ERROR: Constraint failure in RATTLE algorithm for atom 2572!

[bhavranek]

I have tried to create simulation files using both BFEE and BFEE2 and no matter what whenever I run simulation on step 002_euler_theta I get the same error:

ERROR: Constraint failure in RATTLE algorithm for atom 2572!
ERROR: Constraint failure; simulation has become unstable.

I have tried everything suggested online. Including changing the timestep, increasing minimization to get rid of unfavorable contacts, increasing margin, etc. However, nothing has worked.

Are there any suggestions?

@fhh2626 @HanatoK
Thank you!

[fhh2626]

Did you use the latest version of NAMD, i.e. 2.15 alpha?
I also suggest checking the structure of the complex in eq.coor to see if there is anything wrong. If everything is correct, the protein-ligand complex should be located at the center of the box

[bhavranek]

I used NAMD 2.14 and I also tried NAMD 3.0. Is there a difference with NAMD 2.15 that would make it work better? Everything looks okay when I view the equilibration simulation. I am not sure what else to do. I tried everything. The system is not very big either only 169 residues with a small ligand. @fhh2626

I downloaded Linux-x86_64-multicore-CUDA (NVIDIA CUDA acceleration) under Version Nightly Build (2021-07-14) Platforms:
https://www.ks.uiuc.edu/Development/Download/download.cgi?PackageName=NAMD

I tried the non CUDA accelerated build with no luck.

[fhh2626]

I used NAMD 2.14 and I also tried NAMD 3.0. Is there a difference with NAMD 2.15 that would make it work better? Everything looks okay when I view the equilibration simulation. I am not sure what else to do. I tried everything. The system is not very big either only 169 residues with a small ligand. @fhh2626

I downloaded Linux-x86_64-multicore-CUDA (NVIDIA CUDA acceleration) under Version Nightly Build (2021-07-14) Platforms:
https://www.ks.uiuc.edu/Development/Download/download.cgi?PackageName=NAMD

I tried the non CUDA accelerated build with no luck.

I am not sure what happened on your system. You can:

i) check if there are atoms with a coordinate of (0, 0, 0) in your PDB file;
ii) try to output eulerTheta CV during the equilibration step, and see what will happen