Content-Type: multipart/alternative; boundary="----=_Part_752_31450296.1368181863770" ------=_Part_752_31450296.1368181863770 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Dear Professor Baker, Thanks for your reply. I will try these suggestions, and will post here the= =20 results if I succeed. I actually want to compare this binding energy=20 between a set of systems with different mutations and conformations, and=20 was trying to find the best settings in one example first. So, at the end I= =20 will be computing relative energies. One more question, for these=20 comparisons, all the calculations should be done exactly on the same grids,= =20 not only grid spacings, but also dimensions and centers, right? Thanks, Rosana On Friday, May 10, 2013 12:27:45 AM UTC+2, Baker, Nathan wrote: > > Hi Rosana =E2=80=93 > > =20 > > This is a warning message intended to encourage you to explore the=20 > sensitivity of your results to the calculation parameters. In your case,= I=20 > would recommend varying the =E2=80=9Cpdime=E2=80=9D (increase it) and =E2= =80=9Cofrac=E2=80=9D (increase it)=20 > parameters to see how much they influence your results. > > =20 > > However, I would also encourage you to clarify the problem you=E2=80=99re= solving=20 > with APBS. Absolute electrostatic binding energies aren=E2=80=99t usuall= y very=20 > meaningful =E2=80=93 quite often you=E2=80=99re more interested in relati= ve values of=20 > binding energies as differences calculated when testing multiple mutation= s,=20 > solution conditions, titration states, etc. It is likely that these=20 > relative energies are more stable with respect to grid setup (within=20 > reason). > > =20 > > Thanks, > > =20 > > -- > > Nathan Baker > > Pacific Northwest National Laboratory > > +1-509-375-3997 > > http://nabaker.me=20 > > =20 > > *From:* Rosana Collepardo [mailto:r.coll...@gmail.com ]=20 > *Sent:* Thursday, May 09, 2013 2:56 AM > *To:* apbs-...@googlegroups.com > *Subject:* [apbs-users] Convergence problems in very large calculation > > =20 > > Hello, > > =20 > > I'm trying to calculate the electrostatic energy between two nucleosomes= =20 > in a dinucleosome complex using apbs, and I'm having problems converging= =20 > the calculation against grid size. > > =20 > > The calculation requires a lot of memory, but using fine grid spacings of= =20 > 0.5 and 0.35 A requires less than the ~28GB apbs limit. > > =20 > > For those grid spacings the resulting internucleosome electrostatic energ= y=20 > (dinucleosome - nucleosome1 - nucleosome2) is not well converged and give= s=20 > very large values. For instance, it changes from -341 kJ/mol for 0.5A to= =20 > -563 kJ/mol for 0.35 A. (No evident errors or warning messages are=20 > reported by apbs) > > =20 > > The problem is that if I further decrease the grid spacing, I need to=20 > split the calculation as explained in the apbs website. I used=20 > the asynchronous parallel procedure and prize.py to calculate the=20 > dimensions of the sub grids, and then summed the energies for each=20 > calculation. A new problem is that when I split the calculation I start= =20 > getting warning messages for the subgrid calculations (Unusually large=20 > potential values detected on the focusing boundary! Convergence not=20 > guaranteed for NPBE/NRPBE calculations!). > > =20 > > Furthermore, the resulting energy (direct sum of the energies given by=20 > each of the subgrid calculations) seems to diverge now, it is -1740 kJ/mo= l=20 > for a grid spacing of 0.2A and to -3697 kJ/mol for a grid spacing of 0.15= A.=20 > I also tried changing ofrac, but without any noticeable improvement.=20 > > =20 > > I wanted to know if anyone has experience with very large systems and if= =20 > you could give me a few hints as to what I could do in order to converge= =20 > this calculation. > > =20 > > With thanks and best wishes, > > Rosana > ------=_Part_752_31450296.1368181863770 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable Dear Professor Baker,

Thanks for your reply. I will try = these suggestions, and will post here the results if I succeed. I actually = want to compare this binding energy between a set of systems with different= mutations and conformations, and was trying to find the best settings in o= ne example first. So, at the end I will be computing relative energies. One= more question, for these comparisons, all the calculations should be done = exactly on the same grids, not only grid spacings, but also dimensions and = centers, right?

Thanks,
Rosana

On Friday, May 10, 2013 12:27:45 AM UTC+2, Baker, Nathan wrote:

Hi= Rosana =E2=80=93

This = is a warning message intended to encourage you to explore the sensitivity o= f your results to the calculation parameters.  In your case, I would r= ecommend varying the =E2=80=9Cpdime=E2=80=9D (increase it) and =E2=80=9Cofr= ac=E2=80=9D (increase it) parameters to see how much they influence your re= sults.

However, I would= also encourage you to clarify the problem you=E2=80=99re solving with APBS= .  Absolute electrostatic binding energies aren=E2=80=99t usually very= meaningful =E2=80=93 quite often you=E2=80=99re more interested in relativ= e values of binding energies as differences calculated when testing multipl= e mutations, solution conditions, titration states, etc.  It is likely= that these relative energies are more stable with respect to grid setup (w= ithin reason).

=

Thanks,<= /span>

--

Nathan Baker

Pacific Northwest National Labora= tory

+1-509-375= -3997

<= p class=3D"MsoNormal">

From: Rosana Collepard= o [mailto:r.coll...@gmail.com]
Sent: Thursday, = May 09, 2013 2:56 AM
<= b>Subject: [apbs-users] Convergence problems in very large calculation<= /span>

Hello,

= I'm tr= ying to calculate the electrostatic energy between two nucleosomes in a din= ucleosome complex using apbs, and I'm having problems converging the calcul= ation against grid size.

The calculation requires a lot of memory, = but using fine grid spacings of 0.5 and 0.35 A requires less than the = ~28GB apbs limit.

=

For those grid spacings the resulting internucl= eosome electrostatic energy (dinucleosome - nucleosome1 - nucleosome2) is n= ot well converged and gives very large values. For instance, it changes fro= m -341 kJ/mol for 0.5A to -563 kJ/mol for 0.35 A.  (No evident er= rors or warning messages are reported by apbs)

The problem is that = if I further decrease the grid spacing, I need to split the calculation as = explained in the apbs website. I used the asynchronous parallel proced= ure and prize.py to calculate the dimensions of the sub grids, and the= n summed the energies for each calculation. A new problem is that when I sp= lit the calculation I start getting warning messages for the subgrid calcul= ations (Unusually large potential values detected on the focusing boun= dary! Convergence not guaranteed for NPBE/NRPBE calculations!).=

Furthermore, the resulting energy (direct sum of the energies given by ea= ch of the subgrid calculations) seems to diverge now, it is -1740 kJ/mol fo= r a grid spacing of 0.2A and to -3697 kJ/mol for a grid spacing of 0.15A. I= also tried changing ofrac, but without any noticeable improvement.

I wanted to know if anyone has experience with very large syste= ms and if you could give me a few hints as to what I could do in order to c= onverge this calculation.

With thanks and best wishes,

Rosana

------=_Part_752_31450296.1368181863770--