[gmx-users] toluene tutorial sc-power value

Maik Goette mgoette at mpi-bpc.mpg.de
Fri Feb 15 15:36:26 CET 2008 

hi

10 points is strange ;)
11 is the number with equally spaced points. ANyway, it should suffice 
to get a more or less proper number.
I guess, they use simpsons integration, what should be ok. I compared my 
scripts results, which uses simpson, with those of xmgrace and they fit.
The longer you sample each lambda value and the more lambda values you 
sample, the better the results, but with this system, I think, 11 lambda 
values with 1ns each should be enough.
Be aware of the fact, that you should cut away a few picoseconds of each 
dgdl-file for equilibration purposes (100-200).

The topology seems not too bad. I'm not used to that forcefield, though.

As a general suggestion, its always better to sample many lambda points 
and each longer, than wasting time with using sc-power=2 :)

Regards

Maik Goette, Dipl. Biol.
Max Planck Institute for Biophysical Chemistry
Theoretical & computational biophysics department
Am Fassberg 11
37077 Goettingen
Germany
Tel.  : ++49 551 201 2310
Fax   : ++49 551 201 2302
Email : mgoette[at]mpi-bpc.mpg.de
         mgoette2[at]gwdg.de
WWW   : http://www.mpibpc.gwdg.de/groups/grubmueller/


