Re: Dual Topology in FEP Calculations - How to Calculate Electrostatics and van der Waals Free Energies Separately?

From: Jérôme Hénin (jerome.henin_at_ibpc.fr)
Date: Tue Jul 24 2018 - 02:55:40 CDT

With alchVdwLambdaEnd=0 all atoms (outgoing and incoming groups) will be
coupled to the environment at all times through the Lennard-Jones potential
(lambda_LJ = 1).

I'm now realizing that the transformation you had in mind doesn't work,
because it won't sum to the total mutation free energy. What you can do is
a series of three simulations:

1 - discharge group A with only LJ of A present
2 - vanish LJ of A while growing LJ of B
3 - charge group B with only LJ of B present

Then the sum of free energies from steps 1 and 3 will give you an idea of
the electrostatic contribution, and step 2 gives the vdW part.

Jerome

On Tue, 24 Jul 2018 at 02:54, Nassar, Omneya <omnassar_at_utmb.edu> wrote:

> You are correct; my question is more complex. I am asking about the
> outgoing and incoming atoms for the alchemical transformation when I have
> alchVdwLambdaEnd=0 and alchElecLambdaStart=0. So for my system, I am
> starting with NADH and transforming it to ADP. At the first window (0/32 to
> 1/32), does the simulation only use the NADH van der waals shape and
> charges while ignoring ADP’s shape for the Free Energy calculation or is
> the shape of both NADH and ADP used in the calculation and only the charges
> of NADH are used for the simulation?
>
> Attached is a diagram of what I am asking. If it is still confusing,
> please let me know. I really appreciate your patience and thanks again for
> helping me.
>
>
>
> Omneya
>
>
>
> *From: *Jérôme Hénin <jerome.henin_at_ibpc.fr>
> *Sent: *Monday, July 23, 2018 4:15 PM
> *To: *Nassar, Omneya <omnassar_at_UTMB.EDU>
> *Cc: *Namd Mailing List <namd-l_at_ks.uiuc.edu>
> *Subject: *Re: namd-l: Dual Topology in FEP Calculations - How to
> Calculate Electrostatics and van der Waals Free Energies Separately?
>
>
>
> All particles are present at all times; their interactions vary. I think
> there was more to your question, but I didn't get it.
>
>
>
> Jerome
>
>
>
> On Sat, 21 Jul 2018 at 14:59, Nassar, Omneya <omnassar_at_utmb.edu> wrote:
>
> Thank you Jerome for your prompt response. If you could further answer
> this question, I will greatly appreciate it. So at FEP lambda 0.0625 to
> 0.125, if I have 0,0 for alchVdwLambdaEnd, alchElecLambdaStart,
> respectively, do I have the geometry of NADH and ADP present or do I have
> the geometry of the intermediate at 0.0625 to 0.125?
>
>
>
> Thanks!
>
> Omneya
>
>
> On Jul 21, 2018, at 6:02 AM, Jérôme Hénin <jerome.henin_at_ibpc.fr> wrote:
>
> Dear Omneya,
>
>
>
> Your simulation parameters seem correct to me, as does your understanding
> of the alchemical process in this implementation.
>
>
>
> Best,
>
> Jerome
>
>
>
> On Fri, 20 Jul 2018 at 22:07, Nassar, Omneya <omnassar_at_utmb.edu> wrote:
>
> Hello NAMD staff,
>
> I am trying to properly implement the dual topology paradigm in Free
> Energy Perturbation technique for a mutation of NADH to ADP. I want to
> calculate two things: 1) the van der Waals free energy of NADH changing to
> ADP and 2) the electrostatics of NADH changing to ADP. Thus, I should have
> values for dGvdw and dGelec from two separate simulations. However, after
> reading over the documentation (
> http://www.ks.uiuc.edu/Research/namd/2.7/ug/node53.html
