[gmx-users] some prolem about genbox genconf and editconf

[Mark Abraham]

DreamCatcher wrote:
> Hi everybody,
>      I have come across some problem when using genbox, genconf and 
> editconf try to make a box with a certain density and a certain number 
> of polymer chains.
> First, I use genbox to have one polymer chain filled in the box without 
> any solvent.Then I use genconf to pile up 27 polymer chains ,after that 
> I use editconf to designate a density of  0.3 g/ml to attain the final 
> goal. When I finish it, i find that all the bond that connect the atoms 
> are gone and all the atoms disperse in the box.
>     In another try, I first use a pdb file which contains only one 
> polymer chain to genconf a box with 27polymer chain, and then use 
> editconf to designate the density, but finally I get a box with lots of 
> bonds that might not present in a normal structure, like a carbon was 
> bonded with a 3 or 4 or even more atoms distance long and the bonds 
> connect with it is overcount. What is the prolem? How can I fix it? 

As a side issue, read 
http://www.gromacs.org/index.php?title=Download_%26_Installation/Related_Software/Visualization_Software#Topology_bonds_vs_Rendered_bonds

Your scaling by density isn't working. I'd guess that editconf is using
share/gromacs/top/atommass.dat to guess the atom masses and that
your polymer's residue and atom names are not present there. Copy that
file to your working directory, add suitable entries to the end and
see how you go. I think editconf should be confirming which atommass.dat
file it might be using - read it carefully. It should also be warning
you if it doesn't find suitable masses!

Mark