AMBER Archive (2006)

Subject: Re: AMBER: why has "ELE " a large fluctuation in MMP BSA?

From: JunJun Liu (ljjlp03_at_gmail.com)
Date: Tue Jun 20 2006 - 20:08:32 CDT

Dear Prof.Simmerling,

Thanks! But as Prof.Cheatham also noticed, when using snapshots from
explict solvation simulation, Delta(INT) from MM-PBSA is not zero, in
another words, the internal energy from each component can't be canceled.
Although post-processing methods like MM-PBSA may give large fluctuations,
I don't think the internal energy can be affected. How do you think?

Thanks!

On Tue, 20 Jun 2006 22:54:51 -0300, Carlos Simmerling
<carlos_at_csb.sunysb.edu> wrote:

> no, I do not think the explicit water is wrong.
> this is just the way it is- postprocessing
> may give large fluctuations. it doesn't mean anything
> is wrong...
> good luck!
>
> JunJun Liu wrote:
>
>> Thanks a lot to Prof.Simmerling and Prof.Cheatham,
>>
>> I understand now. It may be a problem that caused by too many water
>> molecules, which make a major energy contribution to the whole system.
>> This is possible, because our simulated system contains 11,594
>> DNA+Protein+Ligand atoms v.s. 93,843 water atoms. If this is the case,
>> can I say it's a disadvantage of explicit solvation model of MD, in
>> which systems containing large amount of solvent molecules may
>> possible not be treated correctly? Is it right?
>>
>> Thanks!
>>
>> On Tue, 20 Jun 2006 21:19:10 -0300, Carlos Simmerling
>> <carlos_at_csb.sunysb.edu> wrote:
>>
>>> of course my suggestions are just ideas, it is entirely possible
>>> that you have some error. what i meant was that your
>>> observations are not that surprising.
>>>
>>> JunJun Liu wrote:
>>>
>>>> Dear Dr.Simmerling,
>>>>
>>>> Thanks a lot for the responses. I may not state the problem clearly.
>>>> In this case, GB model is not used. What I did is like the
>>>> followings:
>>>> 1). MD simulation with explicit solvation box
>>>> 2). extract snapshots from stable MD trajectory and remove waters
>>>> and counterions
>>>> 3). Perform 1 step(single point) gas phase minimization against
>>>> those snapshots
>>>>
>>>> I know the energy from MD simulation contains solvent contribution,
>>>> and this energy should be different with energy from gas phase. But
>>>> it's abnormal that the gas-phase electrostatic energy fluctuate too
>>>> much, which can be up to ~600kcal/mol. This is quite strange! We
>>>> checked the RMSD of selected snapshots, the values are less than
>>>> 0.3A, indicating the selected structures are very close! Any
>>>> suggestions? Thanks!
>>>>
>>>> Regards!
>>>>
>>>> Liu
>>>>
>>>> On Tue, 20 Jun 2006 20:10:48 -0300, Carlos Simmerling
>>>> <carlos_at_csb.sunysb.edu> wrote:
>>>>
>>>>> possible things to consider:
>>>>>
>>>>> 1) the GB solvated system is much smaller, so larger fluctuations
>>>>> are expected (check
>>>>> a stat mech book)
>>>>> 2) you are calculating energies with a different energy function
>>>>> than you used to
>>>>> generate the simulation. It is entirely possible that the energy
>>>>> fluctuations for the
>>>>> trajectory from explicit water differ from what you would get if
>>>>> you used structures
>>>>> that were actually sampled in the GB model.
>>>>> 3) the ele in the explicit water includes solvation, and the GB one
>>>>> does not.
>>>>> it is well known that these terms tend to fluctuate in the
>>>>> opposite direction
>>>>> so leaving the solvation energy out will give higher fluctuations.
>>>>>
>>>>>
>>>>>
