Six Essential Improvements
For
Molecular Visualization Technology
An Essay With Examples.
L. Van Warren MS CS, AE
UAMS Biochemistry and Molecular Biology
|
Introduction |
Several breakthroughs in biochemistry and molecular biology have been achieved by building models and using those models to make crucial inferences. Among these breakthroughs have been the mapping of the DNA double helix structure, and enzyme structures. The modern day counterpart of physical models are computer models and simulation. The purpose of this essay is to identify and suggest six improvements to molecular visualization, that have the potential to significantly increase understanding.
|
1) Animation |
Biochemical communication is the richest, most complex subject on the planet. Computers are the sole hope for effective communication, since, biochemistry is first of all, an information management problem. If you think about it for a while, it becomes clear that computers are most of all, about communication. Communication via vivid moving three dimensional images is most effective. Paradoxically, animation is not used sufficiently, convincingly or frequently enough in the area where it is most needed, biochemical communication. However, there may be a "path" out of this situation.
Currently, biochemical pathways, such as those published by Boehringer Mannheim, Lubert Stryer, the author and others depict snapshots of specific molecules and transition states from a keyframe point of view. In traditional animation, a keyframe is something that an animation artist uses to show a key point or position in a story. The entire story is depicted on a large surface called a storyboard. Keyframes, whether factual or fictional, enable common agreement about the state of the story at a given moment among the parties involved in its production. After consensus is achieved between the technical and artistic director, keyframes are manually interpolated by apprentice artist labor, or more recently by computer programs to fill in the in-between frames. The keyframes and in-between frames are then linked together to form the final movie.
A common example of a biochemical "story" is the metabolism of glucose that takes place in the cytosol of most mammalian cells. This process, called glycolysis, is well understood, but difficult to transmit to the uninitiated, in a short period of time. The typical approach is to present the ten steps of glycolysis as a flowchart in 2D or 3D. As in a storyboard, we see the molecules at each stage in the process, without seeing the events that happen to a given molecule from beginning to end. These are only inferred after time consuming study.
As in all fields, we may wish to provide additional representations of a problem, without abandoning the usefulness of former ones. Animation can reduce the time required to perceive a complex idea by a factor of thirty or more. This means that we can potentially be thirty times smarter in a given unit of time.
A color animation is more expensive to produce than a black and white copy, especially when the material has to be collected from scratch. However, once an animation is produced it can be viewed by millions of people, effectively prorating its production cost per viewer to zero.
The advent of high speed computers, chemistry software and inexpensive color printers has made animation a viable alternative to rapidly thrown together Xerox copies for biochemistry communication.
|
2) Electrostatic Interactions |
We now have at our fingertips access to the geometric structures of thousands of organic compounds, including proteins, nucleic acid sequences etc. In the richness of this information however, there is a critical graphical component lacking and that is a representation of the electrostatic forces that influence the behavior of the molecule. At the nanometer scale, it is charge, and not shape alone that dictates the physics of chemical interaction. Even in drawings that we often make, the influence of charge is not well characterized. Take ATP, the principal energy storage compound of the cell for example. It is often schematically drawn like this:
A molecule of Adenosine Triphosphate
Two dimensional schematic representations such as that given above are currently the most common in organic chemistry. These flat representations excel in depicting connectivity between adjacent elements and give some hints as to the nature of the bonds, but do not provide electronic characterization. Three dimensional ball and stick representations suffer from the same electrodynamic limitations, but do depict three dimensional geometry. Even so this tells only only half the story. At the scale of chemical reactions charge is the significant driving force, not just geometry. Consider the following pair of representations. On the left we draw the traditional ball and stick model, on the right we size the ball according to its charge influence. The right drawing, while remaining inadequate, does remind us that the inorganic phosphorus groups exert a massive influence on the behavior of this compound.
ATP as ball and stick |
ATP size computed from partial charges |
At a higher level of complexity, cellular compartments such as mitochondria has on its outer surface an accumulation of electrons resulting in an abundance of negative charge. An observer on the surface might have their virtual hairs standing if they could hear the of accumulating hum and see the flash of ion lightning bolts suddenly transported through perforations in a charge field that permits local ion flow.
|
3) True Electron Density |
A third limitation of conventional ball and stick models, even space filling ones, is that the forces generated by the interaction of electron density functions are not depicted in a realistic fashion. The binding forces that link two elements together are linear only within very limited excursions, and decidedly nonlinear for large ones.
