Hi Ray,
first of all thank you for your prompt answer !
Please see my comments:
Dne Tue, 07 May 2013 23:52:26 +0200 Ray Luo, Ph.D. <ray.luo.uci.edu>
napsal/-a:
> Marek,
>
> Amber PBSA and APBS should give you the same energies, within
> numerical accuracy, if you use the same parameters, such as dielectric
> constants, radius definition, surface definition and tension etc. This
> has nothing to do with which package you use. If you see different
> energies, you are most likely using different parameters, i.e.
> different default values.
I am not sure if providing to APBS exactly the same values of corresponding
parameters as used with Amber PBSA will 100% guarantee the same results.
For example the apolar contribution to the solvation energy is calculated
in both cases (Amber APBS, APBS) using different approaches.
( see here
http://www.poissonboltzmann.org/apbs/examples/solvation-energies#TOC-Apolar-solvation
)
so I can imagine that for example the identical values of "Surface tension
for apolar energies"
i.e. using "cavity_surften" = "gamma" might not in general ensure the same
results in both
softwares even if all the rest of parameters is identical and all the rest
of calculations are
done using exactly the same theories/approaches.
For me would be interesting your opinion especially here in case of
setting cavity_surften
parameter in situation when APBS is used ( sander_apbs ).
There are 3 possible candidates:
0.00542 kcal/mol/A^2 (Amber PBSA default for "inp=1 radiopt=0")
0.0378 kcal/mol/A^2 (Amber PBSA default for "inp=2 radiopt=1")
0.0251 kcal/mol/A^2 (APBS default)
even if it seems from my experience that "sander_apbs=1" setting
is possible to use now just with "inp=1 radiopt=0" setting I
would not consider value 0.00542, simply because the general schema
of calculating whole apolar contribution in APBS seem to me to be closer
to that Amber PBSA with inp=2 setting.
So I would decide between 0.0378 and 0.0251 having tendency to use rather
the APBS default 0.0251 simply beacuse it might be a more suitable to
the theoretic approach used in APBS for which was this parameter optimized.
The confirmation test here might be e.g. comparison of apolar parts of
solvation energies calculated with Amber PBSA and sander.APBS
on some non-problematic systems like peptides.
The more proper value for APBS calculation (from those 0.0378 and 0.0251 )
will be
perhaps that which provides smaller difference in apolar energies between
Amber PBSA and sander.APBS.
What do you think about this idea ?
> Developers often set different default parameters for different
> intended usages. In Amber/PBSA, we set these values for MMPBSA
> applications. In APBS, these values are set for some other
> applications. I believe you can find these in their user manual.
It seems that the main APBS applications are the same as applications of
Amber PBSA
see here:
http://www.poissonboltzmann.org/apbs
http://www.poissonboltzmann.org/apbs/examples/binding-energies
Best wishes,
Marek
>
> Ray
>
> On Tue, May 7, 2013 at 11:27 AM, Marek Maly <marek.maly.ujep.cz> wrote:
>> Hi all,
>>
>> I have question regarding the optimal parameters
>> values when one using APBS/sander.APBS via MMPBSA.py or MMPBSA.pl
>> interface.
>>
>> My guess is that in such a case perhaps the APBS defaults
>> ( which are reported e.g. here
>> http://mccammon.ucsd.edu/iapbs/usersguide/using-amber.html#AEN187 )
>> are the best choice to get the most accurate overall solvation energies
>> using this module.
>>
>> However these defaults differ a bit from those of Amber PB.
>>
>> For example:
>>
>> Internal dielectric constant:
>> Amber default = 1, APBS default = 2
>>
>> External dielectric constant:
>> Amber default = 80, APBS default = 78.4
>>
>> Cavity surface tension :
>> Amber defaults:
>>
>> 0.0378 kcal/mol/A^2 (for INP = 2 and RADIOPT = 1)
>> 0.00542 kcal/mol/A^2 (for INP = 1 and RADIOPT = 0)
>>
>>
>> APBS default = 0.0251 kcal/mol/A^2
>> (It seems that APBS module is compatible actually just with RADIOPT = 0
>> )
>> so here we have quite a big difference 0.00542 versus 0.0251.
>>
>>
>> so to be sure I am just asking this question, also to learn about
>> another important parameters which should be eventually changed
>> from the Amber defaults to APBS ones (if any).
>>
>> I would like also notice probable error in translating ISTRNG parameter
>> into
>> APBS parameter "ionc".
>>
>> In case of PB section of the mm_pbsa.in file (input file for PERL
>> version),
>> there is explicitly stated that the ISTRNG parameter is in mM units
>> which
>> e.g. means that the
>> right value for the physiological ionic strength of Na+,Cl- ions is 150
>> here.
>>
>> However when translated for the sander.APBS (in &apbs section) it should
>> be perhaps converted
>> to APBS native units for this parameter (here "ionc") which is in molar
>> (M) units.
>>
>> So having in PB section "ISTRNG=150", for sander.APBS it should be
>> translated
>> (written in &apbs section) as "ionc=0.15, 0.15" not "ionc=150,150" as it
>> is doing now
>> (Amber12/AT13 fully patched)).
>>
>> In case of MMPBSA.py there is "istrng" value also just simply copied
>> into
>> "ionc"
>> parameter, but here it is OK because here "istrng" is in M units so no
>> "value-conversion"
>> is necessary.
>>
>> Thanks in advance for any comments to this topic.
>>
>> Best wishes,
>>
>> Marek
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> --
>> Tato zpráva byla vytvořena převratným poštovním klientem Opery:
>> http://www.opera.com/mail/
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
> __________ Informace od ESET NOD32 Antivirus, verze databaze 8307
> (20130507) __________
>
> Tuto zpravu proveril ESET NOD32 Antivirus.
>
> http://www.eset.cz
>
>
>
--
Tato zpráva byla vytvořena převratným poštovním klientem Opery:
http://www.opera.com/mail/
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue May 07 2013 - 21:00:02 PDT
