RMSD values are generally used for a ligand, when ligand show different poses at a particular binding site cluster of a protein. For selecting a particular pose of a cluster, two things you must consider 1. how big is the cluster.... 2. what is the Ki value or interaction energy value of the ligand with protein at that particular pose... the pose having lowest Ki value or highest interaction energy in a cluster will have reference value from which RMSD of other poses are calculated...

I hope while selecting a particular pose, you will consider all these factors...

Generally, RMSD values representation mainly for analyzing stability of protein and predicting conformational changes of protein. RMSD values depends up on binding interaction and energy between protein and ligand. Also the optimized protein has lowest RMSD values.

Usually, the lowest RMSD values reflect minimum Ao . The lowest RMSD would be acceptable. The ligand binding poses, energy and interaction fully depends on residual fluctuation (RMSF) values. After, protein ligand docking, the docked conformers were undergone for identification of RMSF analysis. the RMSF plot will depict the ratio of fluctuation in residue level. Hence the both dynamic plot are important to carry out the prediction of structural stability on protein.

Hope the information useful.

Regards

Rajesh

You don't state what you are measuring the RMSD with respect to, so the question is hard to answer.

RMSD has often been used to measure the quality of reproduction of a known (i.e. crystallographic) binding pose by a computational method, such as docking.

In this case a low RMSD with respect to the true binding pose, is good (ideally less than 1.5Angstrom, or even better, less than 1 Angstrom). This  represents good reproduction of the correct pose.  Beware reading much significance into relative values of RMSD's greater than 3. A binding pose with RMSD 4 Angstrom is not better than one of 6 Angstrom. They are both equally poor (i.e. wrong)

Dear John sir,

thanks for answer.

My objective behind the measurement of RMSD is to decide which poses  are accepted/rejected for each ligand docking from initial generation for

refinement.

sir, plz provide literature with respect to following statement,

A binding pose with RMSD 4 Angstrom is not better than one of 6 Angstrom. They are both equally poor (i.e. wrong).

Totaly agree with the reply of John Liebeschuetz.

Usually RMSD value is used to validate your docking protocol. Validation of docking protocol means you need to consider a crystallographic complex protein with ligand in it and perform the docking of the same complex. Then you need to check for the RMSD values. If your docking protocol able to produce similar docking pose of a ligand with respect to the biological configuration of the same ligand in the crystal structure of complex protein then it means your docking is validated. Lower the value of RMSD you are getting higher the accuracy of docking. And so no doubt RMSD values less than 1.5 Angstrom are always wise to consider. There is no sense in speaking RMSD values higher than 3 Angstom (3,4,5,6,..... are the same rather useless to consider).

 It is well known that values of RMSD  above 3.0 Angstrom are pretty meaningless when looking at docking pose quality. This can be easily verified by examining by eye docking runs for which RMSD's have been calculated. Docking poses with RMSD above 3.00 Angstrom will have few, if any, of the key binding interactions of the true binding poses. Since the key binding interactions really define the binding pose much better than any numeric metric such as RMSD, it is meaningless to say a 4angstrom RMSD pose is better than a 6angstrom pose if neither pose make any of the key interactions.

  Some of the publications on docking have been severely criticized for using average RMSD over a test set of different protein/inhibitor complexes, as a criterion for overall docking perfomance. The reason for the criticism is simply that a high RMSD for one complex, say 10.0, will bias the the performance criterion adversely, much more than a lower RMSD of  4.0, say. Yet as previously stated both should be treated as equally 'wrong'.

Regarding the value of RMSD as a metric, it needs to be pointed out that it is by no means perfect. It has to be carefully calculated to ensure that any symmetry in the molecules is taken into account, otherwise high RMSD's can be returned for docking poses that are actually correct. Pseudosymmetric (i.e. nearly-symmetric) molecules cause particular problems because then docking poses can be generated that can be considered very reasonable but which have high RMSD (even if correctly calculated to take account of symmetry). An example is given in Liebeschuetz, Cole and Korb,   J. Comput. Aided. Mol Des (2012) 26:737-748, .

Another problem with RMSD is that it cannot be used for molecules dissimilar to the reference molecule as it relies on a one-to-one atomic correspondence. So it only has value in Virtual Screening if you are docking a chemical library which has a common fragment of known desired binding mode. In that case you can use RMSD of the common fragment to the known binding mode, as an additional criterion for compound selection, alongside the docking score.

thanks all of you.

John Liebeschuetz I have docked a peptide-protein complex and I want to compare it with another, but as you mentioned my peptide that I docked has a common fragment of known desired binding mode with the peptide that I want to compare it to.

I am using pymol to see RMDS. How can I see this desired RMSD in pymol? because aligning the complexes in different ways the RMSD is different...

Ty

John Liebeschuetz I have docked a peptide-protein complex and I want to compare it with another, but as you mentioned my peptide that I docked has a common fragment of known desired binding mode with the peptide that I want to compare it to.

I am using pymol to see RMDS. How can I see this desired RMSD in pymol? because aligning the complexes in different ways the RMSD is different...

Ty

Hi Mafalda Inês Pereira,

https://www.researchgate.net/post/How_can_I_calculate_RMSD_between_two_ligands_crystal_and_docked

this link useful..