By way of a gravitational analogy consider the inverse square law forces that control the motion of the planets. Astronomers using computer animation to depict planetary motion do not use sticks to attach the sun to the planets or the planets to their respective satellites!
An improved way to depict such bonds is to show the electron density of the molecular orbitals. Molecular orbital theory is a step beyond atomic orbital theory in that the electron energies possessed by an electron are not a property of the single elements but rather are a property of the composite molecule. These density functions must depict the probability of finding an electron within at a given point in space. This probability is best represented by a glowing transparent fog whose intensity is proportional to the probability of finding an electron at that point.
Existing software does allow the approximate solution of molecular orbital surfaces such as that shown below for acridine orange:
-- one of 99 molecular orbital configurations for the fluorophore acridine orange (C) 1999 LVW |
But we seek the superposition of all the bonding orbitals such that the full hybrid electron density cloud is depicted. The resulting fog must be carefully rendered so as to produce an intelligible result. This is because the human perceptual system is very sensitive to the meanings inferred from small changes in opacity. Even a 5% variation in transparency can make a large functional and aesthetic difference.
|
4) Showing Actual Reactions |
Quite often when communicating elements of a biochemical pathway, or showing the operation of RNA/DNA polymerases and endonucleases, we show final states, and sometimes transition states. An essential improvement that is quite informative is this; Show reactions in progress, not just vibrating molecular structures. An index to the current state of the art on these matters can be viewed at the chemist's art gallery. Showing the reaction pathway in addition to showing the products, reactants and transition states (in-between frames) creates a motion pattern memory that is useful for further reasoning.
|
5) Stereo Imagery |
We are by nature built to perceive objects in three dimensions using binocular stereopsis. Binocular stereopsis is the combined optical and neural processing that results in the perception of a three dimensional image from objects whose viewing angles vary by from 2 to 7 degrees. One difficulty that this presents in a teaching environment is that it is typically the case that a single projection surface is being utilized by individuals whose distance from the screen my vary by a factor of three. Red Green glasses, polarizing glasses and virtual reality goggles are less than satisfying solutions to this problem. MIT has developed three dimensional holograms from computer displays, but these images are static at the current time. Vibrating mirrors have been used to create effective three dimensional illusions but these have been monochromatic displays with limited viewing angle.
-- a stereo view of warfarin, a blood clotting inhibitor (C) 1999 LVW |
We have ears as well as eyes. In complex situations using both senses provides additional information. Let's say that we generate a sound signal for the DISTANCE between every pair of bonded elements. That makes for one sound generator per bond. We then define a subharmonic mapping that multiply bond vibration frequencies by appropriate scale factors so that they fall within the audio spectrum... so we can HEAR them!
At absolute zero the distance isn't changing, there is silence, otherwise if the molecule is in its normal thermal energy of vibration, there is a signal corresponding to the natural frequency of vibration. As we start looking at more complex molecules, they become like orchestras. They behave in a natural sonic way. If the molecules are far from us, they are quiet, if we are close or zoomed into the action, the harmony for the ensemble of molecules is louder. We could select to hear the vibrations of only certain groups, even species that were undergoing transition.
This could be much more informative and, when combined with the techniques enumerated above and applied to molecules of biological complexity the results could be quite spectacular.
|
Conclusion |
Six essential improvements have been identified that increase the rate of understanding. Information is being produced at avalanche rates. Information is being produced faster than we can assimilate it via traditional methods. The improvements suggested by this essay may ultimately become a necessity rather than a luxury. These improvements will become more practical as increases in processor speed, network bandwidth, and computing algorithms progress. Among these computing algorithms are space and object decomposition strategies that will reduce computational loads by an order of magnitude.
|
Acknowledgments |
The author would like to acknowledge the support, review and kind criticism of the following individuals and institutions: Dr. Barry Hurlburt, Dr. Chidam Bhuvaneswaran, Dr. Kevin Raney, Marilyn Fulper-Smith, and Lynn Warren. Also The Protein Data Bank, its authors and support staff and the individuals and institutions that maintain the websites cited in this essay. All opinions contained herein are those of the author and do not reflect on those of the reviewers, the institution, its practices or its policies.