Re: Output of thermodynamic integration

From: Jérôme Hénin (jhenin_at_ifr88.cnrs-mrs.fr)
Date: Sat Jan 28 2012 - 12:52:45 CST

You are right, I should have seen it in the output you attached. This
is beginning to look like a bug. I don't know the TI code well, and I
don't have the time to go into the details right now. I hope that
another developer can chime in...

Best,
Jerome

On 28 January 2012 17:31, Okazaki, Keiichi (NIH/NIDDK) [F]
<keiichi.okazaki_at_nih.gov> wrote:
> Dear Jerome,
>
> Thank you for your suggestion.
> However, 1 is set in my alchFile for the flag, that is, it goes in the same direction with the global lambda.
> So, I am still not sure why Elec_dU/dl takes some values even though ele is not turned on yet.
>
> Regards,
> Kei-ichi
> ________________________________________
> From: heninj_at_gmail.com [heninj_at_gmail.com] On Behalf Of Jérôme Hénin [jhenin_at_ifr88.cnrs-mrs.fr]
> Sent: Saturday, January 28, 2012 5:25 AM
> To: Okazaki, Keiichi (NIH/NIDDK) [F]
> Cc: namd-l_at_ks.uiuc.edu
> Subject: Re: namd-l: Output of thermodynamic integration
>
> Dear Kei-ichi,
>
> If you are decoupling the solute from the solvent, it probably has -1
> flags in the alchFile. In this case, the lambdas for the solute are
> reversed with respect to the "global" lambda, and the first thing that
> happens is that electrostatics gets decoupled (which is the right
> thing to do).
>
> Best,
> Jerome
>
>
> On 28 January 2012 01:16, Kei-ichi Okazaki <keiichi.okazaki_at_nih.gov> wrote:
>> Dear NAMD users,
>>
>> I am trying to calculate solvation free energy of a small molecule by
>> thermodynamic integration (TI).
>> And I have a question about output file of TI.
>> I set up configuration file for TI with soft-core potential as below.
>> I set "alchElecLambdaStart 0.5" and "alchVdwLambdaEnd 0.5" to separate the
>> scaling of vdW and electrostatics, that is, vdW is turned on first for
>> 0<=lambda<=0.5 and then, ele is turned on for 0.5<=lambda<=1.0.
>> And in the output, Elec_dU/dl and vdW_dU/dl are reported as below.
>> When I saw the output of lambda=0.1, for example, I was confused because
>> Elec_dU/dl takes some values even though ele is not turned on yet.
>> I suspect that it just reports Elec_dU/dl for lambda=0.0, but it cannot be
>> used for TI calculation because there should not be any contribution from
>> ele for lambda<0.5.
>> Is my guess right? I appreciate any suggestions.
>>
>> Thanks in advance.
>> Kei-ichi Okazaki, PhD
>>
>> ## configuration file #########################################
>> # Free energy perturbation
>> alch           On             # Enable alchemical simulation module
>> alchType       ti             # Set method to thermodynamic integration
>> alchFile       ../../Common/h2p.box20.alchem.pdb   # PDB file with
>> perturbation flags
>> alchCol        B              # Perturbation flags in Beta column
>> alchOutfile    $outputname.out
>> alchOutFreq    1000
>> alchEquilSteps 500000
>>
>> alchVdWShiftCoeff    6.0       # Enable soft-core vdW potential
>> alchElecLambdaStart  0.5       # Introduce electrostatics for lambda > 0.0
>> alchVdwLambdaEnd     0.5
>> alchLambda 0.1
>> alchLambda2 0.1
>>
>> ## output file #################################################
>> #       STEP      Elec_dU/dl      Elec_avg        vdW_dU/dl      vdw_avg
>>   Elec_dU/dl      Elec_avg      vdW_dU/dl       vdw_avg       PME_dU/dl
>>  PME_avg
>> # <---------------------PARTITION 1------------------------>
>> <---------------------PARTITION 2--------------------->
>> #NEW TI WINDOW: LAMBDA 0.1
>> #PARTITION 1 VDW LAMBDA 0.2
>> #PARTITION 1 ELEC LAMBDA 0
>> #PARTITION 2 VDW LAMBDA 1
>> #PARTITION 2 ELEC LAMBDA 0.8
>>
>> TI        0       -452.9840      -452.9840        -9.6522        -9.6522
>>     0.0000         0.0000         0.0000         0.0000
>> TI     1000       -228.3854      -228.6098         1.7409         1.7296
>>     0.0000         0.0000         0.0000         0.0000
>>
>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:21:36 CST