Dear all,
I forward an answer below which may be of wider interest.
Best regards
Holger
>
> On 03.02.12, Nadine Homeyer wrote:
> > Dear Therese and Quin,
> >
> > it is reasonable that the results for ECAVITY in the *.all.out files of AMBER9 and AMBER11 differ, because the procedure for the calculation of the nonpolar energy in mm_pbsa.pl has changed. In AMBER9 the nonpolar energy was calculated before the PBSA energies were written to the *.all.out files, whereas in AMBER11 the calculated surface area is printed to the *all.out files and the nonpolar energy is determined in the final step of the calculation and thus only printed to the <PREFIX>_statistics.out file. Thus, in AMBER9 the CAVITY_SURFTEN and CAVITY_OFFSET values are considered before the values are written to the *.all.out files, whereas in AMBER11 the respective CAVITY_SURFTEN and CAVITY_OFFSET values are set to 1 and 0 in the PBSA calculation (pbsa_<com|rec|lig>.in file) and the SURFTEN and SURFOFF values provided in mmpbsa.in (corresponding to the CAVITY_SURTEN and CAVITY_OFFSET) are only applied in the statistical analysis. The values of the nonpolar solvation free energy in the <PREFIX>_statistics.out files of AMBER9 and AMBER11 should not significantly differ and did not do so in my test case. Therefore Bugfixes related to the CAVITY_SURFTEN and CAVITY_OFFSET values are not necessary. You might however have problems with the statistical analysis in AMBER11, when you run PB with the Tan and Luo method and GB set to 0. This is related to an actual bug in AMBER11 mm_pbsa.pl, which has been fixed in the AMBER12 release. The easiest way to circumvent this problem is to set GB to 1 in the mmpbsa.in file. If you do not want to redo the calculation and / or are not interested in the GB energies, please do the following:
> > 1.) Create a folder called "lib" in your analysis directory
> > 2.) Copy the $AMBERHOME/src/mm_pbsa/mm_pbsa_statistics.pm script to the "lib" directory
> > 3.) Change line 1933 of the mm_pbsa_statistics.pm file from
> > $check_MS = 1 if $line eq "MS"
> > into
> > $check_MS = 1 if $line eq "MS" || $line eq "PB";
> > 4.) Generate a copy of the mm_pbsa_statistics.pl file in your analysis directory
> > 5.) In the copied mm_pbsa_statistics.pl file shift the line "use mm_pbsa_statistics;" to the end of the library definition section and add the following line before it:
> > use lib("/<path to your analysis folder>/lib");
> > 6.) Remove the flag "MS" from the headers of the *.all.out files
> > 7.) Run your statistical analysis
> > The nonpolar energies shown in the <PREFIX>_statistics.out of AMBER11 should not differ considerably from those you obtain with AMBER9.
> >
> > Best regards,
> > Nadine
> >
> >
> > >
> > >-------- Original Message --------
> > >From: Qin Cai
> > >Date: Feb 2, 2012 9:38:57 PM
> > >Subject: Re: [AMBER] Fwd: about using MMPBSA Tan and Luo method in AMBER 11: discrepancy in ECAVITY values
> > >To: AMBER Mailing List
> > >Cc: terez.pasteur.fr
> > >
> > >>It's really odd that CAVITY_SURFTEN and CAVITY_OFFSET doesn't change ECAVITY. After looking into the code, I found the CAVITY_SURFTEN and CAVITY_OFFSET are fixed to 1 and 0, respectively, in mm_pbsa.pl. So what you got for ECAVITY was actually the SAS area or the volume enclosed by SAS, depending on whether use_sav is 0 or 1. Please apply the attached patch in your $AMBERHOME so that your own CAVITY_SURFTEN and CAVITY_OFFSET will take effect.
> > >>
> > >>Best,
> > >>Qin
> > >>
> > >>
> > >>
> > >>
> > >>On Feb 1, 2012, at 9:49 AM, Ray Luo, Ph.D. wrote:
> > >>
> > >>>Therese,
> > >>>
> > >>>I'm forwarding this to Qin for a look ...
