[gmx-users] glycam amber in gromacs mixed 1-4 scaling glycam06 fudgeQQ pairs coulomb amberports

[Oliver Grant]

Hi all,

I currently mix GLYCAM and AMBER in gromacs4.0.7 using the amberports and
amb2gmx.pl for the glycan. My problem is that I can't mix the scaling
properly as required for correct rotamer population sampling of the glycan
with the protein present.

>From Glycam_06g.dat:
"Correct rotational behavior for O-C-C-O fragments requires SCEE=SCNB=1.0.
This is in contrast to "standard" AMBER, in which it is normal to set
SCEE=1.2 and SCNB=2.0.  Unless you are attempting to generate rotamer
populations, it is OK to use the "standard" values.  Using non-standard
values (SCEE=SCNB=1.0) may be unacceptable when a protein is also present."

So in gromacs GLYCAM requires a fudgeQQ = 1.0 and a fudgeLJ = 1.0 whereas
the ffamberports requires fudgeQQ = 0.8333 and a fudgeLJ = 0.5.
I've been reading the mailing list archive and chris neale's method from
2006 for combining the Berger lipids and OPLS-AA forcefields however, as he
states, the method won't work in this case as the ratio of the fudgeQQ
factors is not an integer.

My understanding was that it is not possible at the minute to do this
correctly in gromacs. It would require a [ pairtypes ] section for 1-4
coulomb interactions. However I found this discussion from 2008:
http://dev-archive.ambermd.org/200812/0000.html

The relevant part being:
"we're chatting about implementing mixed 1-4 scaling in gromacs. It's easy
to enable on a per-molecule basis.  Enabling on a per-residue and per-atom
basis is certainly possible, but there are a bunch of ways to do it."

So... Is it currently possible to implement mixed 1-4 scaling on a
per-molecule basis? I've checked the 4.5.3 manual but it seems not to have
changed since 4.

Thanks,

Oliver
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110120/94355268/attachment.html>