cilpa at users.csc.fi wrote:
> hi
> 
> concerning the cycle i saw this error since i started the tutorial, i paid
> attention to that but as you said it's not that important since my energy
> value is too high. i'm right now doing the problem again with sc-power=2
> and also 10 points is it enough to have an approximate correct value since
> i have to integrate the average energy plot by xmgrace, i don't know what
> kind of integration they use, trapezoid? or... it could have an effect if
> i don't have enough point. but then .. here attached is my top. file.
> regards
> geraldine
> 
> 
> 
> 
> 
> 
> 
> On Fri, February 15, 2008 11:07 am, Maik Goette wrote:
>> I had a veeery short look into the tutorial and found an error.
>>
>>
>> DG_hyd=DG_1-DG_3-DG_2 for the cycle is wrong.
>>
>>
>> The correct cycle would be:
>>
>>
>> DG_1+DG_3=DG_hyd+DG_2
>>
>>
>> =>DG_hyd=DG_1-DG_2+DG_3 with DG_3 =0
>>
>>
>> So its:
>> DG_hyd=DG_1-DG_2 instead of DG_hyd=DG_1-DG_3
>>
>>
>> Anyway, this won't help you, cause the only effect is a change of the
>> sign to +19 and therefore worse.
>>
>> Can you show me the topology, you made?
>> As far as the free energy experts here discussed it, sc-power=2
>> shouldn't be the lambda-dependance of choice. I can agree to that from my
>> simulations.
>>
>> regards
>>
>> Maik Goette, Dipl. Biol.
>> Max Planck Institute for Biophysical Chemistry
>> Theoretical & computational biophysics department
>> Am Fassberg 11
>> 37077 Goettingen
>> Germany
>> Tel.  : ++49 551 201 2310
>> Fax   : ++49 551 201 2302
>> Email : mgoette[at]mpi-bpc.mpg.de
>> mgoette2[at]gwdg.de WWW   : http://www.mpibpc.gwdg.de/groups/grubmueller/
>>
>>
>>
>> cilpa at users.csc.fi wrote:
>>> Hi maik
>>>
>>>
>>> the energy DEltaG_hyd=deltaG_mutation(vacuo)-deltaGmutation(water)
>>>
>>>
>>> for fisrt one deltaG_mutation(vacuo) =-2,85 for second
>>> deltaGmutation(water) =15,87
>>>
>>> so i found DEltaG_hyd = -19 kj/mol instead of -3,1 kj/mol
>>>
>>> i'll do a series of calculation with sc-power=2 and see the difference
>>> see if there's some significant difference. geraldine
>>>
>>>
>>> On Thu, February 14, 2008 10:57 am, Maik Goette wrote:
>>>
>>>> Hi
>>>>
>>>>
>>>>
>>>> Now, make sure, that you apply the thermodynamic cycle correctly.
>>>> I wouldn't wonder about large errors. Depending on the definition of
>>>> large. Error estimation from free energy calculations can be very
>>>> tricky. Using mpi shouldn't influence your calculations outcome.
>>>> Still I don't know the error.
>>>> How to post the numbers: Exp. value (+error) and your calculated value
>>>>  (+error) ? From those I could tell you if something is terribly
>>>> wrong or not. :)
>>>>
>>>> Regards
>>>>
>>>>
>>>>
>>>> Maik Goette, Dipl. Biol.
>>>> Max Planck Institute for Biophysical Chemistry
>>>> Theoretical & computational biophysics department
>>>> Am Fassberg 11
>>>> 37077 Goettingen
>>>> Germany
>>>> Tel.  : ++49 551 201 2310
>>>> Fax   : ++49 551 201 2302
>>>> Email : mgoette[at]mpi-bpc.mpg.de
>>>> mgoette2[at]gwdg.de WWW   :
>>>> http://www.mpibpc.gwdg.de/groups/grubmueller/
>>>>
>>>>
>>>>
>>>>
>>>> cilpa at users.csc.fi wrote:
>>>>> Hi MaiK,
>>>>> Sorry i forgot the subject. Thank you to have answered this email.
>>>>> This
>>>>> tutorial is about solvation free energy of toluene using
>>>>> thermodynamical cycles. two energies have to be calculated according
>>>>> to this cycle, the hydration energy of toluene and the mutation
>>>>> energy from dummy to toluene in vacuo. For each energy calculation
>>>>> several MD simulation for different lambda value must be performed.
>>>>> in the mdp files given i added the line sc-power=1 and made the
>>>>> calculation. from the file dgdl.xvg obtained for each lambda value i
>>>>> used the g_analyse command to get the average and the estimated
>>>>> error of free energy. i made the plot <dgdl>=f(lambda) for the first
>>>>> part (hydration of toluene) integrated and get the integrated value
>>>>> of the plot and did the same for mutation energy of toluene and made
>>>>> the difference of the two integrated values. I really don't know
>>>>> where the error came from : -version of gromacs, the fact that mpi
>>>>> has been used for the first part and single process for the mutation
>>>>> part (because of shake block problem) won't change anything i guess,
>>>>> I would carefully go through one
>>>>> more time and check why i didn't get the right value. I don't know
>>>>> if this high estimated   error is really problematic. regards
>>>>> geraldine
>>>>>
>>>>> On Wed, February 13, 2008 6:19 pm, Maik Goette wrote:
>>>>>
>>>>>
>>>>>> Hi
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> First of all: You can be quite happy, if anyone reads a "no
>>>>>> subject" mail. sc-power seems to be set to zero in the newer
>>>>>> version, in contrast to the old ones, where it was 1 by default.
>>>>>> Therefore GROMACS
>>>>>> complains. 1 is the value, you should use. I don't know the
>>>>>> tutorial, but if you use exactly the files given there (and add
>>>>>> sc-power = 1), you should get values, which are close to the
>>>>>> experimental one (I guess, cause a tutorial won't make sense, if
