I carefully looked at the trajectory (VMD movie) and saw there are two flexible terminii fluctuating with a very high amplitude contributing to the high value of RMSD. Should I report this problem with this RMSD value?
Well it depends on the model and on what you mean by "report this problem." If the model you built was constructed by extracting the protein from a larger structure or complex (ie. truncating it), then the large movement is likely to be an artifact. (Here I am assuming that the RMSD was computed with reference to the initial input conformation.)
If the model represents the protein in its entirety and uncomplexed to another biomacromolecule, then it is “possible” that such large motion exist in the termini, especially if the terminii have no secondary structure and composed of several amino acids - however, it does not prove that it can be experimentally observed. The question now comes down to if the observed motion is statistically significant - to answer that you have to perform several MD simulations to see if the results converge to a similar observation. One other thing to consider is if the the terminii rearrangement/movement has an impact on what you are interested in (e.g. a ligand binding site far from the terminii). Rigorously speaking, again you have to consider if their impact, or lack thereof, is reproducible.
Ethically speaking, if you mean to report this research to a peer-reviewed journal, then you should make note of this finding and that you have carefully considered it. It could provide the reviewers with important information concerning the methodology, analysis, and drawn conclusions. It is always okay to report findings if they are relevant to the topic, but it is not okay to purposely neglect the reporting of unfavorable relevant results.
Well it depends on the model and on what you mean by "report this problem." If the model you built was constructed by extracting the protein from a larger structure or complex (ie. truncating it), then the large movement is likely to be an artifact. (Here I am assuming that the RMSD was computed with reference to the initial input conformation.)
If the model represents the protein in its entirety and uncomplexed to another biomacromolecule, then it is “possible” that such large motion exist in the termini, especially if the terminii have no secondary structure and composed of several amino acids - however, it does not prove that it can be experimentally observed. The question now comes down to if the observed motion is statistically significant - to answer that you have to perform several MD simulations to see if the results converge to a similar observation. One other thing to consider is if the the terminii rearrangement/movement has an impact on what you are interested in (e.g. a ligand binding site far from the terminii). Rigorously speaking, again you have to consider if their impact, or lack thereof, is reproducible.
Ethically speaking, if you mean to report this research to a peer-reviewed journal, then you should make note of this finding and that you have carefully considered it. It could provide the reviewers with important information concerning the methodology, analysis, and drawn conclusions. It is always okay to report findings if they are relevant to the topic, but it is not okay to purposely neglect the reporting of unfavorable relevant results.
Thank you Sir for the nice description.
The initial structure I took from the Protein Data Bank. There were no missing residue in that pdb coordinates. The only change I did is that I have removed the ligand from the protein so to chek the stability of the free enzyme. Now, by following standard simulation protocol, I performed two simulations of 5 ns each. In both the cases I see these observations. The flexible loop does not have any secondary structure and it is way far from the active site.
It may be that this flexible loop is facing towards the solvent, where it counters strong fluctuations due to the environment. You can also calculate RMSF per residue which
will give the correct idea of residual movement and last but not the least i hope you have aligned the protein before calculating the RMSD.
Thanks
You are right, the loop is getting exposed to the solvent as time goes.
I have alligned the protein while calculating the rmsd. RMSF would be a great idea.
Where this termini present in the crystal coordinates. if so , do they form part of ur protein domain. if its just a extension of ur domain of interest, u can happily report two RMSDs line in the same plot: one with the terminal loops, and one calculated without them. The C-alpha rmsd without the terminal random coil will be much lower intuitively. Its know that solvent exposed terminal always have high rmsf. So plot a RMSF as suggested by Prashant. If your protein has homology, look into those structures, does the relevant region are also terminal loops. If so, u can confirm that u r doing the right thing
I have plotted RMSD in presence and absence of the loop. In absence of the loop, it is coming around 0.3 nm. Now, I suppose everything is fine for the active part of the protein. RMSF plot also indicates the high flexibility of the protein in the two terminii.
look at the ramachandran plots for the fluctuating residues, and if you find the termini virutally uncorrelated to the rest of the protein, you can conveniently ignore them in your analysis as they have little structural or functional importance for you system. But you should be careful and not ignore more than 3 (at max) residues per terminal if the do turn out to be uncorrelated.
One more bit of information. Protein crystal structures are models, and these models have varying degrees of confidence depending on the resolution of the X-ray structure. In addition, crystallographers have data on the motion of residues that are solvated by looking at the temperature factors (or B factors). If the B factors of your flexible termini are much higher than the average B factors for the rest of the model, then the crystal structure itself contains information that shows that these residues are moving. If that is the case, then the flexibility of the MD simulation is, roughly, reproducing the flexibility seen in the X-ray structure. If you read through Crystal Structure Analysis by Glusker and Trueblood, or talk with a crystallographer who can look at the experimental information on the structure, that will help you interpret your MD results.
- Badges
- Science topic