> > >>>
> > >>>All the best,
> > >>>Ray
> > >>>
> > >>>---------- Forwarded message ----------
> > >>>From: Therese Malliavin<terez.pasteur.fr>
> > >>>Date: Wed, Feb 1, 2012 at 12:48 AM
> > >>>Subject: [AMBER] about using MMPBSA Tan and Luo method in AMBER 11:
> > >>>discrepancy in ECAVITY values
> > >>>To: amber.ambermd.org
> > >>>
> > >>>
> > >>>Dear AMBER Netters,
> > >>>
> > >>>I have some problems in using the Tan and Luo method with the
> > >>>mm_pbsa.pl script from the AMBER 11 package. Indeed, the results
> > >>>obtained with AMBER 11 are quite different from the ones obtained
> > >>>with AMBER 9, as shown by the results obtained on the same snapshot:
> > >>>
> > >>>AMBER9:
> > >>>1
> > >>> BOND = 1230.9291 ANGLE = 3154.2737 DIHED = 4051.0296
> > >>> VDWAALS = -3251.8837 EEL = -29793.4172 HBOND = 0.0000
> > >>> 1-4 VDW = 1505.8384 1-4 EEL = 14953.4064 RESTRAINT = 0.0000
> > >>>corrected reaction field energy: -9780.831722
> > >>> ECAVITY = 890.3956 EDISPER = -997.9678
> > >>>
> > >>>
> > >>>AMBER11:
> > >>>1
> > >>> BOND = 1230.9314 ANGLE = 3154.2736 DIHED = 4051.0294
> > >>> VDWAALS = -3251.8840 EEL = -29793.4156 HBOND = 0.0000
> > >>> 1-4 VDW = 1505.8386 1-4 EEL = 14953.4044 RESTRAINT = 0.0000
> > >>>corrected reaction field energy: -9734.978826
> > >>> ECAVITY = 20458.5423 EDISPER = -997.6458
> > >>>
> > >>>As you can see, all energy terms are closely similar, except the
> > >>>ECAVITY which is much more larger in AMBER11 than in AMBER9. This much
> > >>>larger energy
> > >>>term then produce nonpolar solvation term of order of value larger than the
> > >>>solvation polar term, and I find this quite strange.
> > >>>
> > >>>In mm_pbsa.pl with AMBER 11, I have been using the input parameter INP
> > >>>in place of NOPT (NOPT seems to be obsolete, as far as I
> > >>>understood?). Using INP permittted to get non-zero values for ECAVITY energy.
> > >>>Then, I used SURFTEN and OFFSET options in place of CAVITY_SURFTEN
> > >>>and CAVITY_OFFSET options. But, using SURFTEN, OFFSET or
> > >>>CAVITY_SURFTEN, CAVITY_OFFSET does not change the ECAVITY value.
> > >>>
> > >>>Also, concerning the Tan and Luo method, I have seen that a article
> > >>>describing it is still quoted "in preparation" in the
> > >>>documentation. But, I had the impression that the method was published in:
> > >>>.Article{ tan07,
> > >>> author = {Tan C and Tan YH and Luo R},
> > >>> title = {Implicit nonpolar solvent models},
> > >>> journal = {J Phys Chem B},
> > >>> volume = {111},
> > >>> pages = {12263-12274},
> > >>> year = {2007},
> > >>>}
> > >>>
> > >>>I would much appreciate to get some information about using the Tan et
> > >>>Luo method in AMBER 11.
> > >>>
> > >>>Best regards,
> > >>>Thérèse
> > >>>
> > >>>Thérèse Malliavin
> > >>>Unité de Bioinformatique Structurale
> > >>>Institut Pasteur and CNRS UMR 3528
> > >>>Paris, France
> > >>>
>
> --
> ++++++++++++++++++++++++++++++++++++++++++++++++++
> Dr. Holger Gohlke
> Professor fuer Pharmazeutische/Medizinische Chemie
>
> Heinrich-Heine-Universitaet Duesseldorf
> Institut fuer Pharm. und Med. Chemie
> Universitaetsstr. 1
> 40225 Duesseldorf
> Germany
>
> Tel.: (+49) 211-81-13662; Fax: (+49) 211-81-13847
> Email: gohlke.uni-duesseldorf.de
> URL: http://cpclab.uni-duesseldorf.de
> ++++++++++++++++++++++++++++++++++++++++++++++++++
--
++++++++++++++++++++++++++++++++++++++++++++++++++
Dr. Holger Gohlke
Professor fuer Pharmazeutische/Medizinische Chemie
Heinrich-Heine-Universitaet Duesseldorf
Institut fuer Pharm. und Med. Chemie
Universitaetsstr. 1
40225 Duesseldorf
Germany
Tel.: (+49) 211-81-13662; Fax: (+49) 211-81-13847
Email: gohlke.uni-duesseldorf.de
URL: http://cpclab.uni-duesseldorf.de
++++++++++++++++++++++++++++++++++++++++++++++++++
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Feb 04 2012 - 02:00:02 PST