If two same protein structure are superimposed for backbone atoms using any tool, where one is docked with a ligand and the other is without docking. What is the significance of RMSD in case of docked structure?

Does this information has any importance especially when the docking program takes into account the backbone flexibility during docking?

Most of the docking protocols are of rigid mode only. If we try to dock the co-crystallized ligand by following same protocol, how can we expect a low RMSD value?

Flexible docking by Induced-fit mode or by running molecular dynamics, then chances of getting low RMSD will be higher.

If so, can we use RMSD values to validate docking protocol??

So I only partially agree with you Manikanta,

Let's consider two situations. In the first we rigidly dock a ligand back into its own crystal structure; and we also try a flexible docking protocol and compare the results. In general the RMSD from the flexible docking protocol will be WORSE than that from the rigid docking protocol. This is because allowing the protein flexibility will take it away from the starting crystal structure and make it harder to achieve the exact binding pose against this structure.

In the second case we dock a ligand back into a protein structure crystallised using another ligand, but calculate its rmsd against it's position in its own protein -ligand crystal structure (for avoidance of doubt this is the same protein for both ligands) . This is called cross-docking and is often a more difficult problem for rigid docking, than docking back into the same crystal structure (self-docking), as there may be conformational changes in the active site of the second structure, since it binds a different ligand.

Rigid cross-docking quite often gives high rmsds. In those cases where it does give a high rmsd, flexible docking may well then give you a lower one, as then the sidechains can relax and accommodate the ligand in its proper binding mode

In practice we do both 'self docking' and 'cross-docking' . It is important to appreciate that the probability of getting good dockings is significantly lower in the cross-docking case, unless the protein is almost rigid. it is in these cases where the protein is not rigid, that flexible docking may help you.

John Liebeschuetz Thank you for the explanation.

As mentioned, Self-docking and Cross-docking in terms of rigid vs flexible docking results tells that conformations are independent for every ligand and protein in the complex.

So is docking a right way of justifying or correlating with experimental results?

Rigid cross-docking gives high RMSDs in many cases. Then how can we rely on results generated for unknown compounds.

Does low RMSD of rigid self-docking validates the docking process for that protein? What are the possible errors for high RMSD in rigid self-docking and How to correct them?

So I'll try to address your very good questions, Manikanta.

Firstly, you are quite right, Manikanta, if we cannot reproduce the correct pose of a ligand in docking, then we should not trust the results of docking similar molecules to that ligand. We should only trust the results for ligands very similar to those ligands that dock with low RMSD.

However … what we can do is try to improve the docking protocol so that it does reproduce the binding mode of our problem ligand. The first way to do this is to get the docking program to generate a variety of reasonable poses (most docking programs can be configured to do this ) and then rescore the poses using another method, based perhaps on molecular mechanics forcefield, or knowledge based potential or possibly, as Samee suggests, quantum chemical calculations (though this is still difficult at the moment for the protein ligand interface). A rescoring method that ranks the crystallographically observed pose first, can then be used to process the docking results of similar molecules.

Another thing to do sounds like cheating, but actually also works pretty effectively in my experience if you are docking a series of closely related compounds and have a crystal structure for one of them. If there is a constant portion of your designs that you are sure always stays in the same position in the active site, then you get the docking protocol to fix or restrain that portion to its crystallographic position. Not only do you get much better poses this way (naturally) but the docking scores give you a good idea of the validity of a design according to the hypothesis that it binds the same way as your one experimental structure.

There are other ways to bias docking protocols, by, say, ensuring a key hydrogen bond is always made, and such like. Many docking programs allow you to tweak the protocols so that you can do these things.

Biased docking like this, may seem not quite right, but it is a pragmatic and useful thing to do if you want to design new bioactives.

In answer to Amy, I say that in my experience, enantioselectivity is often quite well handled by the docking part of the program although it is very system dependant. However, there are certainly cases where the inactive and the active enantiomers bind with similar docking scores. In those cases the rescoring strategies or biased docking protocols I described in my previous message, will also work well to pick out the right enantiomer.

Thank you John Liebeschuetz for the clarification. I am very much interested to explore and validate parameters in this direction.

Samee Ullah what do you mean by "QM properties"

Hi Samee Ullah ,

First, thank you very much for this informations, I just wanna know what is the best way to evaluate "protein-ligand complex interactions" after docking is it RMSD, or QM approach and what are the tools that you recommend for such kind of approach.

Thank you very much Samee Ullah ,

It's really valuable information thanks again for your time, sir. really the docking is very sensible and not easy. in fact, We are working hard to make docking better. I also recommend GOLD,in your opinion is there any way to get GOLD free?

In docking RMSD value is used to compare the docked conformation with the reference conformation or with other docked conformation.

Maybe this link is useful

https://www.researchgate.net/post/Chirality_in_autodock_rosetta_etc_ligand_protein_docking_simulations

John Liebeschuetz Which ligand is going to be my reference for RMSD calculation? Is it the ligand from the unprocessed/not energy minimized/raw crystal? Or the ligand I split from the energy minimized crystal?

Thanks in advance for your answer.

Do we consider RMSD cluster or refference RMSD in DLG file of autodock which one is the most important