[gmx-users] OPLS to gromos96 problems

[Ken Rotondi]

Hello All,

I've successfully built .gro and .top files for a large riboprotein in 
OPLSAA using GROMACS. The problem is that it is so large I really need 
to use a UAFF to keep this computationally tractable. I tried running 
pdb2gmx and received errors such as

"WARNING: atom H21 is missing..."

I entered the missing hydrogens to the .hdb file and fixed this. 
However this brought forth two additional problems:

------------------------------------------------------------------------
QUESTION 1) long bonds, there must be thousands
eg
Warning: Long Bond (74318-75599 = 13.2452 nm)
Warning: Long Bond (74345-75599 = 13.5764 nm)

Several questions: What do the atom numbers refer to? The input ".gro" 
file numbering? The uncompleted output .gro file? Also, I'd say with 
the exception of ~25 Long Bonds they are all between some atom and atom 
#75599. I find this confusing since I'm using a GROMACS .gro file as 
input. Are there any suggestions as to what this might mean?

------------------------------------------------------------------------
QUESTION 2) "Fatal error: Atom H211 not found in rtp database in 
residue GUA, it looks a bit like H21"

I assume that this is due to the fact that I have >10,000 atoms in my 
system and the atom type and atom number columns are merging. GUA is 
the first nucleotide after 10,000 atoms and the atomtype and number 
appear thusly for the H21 atom in this residue:

H2110013

Is there a way to have GROMACS accept H21 rather than exiting? If this 
is not the case do I need to go back to the .pdb files to get around 
this problem?

Many thanks, as always,

Ken