# Gromacs molecular dynamic simulation throwing error during minimization

I am trying to investigate a membrane proteins influence on a lipid bilayer using molecular dynamics. I am new to MD and trying to run my first simulation. I have built the simulation input files using CHARMM-GUI and now I am using gromacs to perform the simulation.

CHARMM-GUI provides these commands to get gromacs to perform the initial minimization

gmx_d grompp -f step6.0_minimization.mdp -o step6.0_minimization.tpr -c step5_input.gro -r step5_input.gro -p topol.top
gmx_d mdrun -v -deffnm step6.0_minimization


The first command runs fine. The second command starts the minimization, but at step 15 it freezes.

...

Step 15, time 0.015 (ps)  LINCS WARNING
relative constraint deviation after LINCS:
rms -nan, max 0.098209 (between atoms 16596 and 16597)
bonds that rotated more than 30 degrees:
atom 1 atom 2  angle  previous, current, constraint length
15683  15684   39.2    0.1109   0.1121      0.1111
15683  15685   38.4    0.1115   0.1107      0.1111
16596  16597   72.1    0.1111   0.1220      0.1111
16640  16642   35.3    0.1116   0.1120      0.1111
16643  16644   51.9    0.1108   0.1136      0.1111
16643  16645   47.2    0.1114   0.1079      0.1111



In step 15, the rms is -nan, so I'm guessing that is the problem, and I am assuming that is coming from the difference between the "current" and "previous" is greater than constraint length like on this line here

16596  16597   72.1    0.1111   0.1220      0.1111


The problem is I don't have any intuition on how to fix it? The pdb file is a homology model from phyre without any modification other than adding a single chain id to all the residues. I'm using a 1 to 1 POPC and DOPC in both membrane layers with a CHARMM36m force field.

The previous steps also throw LINCS warnings about bonds exceeding 30 degrees, however they produce a real value for rms.

I've counted on tutorials to get this far and would appreciate some suggestions or good resources I can use to continue my simulation.

Thanks!

Try running/resuming using -cpi the simulation again, I got NaN error and rerun kept continued.