[gmx-users] inconsistent shifts warning

[sang eun jee]

Dear gromacs users

Hello.

I have question about inconsistent shifts warning when I am simulating
infinite peptide chains.

My system is (GLY-PRO-HYP)7 and I connected N atom of the first GLY residue
to the C atom of the last HYP atom. Therefore, the chain is
(N-CA-CO)-PRO-HYP-(GLY-PRO-HYP)5-GLY-PRO-(N-CD2-CG-CB-CA-CO) , while
N-CA-CO represents first GLY and N-CD2-CG-CB-CA-CO represents last HYP. N
of GLY is connected to the C of HYP via editing *.ftp, *.hdb and other
related forcefield files. specbond.dat was edited to include this specific
bond.

After editing forcefield, I processed structure with pdb2gmx and checked
all details of the bonds,pairs, dihedrals from topol.itp file that pdb2gmx
generated. I checked structure files with vmd visualizer and pdb visualizer
and confirmed the two atoms in each edge are close enough to form bond in
pbc condition.

In my system, NHCO protein bond between first GLY and last HYP crosses
periodic boundary.
So, I set *.mdp option as pdb=xyz, periodic_molecules =yes. Energy
minimization worked fine.
But, when I ran nvt & npt md, with and without constraining backbones, I
showed warning "There were 2 inconsistent shifts. Check your topology".

When I checked the structure files,  C and O atoms from last HYP is
detached from the last part of the chain and attached to the first part of
the chain.
 (N-CA-CO)-PRO-HYP-(GLY-PRO-HYP)5-GLY-PRO-(N-CD2-CG-CB-CA-CO)
-->  (CO)-(N-CA-CO)-PRO-HYP-(GLY-PRO-HYP)5-GLY-PRO-(N-CD2-CG-CB-CA)
And when I changed the number of chains to 3, then I could observe number
of inconsistent shifts are exactly consistent(6) with the number of C and O
atoms of HYP, which were moved to the first chains.

I searched archives and there were many reports related inconsistent shifts
warning, but none of them are related with my case. From careful
observation of the gromacs codes mshift.cpp, it looks like this warning is
printed when the position of the shifted atom by pbc is not consistent with
the designated position, but I cannot figure out more details.

If some of you have any idea or experience with this error, please help me.
I really appreciate it.

Sangeun Jee


-- 
Post Doctoral Researcher
School of Materials Science and Engineering
Georgia Institute of Technology