Question regarding choice of group for Least Square Fitting Step

User discussions
1

Hello Everyone

So I have completed production MD for my Protein-Ligand System. My question is regarding gmx_rms RMSD calculation steps.

The tutorial on MD Tutorials suggest that if I am trying to calculate RMSD for my LIGAND, I should choose BACKBONE for the least square fitting step and LIGAND for RMSD step. I have 2 questions:

(1) the silly question : if I choose BACKBONE for the least square fitting step, and LIGAND for RMSD step, will the RMSD be calculated by comparing position of LIGAND in the trajectory compared the the reference position of the LIGAND, right? Whatever (BACKBONE/PROTEIN/LIGAND) I choose for least square fitting step, the RMSD is always calculated by comparing value of X in trajectory vs value of X in reference, right? The tutorial states “By doing so, the overall rotation and translation of the protein is removed via fitting and the RMSD reported is how much the JZ4 (LIGAND) position has varied relative to the protein” - this confused me, isn’t RMSD telling position of JZ4 compared to its position in reference?

(2) why isn’t it more apt to choose LIGAND in the least square fitting step too in addition to choosing LIGAND in the RMSD step.

(3) As I understand, the least square fit works on the trajectory file and then aligns the least square fitted structure to the reference structure. Am I getting this correctly.

Thank you
DJP

GROMACS version:Latest
GROMACS modification: Yes/No
Here post your question

2

If you perform fitting on a very rigid ligand like the one in the tutorial, the RMSD is likely to be close to zero for all frames. Fitting to the protein allows you to see if the ligand position changed (diffusion), which would be eliminated in this case. If you fit to the protein, you can see how the ligand moves relative to the protein.

See answer to (1). It might make sense in some cases. For the tutorial, it really doesn’t.

Yes.

1 Like
3

Thank you so much for the clarifications, Dr Lemkul.

4

Hi Dr.Lemkul,
When might it make sense to select the ligand for least-squares-fitting?

Based on your previous answer, I am guessing we can choose ligand for lsq if the ligand is not so rigid or if we see the ligand actually change its binding pose in the movie, is this correct?

5

Hi,

Sorry for jumping into this topic.
For my simulation, it’s a small protein of 250 residue and ligand has 60 atoms. While watching the trajectory I found ligands moves a lot from the binding site. So choosing “backbone” for lsq fit, and “LIG” for RMSD calculation gives a very high RMSD value. In such cases, should I select “LIG” for both lsq fit and RMSD calculation?

Thank you