>>>>>> the files don't yield the desired value).
>>>>>>
>>>>>> Sorry, but more help is hard to give with so few information.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Regards
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Maik Goette, Dipl. Biol.
>>>>>> Max Planck Institute for Biophysical Chemistry
>>>>>> Theoretical & computational biophysics department
>>>>>> Am Fassberg 11
>>>>>> 37077 Goettingen
>>>>>> Germany
>>>>>> Tel.  : ++49 551 201 2310
>>>>>> Fax   : ++49 551 201 2302
>>>>>> Email : mgoette[at]mpi-bpc.mpg.de
>>>>>> mgoette2[at]gwdg.de WWW   :
>>>>>> http://www.mpibpc.gwdg.de/groups/grubmueller/
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> cilpa at users.csc.fi wrote:
>>>>>>> Hi all,
>>>>>>> i 'm a beginner in molecular dynamics. I was doing the tutorial
>>>>>>> of md group, hydration free energy of toluene :
>>>>>>> http://md.chem.rug.nl/education/Free-Energy_Course/2.hydration-
>>>>>>> fe.h tml
>>>>>>>
>>>>>>>
>>>>>>> i have some problems regarding the calculation of toluene in
>>>>>>> water. the thing is that in the examples files equ-10.00.mdp and
>>>>>>>  data-10.00.mdp no sc-power were mentioned and doing the
>>>>>>> calculation the error files mentioned about sc-power must be >0
>>>>>>> so i added this line sc-power=1 to those files. Analyzing the
>>>>>>> dgdl.xvg files the estimate error(err. est.) is about 5.79....
>>>>>>> for each lambda value studied while the mutation of toluene to
>>>>>>> dummy in vacuo gave really reasonable rms of about 10E-5.
>>>>>>> Combining the integration value of
>>>>>>> the plot in water and vacuo did not give me the right solvation
>>>>>>> free energy value (compare to experimental one). i assumed that
>>>>>>> this is the free calculation in water which is problematic. I
>>>>>>> tried to play with sc-power and sc-alpha but still even if those
>>>>>>> parameters seem to influence a lot on the energy no real
>>>>>>> improvement are visible. Would you have any idea what could be
>>>>>>> the problem? geraldine Helsinki university
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>>>>> http://www.gromacs.org/mailman/listinfo/gmx-users
>>>>>>> Please search the archive at http://www.gromacs.org/search
>>>>>>> before posting! Please don't post (un)subscribe requests to the
>>>>>>> list. Use the www interface or send it to
>>>>>>> gmx-users-request at gromacs.org. Can't post? Read
>>>>>>> http://www.gromacs.org/mailing_lists/users.php
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> .
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>> _______________________________________________
>>>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>>>> http://www.gromacs.org/mailman/listinfo/gmx-users
>>>>>> Please search the archive at http://www.gromacs.org/search before
>>>>>> posting! Please don't post (un)subscribe requests to the list. Use
>>>>>> the www interface or send it to gmx-users-request at gromacs.org.
>>>>>> Can't
>>>>>> post? Read http://www.gromacs.org/mailing_lists/users.php
>>>>>>
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>>> http://www.gromacs.org/mailman/listinfo/gmx-users
>>>>> Please search the archive at http://www.gromacs.org/search before
>>>>> posting! Please don't post (un)subscribe requests to the list. Use
>>>>> the www interface or send it to gmx-users-request at gromacs.org. Can't
>>>>> post? Read http://www.gromacs.org/mailing_lists/users.php
>>>>>
>>>>>
>>>>>
>>>>> .
>>>>>
>>>>>
>>>>>
>>>> _______________________________________________
>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>> http://www.gromacs.org/mailman/listinfo/gmx-users
>>>> Please search the archive at http://www.gromacs.org/search before
>>>> posting! Please don't post (un)subscribe requests to the list. Use the
>>>>  www interface or send it to gmx-users-request at gromacs.org. Can't
>>>> post? Read http://www.gromacs.org/mailing_lists/users.php
>>>>
>>>>
>>>>
>>>
>>> _______________________________________________
>>> gmx-users mailing list    gmx-users at gromacs.org
>>> http://www.gromacs.org/mailman/listinfo/gmx-users
>>> Please search the archive at http://www.gromacs.org/search before
>>> posting! Please don't post (un)subscribe requests to the list. Use the
>>> www interface or send it to gmx-users-request at gromacs.org. Can't post?
>>> Read http://www.gromacs.org/mailing_lists/users.php
>>>
>>>
>>> .
>>>
>>>
>> _______________________________________________
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://www.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at http://www.gromacs.org/search before posting!
>>  Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org. Can't post? Read
>> http://www.gromacs.org/mailing_lists/users.php
>>
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://www.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at http://www.gromacs.org/search before posting!
>> Please don't post (un)subscribe requests to the list. Use the 
>> www interface or send it to gmx-users-request at gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php

More information about the gromacs.org_gmx-users mailing list