> <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.proofpoint.com%2Fv2%2Furl%3Fu%3Dhttps-3A__na01.safelinks.protection.outlook.com_-3Furl-3Dhttp-253A-252F-252Fwww.ks.uiuc.edu-252FResearch-252Fnamd-252F2.7-252Fug-252Fnode53.html-26data-3D02-257C01-257Comnassar-2540UTMB.EDU-257C099dbcc3f2c0496789c708d5eef97f81-257C7bef256d85db4526a72d31aea2546852-257C0-257C1-257C636677677711073805-26sdata-3DahfoJSCH4i7e6sVP6dY4rWsBU5WB372Idz9USilrGXE-253D-26reserved-3D0%26d%3DDwMFaQ%26c%3DOCIEmEwdEq_aNlsP4fF3gFqSN-E3mlr2t9JcDdfOZag%26r%3DjUfnSyKkfkyVRBIUzlG1GSGGZAZGcznwr8YliSSCjPc%26m%3Dojx3DcDQE2hr2yiIPoSjDlagDXtFEW9J1hoRZUO6cVc%26s%3DiFIxvgyH8sawhze2_wMWmonbVfmpzGFOA8R3KlI9fgk%26e%3D&data=02%7C01%7Comnassar%40UTMB.EDU%7Cdfc37453346e4ab070bd08d5f0e16c7b%7C7bef256d85db4526a72d31aea2546852%7C0%7C0%7C636679773341569223&sdata=ovmkZX%2B6ciRgIIeBp%2FzVWFEPG28gn2iJqlLZZ5w8Kcc%3D&reserved=0>),
> it seems that it is not possible to properly separate van der Waals and
> electrostatics calculations using FEP.
>
> However, in contrast to my understanding of the above documuntation and
> according to this very general response (
> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2010-2011/1722.html
> <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.proofpoint.com%2Fv2%2Furl%3Fu%3Dhttps-3A__na01.safelinks.protection.outlook.com_-3Furl-3Dhttp-253A-252F-252Fwww.ks.uiuc.edu-252FResearch-252Fnamd-252Fmailing-5Flist-252Fnamd-2Dl.2010-2D2011-252F1722.html-26data-3D02-257C01-257Comnassar-2540UTMB.EDU-257C099dbcc3f2c0496789c708d5eef97f81-257C7bef256d85db4526a72d31aea2546852-257C0-257C1-257C636677677711073805-26sdata-3DWOaQxK6X-252B038lGSGPbfUnswBbf0vH7-252Bwp-252Ffqfr42-252B74-253D-26reserved-3D0%26d%3DDwMFaQ%26c%3DOCIEmEwdEq_aNlsP4fF3gFqSN-E3mlr2t9JcDdfOZag%26r%3DjUfnSyKkfkyVRBIUzlG1GSGGZAZGcznwr8YliSSCjPc%26m%3Dojx3DcDQE2hr2yiIPoSjDlagDXtFEW9J1hoRZUO6cVc%26s%3D9dY63UuQAwNwG25E8c_uCCkkkQg-oVhpS_0XGIMXKjM%26e%3D&data=02%7C01%7Comnassar%40UTMB.EDU%7Cdfc37453346e4ab070bd08d5f0e16c7b%7C7bef256d85db4526a72d31aea2546852%7C0%7C0%7C636679773341579232&sdata=9MVdvK6kjkvEt2TmdQHc8LK%2BwknzcMlsDGvGc9XE%2BfU%3D&reserved=0>),
> it is claimed that it is possible to separate vanderwaals and
> electrostatics.
>
> Below is the input for van der Waals and electrostatics for each lambda
> value (I am simulating each lambda window independently).
>
> ____________________________________________________________
> VANDERWAALS CALCULATION/SIMULATION:
>
> # FEP PARAMETERS
>
> source fep.tcl
>
> alch on
> alchType FEP
> alchFile N3A_ion.fep
> alchCol B
> alchOutFile $outname.fepout
> alchOutFreq 500
>
> alchVdwLambdaEnd 1.0
> alchElecLambdaStart 1.0
> alchVdWShiftCoeff 5.0
> alchDecouple on
> alchEquilSteps 1000
> set numSteps 6251000
>
>
> runFEP 0.6875 0.71875 0.03125 $numSteps
> ______________________________________________________________
> ELECTROSTATICS CALCULATION/SIMULATION:
>
> # FEP PARAMETERS
>
> source fep.tcl
>
> alch on
> alchType FEP
> alchFile N3A_ion.fep
> alchCol B
> alchOutFile $outname.fepout
> alchOutFreq 500
>
> alchVdwLambdaEnd 0.0
> alchElecLambdaStart 0.0
> alchVdWShiftCoeff 5.0
> alchDecouple on
> alchEquilSteps 1000
> set numSteps 6251000
>
>
> runFEP 0.6875 0.71875 0.03125 $numSteps
>
> Using the input examples above, am I properly separating electrostatics
> and van der waals calculations? If so, how is the geometry of the molecule
> affected when I run the electrostatics calculation? Do I have both NADH and
> ADP present geometrically, but only the charges are changing?
>
> Your help is invaluable and I really appreciate your response.
>
> Thank you!
>
> Omneya Nassar
> PhD Candidate in Pharmacology and Toxicology
> University of Texas Medical Branch
> 301 University Blvd
> Galveston, TX 77555
> E-mail: omnassar_at_utmb.edu
> Phone: 409.772.0731
>
>
>

This archive was generated by hypermail 2.1.6 : Tue Sep 17 2019 - 23:20:05 CDT