[gmx-users] Atom names mismatch

[Mark Abraham]

George Abadir wrote:
> Thank you very much for your reply. Actually I ran the whole simulations 
> (1000000 time steps), and the system did not crash. The molecule still 

I'm extremely surprised. Perhaps bond constrains on all atoms were your 
friends here. :-) Nonetheless, it seems likely from the mismatches that 
some incorrect van der Waals models were applied.

> has the structure of phenylalanine. Does this mean that the final 
> coordinates are ok, or should I redo the simulation again after 
> resorting the atoms.

Without understanding the reason for the mismatch, you can't say the 
simulation is OK with certainty. It looks like the phenylalanine 
structure you generated for the simulation ordered its atoms differently 
from the one you used to generate its topology. If you go and re-order 
the atoms in the structure file according to the topology file, you 
should see these warnings go away, and now you can be confident a 
resulting simulation is OK. I wouldn't trust the first one, though.

Mark