimic Geis’ myoglobin image by using UCSF Chimera and Maxon Cinema 4D. It took me quite some time but the result is far inferior to the original masterpiece. The heme group seems OK, but the alpha helices are hard to recognize even though I used the same color scheme as Geis’ painting, depth cueing effect, and some lighting techniques in Cinema 4D. In addition, I couldn’t add hydrogen atoms except those in hydrogen bonds in my version, as the model would become overwhelmingly complicated to be meaningful at all. On the other hand, Geis’s painting clearly visualizes the main structural features (the heme group and alpha helices) while giving an overall sense of the structural complexity of the myoglobin protein.

My attempt to mimic Geis’ painting using UCSF Chimera and Maxon Cinema 4D (PDB ID: 1MBN). The alpha helices are difficult to recognize because they overlap with the atoms in the back.

To achieve this, Geis used a process he called “selected lying” [2-4], in which he made small adjustments to avoid structural overlapping. He might distort the protein a little bit here and there, or he might use slightly different viewing angles or perspectives for different parts of the protein. In the end, this process resulted in a structure that was not so different from the real structure, but much easier to understand on a flat paper. A computer, on the other hand, draws everything based on the given coordinates and the image it produces is usually not very comprehensible, especially for complicated protein structures.

“Crystal structure of lysozyme (1966)” from the Irving Geis Collection. Rights owned and administered by the Howard Hughes Medical Institute. Reproduction by permission only.

Besides his excellence in visualizing complicated 3D structures, what makes Geis’ paintings special is his belief that “his job was not to draw a protein exactly as it was, but to show how it worked”[3]. Often Geis would add additional layers of information on top of protein structures, making protein functions and mechanisms understandable. This is the reason why his protein paintings are still appreciated and used in some textbooks, even today.

Colors are very important to Geis’ painting. He carefully chose the proper colors to illustrate the inner workings of proteins. “Color is a language”, he said, “and as with any other language, one mustn’t babble!” [3]

“Cytochrome C (1989)” from the Irving Geis Collection. Rights owned and administered by the Howard Hughes Medical Institute. Reproduction by permission only.

Geis had worked with many scientists in his career. Among them, Dr. Richard Dickerson was his long time collaborator and friend. The two co-authored several books, including The Structure and Action of Proteins (1969), Hemoglobin (1983), and a chemistry textbook: Chemistry, Matter and the Universe (1976). The Structure and Action of Proteins became classic after published and inspired a generation of young biochemists.

Irving Geis was born in 1908 in New York City and died in 1997. He was 88 and lived in Manhattan.

“Irving Geis and his work-in-progress 1961 myoglobin painting” from the Irving Geis Collection. Rights owned and administered by the Howard Hughes Medical Institute. Reproduction by permission only.

---

Interview with Dr. Richard E. Dickerson about Mr. Geis

1. It’s rare that an artist shares the authorship with a text writer. However, you and Mr. Geis coauthored 3 books together [The Structure and Action of Proteins (1969), Chemistry, Matter and the Universe (1976), Hemoglobin (1983)]. As you wrote, “I could describe what needed to be illustrated about protein structure, and Irv would come up with clever and original graphic methods of putting the point across. It was never clear whether Irv illustrated my books, or I wrote Irv’s captions.”[3] Could you please share with us how Mr. Geis was able to get such a deep understanding of proteins?

It is my understanding that he acquired his background knowledge slowly by illustrating for magazines such as Scientific American.  Indeed, it was his illustrations for John Kendrew’s myoglobin structure that first caused us to meet in New York in 1963.  He was a very intelligent man, and simply read a lot about scientific matters in order to make his diagrams more meaningful.

2. In you books, I am really amazed that Mr. Geis was able to depict sophisticated protein structures and functions with only black/gray and another color. The images are not only easy to understand but also elegant. Could you please comment on this?

Irving Geis was a true artist.  A skillful artist can convey information that a simple draftsman cannot.  His use of color and shading for emphasis was masterful. 

3. As you wrote, Mr. Geis used a process he called “Selective Lying” to tweak the protein representation if “some key aspect of protein structure was eclipsed and out of sight”.[3] The end result is a protein structure and its molecular mechanism easy to understand. Personally, I am firm believer of this approach. But others might think we should not alter the structure at all. What is your opinion on this?

The purpose of Irv’s scientific drawings was to convey information about how the molecules work, and not simply to illustrate the fine-structure details of the molecules.  For the latter, one should turn to the Protein Data Bank (PDB) or other such archives.  If a certain amount of artistic licence would aid the reader in understanding how the molecule worked, then such behaviour is not only legitimate; it is constructive.

4. As our knowledge of protein structure and function grow rapidly, computer animation becomes a common media to depict the dynamic protein behaviours in the cell. Compared with traditional protein illustrations, what are the pros and cons of computer animations? What would be the basic requirements for artists/scientists creating these animations?

I have no strong opinions on this.  If computer animation is available to a research worker or a student, then it obviously could assist him in understanding the macromolecule.

