Hi All,
It appears that my last reply on the use of new UA force field with GB
was not archived. Here I'm sending it again ...
Sorry for the late reply on this. Yes, as Dave commented, the default
carbon atom radii would cause electrostatic solvation free energies in
ff03ua to be a bit too negative. For igb=5 applications with the mbondi2
radius set, the recommended radii for all three united carbon atoms are
2.2 A. In contrast, the all-atom carbon atoms is 1.7 A. Dr. Chuck Tan at
UC-IRVINE has done a systematic analysis on this and the following is
his report ...
++++++++++++++++++++
We have compared both GB (igb=5) and PB with TIP3P with respect to the
reaction field energies (relative charging free energies between in
water and in vaccum) for 13 neutral amino acid sidechain analogs in
ff03aa and ff03ua. For both GB and PB, the modified Bondi radii was
assigned by "set default PBradii mbondi2" in tleap.
ff03AA:
GB(igb=5) PB
CC: 0.9960 0.9917
RMSD: 1.3346 1.1575
Slope: 1.1487 1.0849
Offset: -0.5310 -0.6333
ff03UA: (with the recommended radii 2.2A for the united carbons)
GB(igb=5) PB
CC: 0.9907 0.9885
RMSD: 1.1178 0.9115
Slope: 1.1275 1.0833
Offset: -0.2932 -0.2496
CC: correlation coefficient; RMSD, room-mean-squared deviation; Slope
and Offset are for linear regressions with respect to the TI reaction
field energies in TIP3P.
++++++++++++++++++++
Note that Chuck did not try to minimize RMSD for ff03ua with respect to
TI reaction field energies, but simply to make its regression slope
similar to that in ff03aa. Let us know if you have any more comments and
suggestions.
All the best,
Ray
David A. Case wrote:
>On Wed, May 24, 2006, Ray Luo wrote:
>
>
>>The paper is now in press in JPC-B, to appear in June, 2006. Since
>>ff03ua is very similar to ff03, please find out what GB parameters have
>>been used with ff03, for example, from Y. Duan's group. Maybe Dr. Yong
>>Duan can comment on this.
>>
>>
>
>Let me comment on this point: the key item would be the intrinsic Born radii
>for the "united atom" carbons. There aren't any analogs of this type of atom
>in ff03, so some exploration and parameterization seems inevitable.
>
>Maybe I am missing something here: I haven't yet had a chance to use ff03ua
>myself.
>
>....regards...dac
>
>
>
>
--
====================================================
Ray Luo, Ph.D.
Department of Molecular Biology and Biochemistry
University of California, Irvine, CA 92697-3900
Office: (949)824-9528 Lab: (949)824-9562
Fax: (949)824-8551 e-mail: rluo.uci.edu
Home page: http://rayl0.bio.uci.edu/rayl/
====================================================
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Jul 09 2006 - 06:07:08 PDT
