what is the difference in each replica? starting structure was same? all paramters are same?

Hi @Nikhil Maroli

thank you for your fast response.

yeah, the pdb structure is the same (wild type monomer protein), I used tleap to prepare the structure (solvation and neutralization of the charge adding Na+ and Cl- ions). For each replica, I did the following: two minimizations (one with restrains), a heating phase, an equilibration phase and a production run phase. The input file .in for these phases are the same. I don't know if running again "ex novo" the 2 step minimization for each replica before doing the real MD simulation caused this difference, but I'm not sure. Furtnermore during the 3 MD simulations phases, I set ig = -1, in order to have different initial velocities and different trajectories and so explore better the space, though the differences in RMSD seem huge. The protein I'm studying is SOD1, which theoretically is really stable.

Hi,

SOD1 contains two metals, did you add them? how did you model them? If not, then SOD1 has a longer loop that can cause RMSD fluctuation. Did you compare RMSF and see what region is fluctuating? Finally, SOD1 is a dimer protein, if I, if I remember correctly, SOD1 is a dimer protein remember correctly, so running a monomeric form makes less sense.

More points:

1) Plus, 1-3 A is not a big issue, so most of these are relatively stable.

2) 3000 frames are not enough to make this conclusion.

3) Just look into the yellow replica, around 2000 frames, why it is diverging. it might be the loop.

Run a longer simulation and then compare RMSD.

Hi Ayaz Anwar

In PDB RCSB I didn't found a human SOD1 with Zn and Cu inside, , but I found a animal one with those metals. So, because the structures were very similar after the superposition, I simply loaded the two proteins in MOE, removed the animal protein and saved the human protein with the metals derived from the animal one.

I used the instructions saved in the code.txt file to prepare the PDB for the simulation and models the ions.

Also because I had an issue with the hydrogens whose type had not be recognized by Amber (FATAL: Atom .R does not have a type). SO, i SIMPLY trimmed them with the command: reduce -Trim monomer_amber.pdb > monomer_amber_noH.pdb; but I don't know if this does make any sense (maybe I need to re-add them with "addions monomer H" in tleap.

I created also several dimers with Haddock online tool (even those with metals inside following the same procedure as before) and performed some simulations on those, but I wanted to check if there were any problems with the monomers before analysing the dimers.

Thank you for your help

Furtnermore during the 3 MD simulations phases, I set ig = -1, in order to have different initial velocities>> this will give you different trajectories, you may need to extend the simulation time for better consvergence

There are many human SOD1 dimer with metal ions, here is one:

https://www.rcsb.org/structure/2c9v

No need to use docking structure when there is a crystal strucutre available.

Hydrogen usually creates problems, so rebuilding as per FF definition is fine (for many applications).

Finaly, as I already said, the length of simulation is short, so extend the simulation and then compare the RMSD.

Thank you Ayaz Anwar. 900ns can be enough for what I'm looking to achieve?

Yes, that should be enough, but if you can run for that long, go for 1100ns, and do not use the first 100ns in any analysis. This is the period where the protein relaxes and sometimes fluctuates dramatically. If not, you can use the last 500ns of trajectory for your analysis.

Good luck.

Ayaz Anwar I just did 900ns simulations of wild type monomers and mutated ones. It seems some converge, some other not. Maybe using dimers as you suggested, the stability will be reached sooner?

I will try 1100ns or more for the simulations.

It seems that the wt becomes quite stable, while some mutated types did not, so I still don't know if I should consider those results ok to extract some other features like radius of gyration or RMSF. The structures differ only the structures differ only by one amino acid (point mutation).

I specify that all these simulations are done on the structures I prepared removing hydrogens (with the instruction "reduce -Trim monomer_amber.pdb > monomer_amber_noH.pdb"), so I don't know if, not having rebuild the hydrogens with "addions H" in tleap, caused some problems, regarding the stability of the MD simulations, thanks.

I don't want to be too pushy, please respond when you can, if it's possible.

Thank you

while some mutated types did not, so I still don't know if I should consider those results ok to extract some other features>> why you do this simulation? what is your aim? why you mutated? isnt obvios that muation shows higher or lower stability?

It is generally accepted that the RMSD fluctuation ranging between 1-3 Angstrom is normal and accepted. So, for most of the simulations, this criterion was met.

Building H-atoms in tleap does not have anything to do with it. However, the starting structure that you build might be the reason.

Moreover, what atoms were used for RMSD calculation? Ca or backbone or whole heavy atoms?

I am attaching one figure as an example, where the blue is good. If you look at the yellow replica, beyond 2000 or 3000 frames, it also acquired stability or fluctuated within the acceptable range. Like, I have already mentioned you can ignore the starting trajectory.

For the dark brown replica, it is also stable but somehow, you need to check the trajectory at around 6000-8000 frames what is causing the spike in RMSD? I guess there must be some loop that is swinging away from the core protein region and causing this spike in RMSD.

Finally, MD is a chaotic/random process as evidenced by these 3 replicas, where 1 did not show any sign of movement/fluctuation (blue), one was slightly chaotic (yellow), and the third (dark brown) showed another aspect of the same protein (maybe loop movement; guess).

That's all that I can help.