5. You mentioned in an article[5] that you regretted you were not able to write Atlas of Protein Structure with Mr. Geis at a time when there were only 8 protein structures. Nowadays, we have nearly 100 thousands protein structures in the Protein Data Bank. Do you think it is still necessary write a new Atlas to give students and hopefully the general public a feel of the richness and magic of protein structures? If you do, how many proteins should be included in this challenging book?

I think that the days of an atlas that contains every known structure are over.  No one can write a book that discusses 100,000 different protein structures.  The PDB is a resource that allows people interested in a certain class of protein to learn how many such proteins have been solved and what each of them looks like.  It is true that one could write a general book on protein structure/function relationships, and indeed some have done so.  Such authors would select those of the 100,000 different structures that pertained to the matter at hand, or a set of structures that gave a general impression of protein structure.  But to discuss all 100,000 structures individually:  No.

6. Finally, what is your favorite illustration(s) of Mr. Geis and why?

I simply could not come up with one particular “favorite” illustration; Irv has done so much and done it so well.  I like his haemoglobin/myoglobin and his DNA illustrations, because these concern subjects in which I had a scientific relationship.  But someone else would undoubtedly give you a different list. 

---

Acknowledgement

I would like to thank Sandy Geis for reviewing this article and generously providing many reference materials about Mr. Geis’s works. I would also like to thank Howard Hughes Medical Institute for granting permissions to use the above images from the Irving Geis Collection. Finally, I am grateful that Dr. Dickerson took the time to answer my questions.

Rreferences

[1] S. de Chadarevian, Models and the Making of Molecular Biology, in Models: The Third Demension of Science (eds. S. de Chadarevian and N. Hopwood) 339-368 (Stanford Univeristy Press, Stanford, California, USA, 2004)

[2] B. P. Gaber and D. S. Goodsell, Irving Geis: Dean of Molecular Illustration. Journal of Molecular Graphics and Modeling 15, 57-59 (1997)

[3] R. E. Dickerson, Irving Geis, Molecular artist, 1908-1997. Protein Science 6, 2483-2484 (1997)

[4] D. S. Goodsell and G. T. Johnson, Filling in the Gaps: Artistic License in Education and Outreach. PLoS Biology 5, 2759-2762 (2007)

[5] R. E. Dickerson, Obituary: Irving Geis, 1908-1997. Structure 5, 1247-1249 (1997).

Roger Hayward and His Pastel Drawings of Molecules

Roger Hayward (1899-1979) was an architect, artist, scientific illustrator, and inventor. Today, he is probably best remembered for the pastel drawings he created for The Architecture of Molecules, a book coauthored by Linus Pauling and Hayward, published in 1964. Hayward drew 57 beautiful plates and Pauling wrote easy-to-understand texts accompanying each plates. The book introduced various topics in chemistry to the general public, including atomic structure, structure of small molecules, crystal structure, and protein structure. It was one of the first books in which art and science are perfectly blended together. Nature magazine called it “a fascinating work of art…”

Roger Hayward (1899-1979). Image source: Special Collections & Archives Research Center, Oregon State University Libraries.

Collaboration between Pauling and Hayward

The collaboration between Pauling and Hayward started in the early of 1930s. During the years, they collaborated on the figures of many scientific publications and Pauling’s General Chemistry textbook published in 1948. Hayward was no ordinary artist as he tried to understand what he illustrated beneath the surface. For this reason and his skills to visualize 3D structures, Pauling had a lot of confidence in Hayward’s abilities to illustrate molecules and called him a scientist. The two enjoyed an evolving friendship throughout the years and The Architecture of Molecules was the apex of their collaboration.

Top: Pauling’s draft image of the proposed structure of feather rachis keratin; bottom: Hayward’s drawing appeared in the final PNAS publication (PNAS 37, 265-261, 1951). Image source: Special Collections & Archives Research Center, Oregon State University Libraries.

The Visual Aspect of The Architecture of Molecules

Roger Hayward’s pastel drawings for Linus Pauling. Image source: Special Collections & Archives Research Center, Oregon State University Libraries.

What made the book so special was the use of pastel as drawing media. Most molecules in the book are foreign to the general public. The pastel rendering of these molecules, however, brought a sense of familiarity to readers. Hayward also paid a lot of attention to the colors of the molecules. Although the color palette are based on the one used by scientists, Hayward made some subtle changes so that the colors are soft and pleasant. The pastel rendering and soft colors made the molecules look like something we saw in dreams.

Roger Hayward’s pastel drawings for Linus Pauling. Image source: Special Collections & Archives Research Center, Oregon State University Libraries.

In these beautiful plates, Hayward also displayed his extraordinary skills for visualize complicated 3D structures such as crystals. We have to remember that these drawings were done long before the availability of computer graphics. 

Roger Hayward’s draft image on the crystal structure of Ice II. Image source: Special Collections & Archives Research Center, Oregon State University Libraries.

Some drawings like the crystal structure of Ice II might take a lot of effort to make them scientifically accurate.