[gmx-users] Bad .tpr or .xtc files?

[Mark Abraham]

Peggy Yao wrote:
> Thank you, Tsjerk. I will check them out.
> 
> But I am still confursed with what has gone wrong. As I said in my 
> previous eamil, in fact, 2 Ca atoms are a part of the protein. They are 
> in the PDB file passed to pdb2gmx. Water molecules are added after the 
> protein atoms and the Ca atoms, as in the PDB file after genbox. So I 
> suppose, in the .tpr file, atoms should be ordered as 
> protein-Ca-water-Cl?

The molecules are ordered according to the [molecules] section of your 
.top file. The atoms within the molecules are ordered according to the 
corresponding [moleculetype] sections. So look it up, don't suppose it :-)

> If so, I don't know why when writing out Ca or Cl 
> atoms, I got segmentation fault and index out of bound errors.

Like Tsjerk said, you have a mismatch between your trajectory file and 
the structure file you're using. GROMACS will auto-generate an index 
file from that structure file if you don't provide a separate index 
file. So either you need to make an index file that corresponds, or get 
a structure file that corresponds. Tsjerk's told you how to do these.

One issue to be wary of is that case can be relevant and Ca does get 
used in some places either for calcium or for alpha-carbons.

Mark