A classical lay summary by Dr. Ragothaman M Yennamalli, from SASTRA University.
Have you seen viral videos on TikTok and X/Twitter that show someone zooming in and a new image is seen, like the quantum world of Ant Man movies? Imagine that happening in both directions. We refer to the abstraction of grouping a set of atoms as one entity as graininess in the biophysical world. Graininess in biomolecules is useful when we attempt to simplify a complex system. For example, the amino acid Lysine consists of 24 atoms (inclusive of hydrogen), but if we know that depending on the local environment, some lysines may not have space to move around and it is the same if we consider the alpha carbon as a representation of Lysine. Thus, by making all the atoms of a molecule explicit (including hydrogens) is called a fine-grained model, and abstracting to (in this case) alpha carbons is called coarse-grained model. Some refer to this new entity representing the amino acid as a “pseudo atom”.
Molecular simulation involves mimicking the real world scenario of the biomolecule in a virtual setup. Couple of decades ago, when computational power was expensive, biophysicists would usually face the limitation of running their simulations for a long time scale. Here, the time-scale refers to picoseconds to nanoseconds. But, proteins undergo conformational changes and exhibit function at microseconds, milliseconds, and in some cases in the time scale of seconds. One of the ways to overcome this limitation is to coarse-grain your biomolecule. This technique allows modeling large molecular systems that test the capabilities of computational infrastructure. In other words, coarse grained models allow a biophysicist to simulate large systems within a short period of time with limited computational resources.
The popular coarse grained method used by biophysicists is the Normal Mode Analysis, where the system is made of alpha carbons of a protein and connected by “springs” and the biomolecule has inherent harmonic displacements. In other words, if two people are holding a rope and there is a sinusoidal vibration along the rope, where one moves the rope up and down, one could see vibrations ranging from longer amplitude (or low frequency) to shorter ones (or high frequency). These are called modes and the low frequency modes (also called as slow modes) are seen to be biologically relevant and as close as the real conformational changes that the molecule undergoes. Thus, they give information about the biomolecule’s fundamental characteristics. These slow modes are described with various adjectives like titles, twists, open/close, and many more.
Given all the above, coarse grained modeling does have limitations. For example, in many cases the biomolecule has large non-harmonic (or anharmonic) low frequency behavior due to solvent and uneven energy landscape. Do coarse grained models capture these? The answer is that they are good approximations and need to be verified. Depending on the biomolecule, either one slowest mode is enough to explain the conformational changes and in other cases a couple of slow modes are needed to get the “full picture”. Nevertheless, coarse grained modeling has been applied to ribosomes, GroEL, and other large systems.
References:
1. Ma J. Usefulness and limitations of normal mode analysis in modeling dynamics of Biomolecular Complexes. Structure. 2005 Mar;13(3):373–80. doi:10.1016/j.str.2005.02.002
2. Wang Y, Rader AJ, Bahar I, Jernigan RL. Global ribosome motions revealed with Elastic Network Model. Journal of Structural Biology. 2004 Sept;147(3):302–14. doi:10.1016/j.jsb.2004.01.005
3. Bauer JA, Pavlović J, Bauerová-Hlinková V. Normal mode analysis as a routine part of a structural investigation. Molecules. 2019 Sept 10;24(18):3293. doi:10.3390/molecules24183293