# Unusual RMSD of ligand-protein **URL:** \ **Category:** User discussions\ **Tags:** mdrun, analysis-tools\ **Created:** [May 4, 2025, 12:16pm UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031 "2025-05-04T12:16:34Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![nilanjana88](https://avatars.discourse-cdn.com/v4/letter/n/59ef9b/32.png) [@nilanjana88](https://gromacs.bioexcel.eu/u/nilanjana88)\ **Post date:** [May 4, 2025, 12:16pm UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031/1 "2025-05-04T12:16:35Z")
GROMACS version: 2024.2 GROMACS modification: No I have run a protein-ligand simulation for which the backbone RMSD of protein (lsq fit of backbone to backbone of protein) and RMSD of ligand (lsq fit of ligand to backbone) have been calculated. The ligand RMSD being very high (around 9nm), but doesn’t show any major fluctuations or peak, so probably not a trajectory issue. Also, the ligand seems to be thoroughly bound to the receptor during the 100ns simulation when visualized in VMD. Below is the plot showing ligand RMSD (in red). and protein backbone RMSD (in black). Could you please shed some light on why the ligand RMSD is so high? ![image](https://europe1.discourse-cdn.com/flex017/uploads/bioexcel1/original/2X/7/70105d72fabefde18bc822f1635ba22ab4d7d293.png) Thank you! ---
**Author:** ![Marius](https://avatars.discourse-cdn.com/v4/letter/m/ed655f/32.png) [@Marius](https://gromacs.bioexcel.eu/u/Marius)\ **Post date:** [May 4, 2025, 2:04pm UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031/2 "2025-05-04T14:04:20Z")
Hi, I’m not entirely sure about the wording here: > RMSD of ligand (lsq fit of ligand to backbone) have been calculated. You probably used `gmx rms` for this. Did you select “Backbone” first and then “Ligand,” or vice versa? Best regards, Marius ---
**Author:** ![nilanjana88](https://avatars.discourse-cdn.com/v4/letter/n/59ef9b/32.png) [@nilanjana88](https://gromacs.bioexcel.eu/u/nilanjana88)\ **Post date:** [May 4, 2025, 2:15pm UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031/3 "2025-05-04T14:15:34Z")
Yes, backbone first, and then ligand for calculation of ligand RMSD. ---
**Author:** ![Marius](https://avatars.discourse-cdn.com/v4/letter/m/ed655f/32.png) [@Marius](https://gromacs.bioexcel.eu/u/Marius)\ **Post date:** [May 4, 2025, 2:33pm UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031/4 "2025-05-04T14:33:36Z")
This could explain your observation. With the options you chose, you are calculating the ligand’s RMSD **relative** to the protein backbone. Consequently, your ligand curve no longer reflects deviations in the ligand’s internal coordinates, but rather the ligand’s displacement with respect to the protein backbone. Did you perform your analysis on the raw trajectory, or did you first correct for periodic boundary conditions? Which reference structure did you take for the RMSD calculation? Best, Marius ---
**Author:** ![nilanjana88](https://avatars.discourse-cdn.com/v4/letter/n/59ef9b/32.png) [@nilanjana88](https://gromacs.bioexcel.eu/u/nilanjana88)\ **Post date:** [May 4, 2025, 2:47pm UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031/5 "2025-05-04T14:47:27Z")
I corrected for periodic boundary condition with “-pbc mol -center -ur compact” flag of gmx trjconv, finally applied “-fit rot+trans”, checked if the ligand is leaving the binding pocket, but it was consistent within it. Then I calculated RMSD on this fitted trajectory file. ---
**Author:** ![Marius](https://avatars.discourse-cdn.com/v4/letter/m/ed655f/32.png) [@Marius](https://gromacs.bioexcel.eu/u/Marius)\ **Post date:** [May 4, 2025, 4:47pm UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031/6 "2025-05-04T16:47:54Z")
Okay, but nevertheless I would guess that the actual calculation of your ligand RMSD is relative to the backbone of the protein. If you select `Ligand` for both the fitting as well as the RMSD calculation the value should become much smaller (depending on the size and flexibility of your ligand). You can also compare your reference structure (using `gmx rms -s` ) with individual frames from the fitted trajectory (using `gmx rms -f` ) to pinpoint which motions could cause this large deviations. ---
**Author:** ![nilanjana88](https://avatars.discourse-cdn.com/v4/letter/n/59ef9b/32.png) [@nilanjana88](https://gromacs.bioexcel.eu/u/nilanjana88)\ **Post date:** [May 23, 2025, 8:33am UTC](https://gromacs.bioexcel.eu/t/unusual-rmsd-of-ligand-protein/12031/7 "2025-05-23T08:33:34Z")
Hi @Marius Sorry for my late response. I calculated ligand RMSD with ligand (group 13) for both the “least square fit” and RMSD calculation. The value actually came much smaller (~ 0.2 nm). Although I have seen people calculating ligand RMSD with both “least square fit to the protein backbone” and “least square fit to their own”, I would like to know which is more standardized method to calculate ligand RMSD? J.lemkuls’ MD tutorial had described RMSD calculation with ligand least square fit to the protein backbone, which I followed in many of my simulations, and mostly it gave RMSD values around 0.4 nm. As per my understanding, we are calculating root mean square deviation of the ligand after it is bound to the protein, and backbone is selected for “least square fit” because it is more rigid, and less prone to changes during the simulation trajectory, so it’s easy to align the backbone structure to the reference structure (I took the starting structure as reference). Please forgive me if i’m wrong.