The overall goal of our Molecular Dynamics (MD) experiments is to simulate the decoding center of the ribosome so we can extract useful properties that tell us something about translation. We focus on ribosome-mRNA interactions, and how those interactions change under different conditions such as codon context, stress, and growth phase.
This process begins with a static structure, a 3D model of the translocating ribosome, frozen in time (via Cryo-EM). This comes to us as a .pdb file. This file specifies the positions of most of the atoms of the entire ribosome, a truly enormous structure. To reduce computing requirements and avoid overflow issues, we simulate only the area immediately surrounding the decoding center, about 5% of the total atoms. The atoms of this subsystem have already been extracted for you (see Subsystem Initialization).
The first step of your experiment is, unfortunately, often the trickiest - making the necessary modifications to the subsystem. This begins with the default subsystem .pdb file, which you will modify according to your research question. This involves using a text editor to change the starting .pdb file to update the identity of key nucleotides of the mRNA/tRNA, add post-transcriptional and post-translational modifications to the ribosome/tRNA, or a combination of the two (see Substitutions and Modifications). At this point, you have a list of x,y,z coordinates for each atom but more information is needed to run a simulation. Some atoms may be missing and need to be added, there is no information about bonding, the system has an overall charge from the nucleotide phosphate groups and amino acid side-chains, and there is no solvent. The .pdb file you modified will be loaded into tLEaP to address these issues. From here you will have a .prmtop file, which contains the extra information about bonds and therefore forces you added in tLEaP, and an .rst file, which contains your updated atomic coordinates.
Fortunately, for much of the rest of the process, the workflow becomes much more automated. For the remaining steps, in general, there will be a python script (.py) that will require you to change the input variables specified at the top. The purpose of the python script is to generate a shell script (.sh) pointing to the MD files (.prmtop, .rst) and commands in the format required by the cluster scheduler. You will then submit the shell script to the scheduler, which will process the MD files and commands and provide your output. For Energy Minimization, Heating, and Equilibration, the key output will be an .rst file of updated atomic coordinates that will be the starting point of the next step of the process. The key output of Neutral Dynamics is the trajectories you will use in your analysis.
Using this streamlined process, you will take your structure from tLEaP through Energy Minimization. The original structure obtained from cryo-EM has an average resolution, meaning that the x,y,z coordinates of the atoms of that structure are sometimes not exactly where they should be. In some cases atoms may be placed too close together, resulting in repulsion. Energy Minimization is the process of progressively resolving these high-energy interactions, starting with the most unfavorable, by tweaking atomic positions.
It is now time to add motion to the static structure. During the Heating step, velocities are randomly assigned to each atom based on the Boltzmann distribution to bring the temperature of the system up to 300K. Because the assignment is random, each experiment that has its own heating step will be unique in the way that the motion plays out. Having multiple trajectories of the same system means that our replicates have a distribution of behaviors, which is useful for statistical analysis.
Once the system has been heated, it will need some time to settle into stable behavior. The Equilibration step allows the system to relax and waters to distribute while still anchoring all the atoms of the ribosome to their positions at the end of heating. The final step is Neutral Dynamics, so called because the residues of interest are finally allowed to move without restraint. The resulting trajectories will be the subject of our analysis of CAR behavior. The residues in the outer shell of the subsystem (the "onion shell") remain anchored to the atomic coordinates at the end of equilibration, and so are not used in analysis. Also, perhaps confusingly, we normally discard the first 20ns of a neutral dynamics trajectory because the system is still "equilibrating." The difference between equilibration proper and this initial period of neutral dynamics is that a restraint is applied to all ribosome atoms in the former and only atoms of the onion shell in the latter. The discarded part of the trajectory can be thought of as a final period of system relaxation/settling before data collection begins.
Our normal protocol is to perform 20 60ns experiments and 10 100ns experiments per experimental construct. Once we have these 30 trajectories in hand, we have various Analyses we use to understand CAR behavior. There are individual pages on each of the types of analysis we do with more details on corresponding processes and interpretation.
Continue on to Cryo-EM PDBs to learn more about the sourcing of our atomic coordinates, or skip ahead to Substitutions and Modifications to begin your experiments.