>>>>> JunJun Liu wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> It's acutally a sander related problem. The sander program
>>>>>> produces a stable trajectory when performing a MD simulation in
>>>>>> explicit water box, but gives out very unstable energy curve if
>>>>>> doing vacuum single-point energy calculation against snapshots
>>>>>> of this MD trajectory(with waters and counterions removed). What
>>>>>> could be the reason? Since it's a DNA+Protein system, does this
>>>>>> mean the force field(ff99) can't describe the system well? Or
>>>>>> caused by other problem? Any suggestions? Thanks!
>>>>>>
>>>>>> Regards!
>>>>>>
>>>>>> Liu
>>>>>>
>>>>>> On Tue, 20 Jun 2006 10:47:44 -0300, yxiong99 <yxiong99_at_163.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Dear sir,
>>>>>>> I have a strange problem when I did MM_PBSA. My system is
>>>>>>> protein-DNA-ligand complex. The receptor is protein-DNA complex,
>>>>>>> and the ligand is a small molecule. I get 50 points from a
>>>>>>> stable 200 ps MD simulation. In that simulation, it is obvious
>>>>>>> that energy for every step in 200ps simulation don't change so
>>>>>>> much. However after I did MM_PBSA(I only did MM section to
>>>>>>> test), I found "ELE " has a large fluctuation.
>>>>>>> Could you please give me some suggestion on it?
>>>>>>> MD results:
>>>>>>> -------------------------------------------------------------------------------------
>>>>>>> A V E R A G E S O V E R 100000 S T E P S
>>>>>>>
>>>>>>> NSTEP = 100000 TIME(PS) = 1200.000 TEMP(K) = 297.94
>>>>>>> PRESS = 1.0
>>>>>>> Etot = -1755294.9287 EKtot = 6319.6873 EPtot
>>>>>>> = -1761614.6160
>>>>>>> BOND = 1374.4608 ANGLE = 5579.4236 DIHED
>>>>>>> = 2833.4301
>>>>>>> 1-4 NB = 2212.7443 1-4 EEL = -5183.5817 VDWAALS
>>>>>>> = -5710.5417
>>>>>>> EELEC = -1762720.5514 EHBOND = 0.0000 RESTRAINT
>>>>>>> = 0.0000
>>>>>>> EKCMT = 1.9203 VIRIAL = -24.2883 VOLUME
>>>>>>> = 1239972.2032
>>>>>>> Density
>>>>>>> = 0.8688
>>>>>>> Ewald error estimate: 0.9679E+00
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> R M S F L U C T U A T I O N S
>>>>>>>
>>>>>>> NSTEP = 100000 TIME(PS) = 1200.000 TEMP(K) = 2.04
>>>>>>> PRESS = 51.9
>>>>>>> Etot = 3.7961 EKtot = 43.3440 EPtot
>>>>>>> = 45.1668
>>>>>>> BOND = 29.6708 ANGLE = 45.6035 DIHED
>>>>>>> = 26.7273
>>>>>>> 1-4 NB = 13.7112 1-4 EEL = 38.6013 VDWAALS
>>>>>>> = 35.0353
>>>>>>> EELEC = 47.8923 EHBOND = 0.0000 RESTRAINT
>>>>>>> = 0.0000
>>>>>>> EKCMT = 0.9180 VIRIAL = 1390.4239 VOLUME
>>>>>>> = 142.3465
>>>>>>> Density
>>>>>>> = 0.0001
>>>>>>> Ewald error estimate: 0.2632E-04
>>>>>>>
>>>>>>> MMPBSA results:
>>>>>>> -------------------------------------------------------------------------------------
>>>>>>> # COMPLEX
>>>>>>> RECEPTOR LIGAND
>>>>>>> # ----------------------- -----------------------
>>>>>>> -----------------------
>>>>>>> # MEAN STD MEAN
>>>>>>> STD MEAN STD
>>>>>>> # ======================= =======================
>>>>>>> =======================
>>>>>>> ELE -23558.26 691.21 -23410.78
>>>>>>> 683.97 61.30 0.02
>>>>>>> VDW -2745.00 28.83 -2730.24
>>>>>>> 29.06 16.89 0.01
>>>>>>> INT 13160.44 52.99 13382.69
>>>>>>> 54.40 54.10 0.04
>>>>>>> GAS -13142.81 696.51 -12758.34
>>>>>>> 689.12 132.29 0.03
>>>>>>> # DELTA
>>>>>>> # -----------------------
>>>>>>> # MEAN STD
>>>>>>> # =======================
>>>>>>> ELE -208.78 64.66
>>>>>>> VDW -31.64 3.68
>>>>>>> INT -276.34 17.50
>>>>>>> GAS -516.77 67.28
>>>>>>>
>>>>>>> 　　　　　　　　Ying Xiong
>>>>>>> 　　　　　　　　yxiong99_at_163.com
>>>>>>> 　　　　　　　　　　2006-06-19
>>>>>>>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu