Asked 22nd Nov, 2016 in the project Docking & Molecular Dynamic Simulation
  • AstraZeneca, Chennai, India

Water molecule starting at atom 14394 can not be settled- GROMACS?

Thanks in advance for your help. 
ERROR: Water molecule starting at atom 14394 can not be settled. Check for bad contacts and/or reduce the timestep if appropriate.
I tried changing mdp file and re-running it. This error hasn't appeared for the earlier simulations done for the same protein. I am just using a different ligand now.
How to mitigate this error in Gromacs protein-ligand simulation when running nvt equilibrium step as per Justin's tutorial?

Most recent answer

27th Nov, 2019
Caroline Freitas
Federal University of Rio de Janeiro
Manasa Surakala , I'm facing the same problem... Did you generate the topology using CGENFF? Did you do anything different in the second time?
Thanks in advance.

All Answers (8)

22nd Nov, 2016
Manasa Surakala
AstraZeneca, Chennai, India
 The topology file had to be generated properly, i.e., problem with the coordinates (itp file) . Used the regenerated file, that resolved the error. 
1 Recommendation
23rd Nov, 2016
Hiqmet Kamberaj
International Balkan University
I would suggest to first run energy minimization module (e.g., using Conjugate Gradient method or others available in Gromacs), then using the final minimized structure you can start standard MD simulations.
The error could happen because of the bad contacts between atoms (i.e., distance between them is too small) and the forces become too large, thus the numerical integrators fails (think, the velocities are updated according: v(dt/2) = v(0) + (dt/2) F(0)/m, where dt is the integration step size and F(0) is the force at previous iteration step, if F(0) is too large, then the velocity becomes to large, which is equivalently adding too much energy into the system in the form of kinetic energy. Of course, reducing time step would help, but it all depends on the product (dt/2)*F(0) at the end!)
29th Nov, 2016
Yaojia Chen
Nanjing Tech University
I met the similar problem, could you give me some advice? I had check my topology file but found no mistake.
29th Nov, 2016
Dr Thirumal Kumar D
Meenakshi Academy of Higher Education and Research
Bad contacts refer to atomic clashes that arise either due to an insufficient  minimization or broken model physics that cause molecules to collide with
one another. So increase the number of minimization steps. If it is your modeled protein. Trying remodeling with some other tools.
1 Recommendation
2nd Dec, 2016
Nikhil Maroli
Indian Institute of Science
Are you running in GPU? make sure the topology is fine and reduce the time step and increase the number of steps in minimisation
Still not resolved give the full command you used.
16th Dec, 2016
Manasa Surakala
AstraZeneca, Chennai, India
The problem has been resolved. Thanks to all for your answers.
@Yaojia Chen - regenerate your topology file, even I couldn't find any fault but the newly generated file worked. Sometimes some elements in the topology are not generated.
1 Recommendation
24th May, 2017
Antonio Di Dato
University of Naples Federico II
I am crazy! How to generate your topology file?

Similar questions and discussions

How to remove GROMACS periodic boundary conditions?
23 answers
  • Anthony NashAnthony Nash
Hi all, I have done this countless time and I've "googled" this a lot. Unfortunately I have a simulation that simply won't behave. 
I am trying to visualise (and analyse) a trajectory in VMD after removing the periodic boundary conditions. It is a helical dimer (important, as I need to consider bonds over the box and also two separate peptides separating over the box), in a box of water. During the first frame the dimer is whole in the box.
I have tried many combinations of trjconv with little success. The most logical, as per the Gromacs web site was: 
trjconv -f human_npt_0_40ns.trr -s human_npt_0_40ns.tpr -pbc whole -e 2-o firstframe.gro
I checked the first frame - it is present as t=0.000. The protein (dimer) is whole in the box.
trjconv -f human_npt_0_40ns.trr -s firstframe.gro -pbc nojump -o nojump.gro
In VMD this displays lots of very long bonds i.e., passing through the unit cell to the other side - no good. I tried this option again with -center on the protein. Still no good. I tried to centre the protein for the trajectory after -mol nojump. Still wouldn't work. I am technically not sure what has happened here. On a similar simulation of a dimer, from the above two commands I was getting a beautiful trajectory in VMD. This would suggest that my reference file to the -pbc whole is incorrect, it is not. 
Short of listing every possible combination of -pbc I will leave this here and hope that someone has some insight into removing tricky periodicity from GROMACS trajectories. 

Related Publications

Conference Paper
This article presents the GPU parallelization of new algorithms SD and DPD types for molecular dynamics systems developed by the Molecular Dynamics Group, University of Groningen, the Netherlands. One should note that molecular dynamics simulations are time-consuming simulations of systems, running time ranging from days to weeks and months. Theref...
Mutations in leucine-rich repeat kinase 2 (LRRK2) are the most common cause of Parkinson's disease (PD). LRRK2 contains a Ras of complex proteins (ROC) domain that may acts as a GTPase to regulate its protein kinase activity. Here, we performed 10 ns molecular dynamics simulations on LRRK2 Apo, complex with GDP and mutations (R1441C, R1441G and R14...
This is the shell script for carrying out the molecular dynamics simulations and analysis of the temperature dependent unfolding of crambin in Gromacs.
Got a technical question?
Get high-quality answers from experts.