[gmx-developers] How to make protein whole in do_force, Warning: 1-4 interaction between 567 and 575 at distance 5.695 which is larger than the 1-4 table size 1.000 nm

[mchen2 at mailer.fsu.edu]

you guys,
My system always crashes with /*
Warning: 1-4 interaction between 567 and 575 at distance 5.695 which is 
larger than the 1-4 table size 1.000 nm
These are ignored for the rest of the simulation
This usually means your system is exploding,
if not, you should increase table-extension in your mdp file
Warning: 1-4 interaction between 567 and 575 at distance 6.236 which is 
larger than the 1-4 table size 1.000 nm
and LINCS errors ... */
I'm pretty sure that it's caused by the broken topology of protein. 
Because it happens when partial structure of protein(side chain) went 
out the box(my water box is 6.5, 6.5, 6.5, and here 567 and 575 are the 
"CD" and "HZ3" atoms of LYS, and the "NH3" is out of box).
It say "The coordinates (x) are shifted (to get whole molecules) in 
do_force" in "md.c" sourcecode, but I cannot find any shifts in 
"do_force()" except the routines of put_charge_groups_in_box() and 
mk_mshift(). Actually I did a little adaption based on the 
sourcecode(ver3.3.1). I'm sure I alternate the code properly except: I 
call "do_force()" for many times in one step md loop like that 
"minimize.c" do;  After each "do_force()", "unshift_self()" was called.
It's really vexatious, help me!
Can you guys check the "do_force()" section, tell me how to make the 
molecule whole properly or it's my misunderstanding totaly.
Thanks
Mengen Chen

----------------------------------------------------------------