From owner-chemistry@ccl.net Wed Jul 11 07:44:00 2018 From: "Jorge Iulek iulek*|*uepg.br" To: CCL Subject: CCL: get electrostatic potential at atoms and molecular polarizability Message-Id: <-53389-180710212255-31452-BD4hHRCAZ8uGtGeniPptCA-$-server.ccl.net> X-Original-From: Jorge Iulek Content-Language: en-US Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=utf-8; format=flowed Date: Tue, 10 Jul 2018 22:22:39 -0300 MIME-Version: 1.0 Sent to CCL by: Jorge Iulek [iulek]=[uepg.br] Thanks Dr. Mazur and Dr. Shobe, I am right now testing the question of the B3LYP variants (from gamess manual I think I understand how to do it), the question of cartesian vs spherical I will study a bit more now and then try to find how to do it in gamess. I will let you know of the results. Sincerely yours, Jorge On 07/10/2018 03:45 PM, Grzegorz Mazur mazur###chemia.uj.edu.pl wrote: > > Sent to CCL by: Grzegorz Mazur [mazur:_:chemia.uj.edu.pl] > Hi, > > Sorry, my first answer got mangled somehow. Two obvious suspects are > Cartesian vs spherical basis set, as already suggested by David, and > different variants of B3LYP (VWN3 vs VWN5 based). > > Grzesiek > > On 10.07.2018 14:38, Jorge Iulek iulek]~[uepg.br wrote: >> >> Sent to CCL by: Jorge Iulek [iulek(~)uepg.br] >> Dear all, >> >> >>      I have some molecules structurally similar to ones that were >> used in a previous (not by me) QSAR study. To these, a model >> (equation to estimate biological activity) was proposed (very good >> correlation) to explain (reasonably well) activities based on QM >> descriptors. Therefore I want to check what the biological activity >> is estimated for my similar molecules. I need to calculate the >> electrostatic potential (EP) at two specific atoms and the molecule >> polarizability, to make the prediction. I should say that the model >> was established using other QM program, nevertheless, but I must go >> for the free academic program gamess. The article states that the >> molecules were structure optimized (3D gemometry) using  DFT theory, >> the 6-31G(d) basis set at the hybrid functional B3LYP. >> >>      My gamess input for optimization contains: >> ------------------------ >>   $contrl dfttyp=B3LYP runtyp=optimize nzvar=72 coord=zmt icharg=0 $end >>   $system timlim=525600.0 MWORDS=6 $end >>   $statpt hess=guess nstep=100 nprt=-1 npun=0 $end >>   $basis  GBASIS=N31 NGAUSS=6 NDFUNC=1 $end >>   $guess  guess=huckel $end >>   $SCF DIRSCF=.true. CONV=1d-7 FDIFF=.FALSE. $end >>   $DFT DFTTYP=B3LYP $end >> ... here comes coordinates (as zmatrix), etc... >> --------------------------- >> >>      And then from the resulting .dat file I extract both $DATA and >> $VEC lines to input with lines below to calculate EP... >> ---------------- >>   $contrl runtyp=prop dfttyp=B3LYP nzvar=0 coord=unique icharg=0 $end >>   $system timlim=525600.0 MWORDS=6 $end >>   $basis  GBASIS=N31 NGAUSS=6 NDFUNC=1 $end >>   $guess guess=moread $end >>   $DFT DFTTYP=B3LYP $end >>   $ELPOT IEPOT=1 WHERE=NUCLEI OUTPUT=BOTH $end >>   $FFCALC OFFDIA=.TRUE. $end >> ... $DATA and $VEC ... >> ------- >> >>      And molecular polarizability with... >> -------------- >>   $contrl runtyp=energy local=boys $end >>   $local  poldcm=.true. polnum=.true.   fcore=.false.  $end >>   $contrl dfttyp=B3LYP nzvar=0 coord=unique icharg=0 $end >>   $SYSTEM MWORDS=6 $END >>   $BASIS  GBASIS=N31 NGAUSS=6 NDFUNC=1 $END >>   $GUESS  GUESS=moread $END >>   $SCF DIRSCF=.true. CONV=1d-7 FDIFF=.FALSE. $end >>   $DFT DFTTYP=B3LYP $end >> ... $DATA and $VEC ... >> ------------ >> >>      The matter is that with these instructions, even for the >> original molecules (of the QSAR study), I cannot reproduce author's >> results. In fact, the QSAR model (equation) seems to be quite >> sensitive to the resulting EP (their coefficients in the equation are >> large, so minor differences have large impact). Therefore, possibly I >> miss something here at my input files for gamess. I would much thank >> if someone would criticize my inputs and point to some error. Should >> I expect, in any case, to get very approximate results, once the >> calculation implies the same theory, basis set, functional, etc. (I >> guess), at using different QM programs? >> >>      Yours, >> >> >> Jorge>      >          Conferences: >> http://server.ccl.net/chemistry/announcements/conferences/>     >     >     http://www.ccl.net/chemistry/sub_unsub.shtmlConferences: > http://server.ccl.net/chemistry/announcements/conferences/>     > > > From owner-chemistry@ccl.net Wed Jul 11 09:56:00 2018 From: "Visvaldas K. coyote_v2002.^.yahoo.com" To: CCL Subject: CCL: get electrostatic potential at atoms and molecular polarizability Message-Id: <-53390-180711032806-18237-IOI/R7i+6XsUv1C5c7RWSg^server.ccl.net> X-Original-From: "Visvaldas K." Content-Type: multipart/alternative; boundary="----=_Part_2092094_510730175.1531294078531" Date: Wed, 11 Jul 2018 07:27:58 +0000 (UTC) MIME-Version: 1.0 Sent to CCL by: "Visvaldas K." [coyote_v2002[a]yahoo.com] ------=_Part_2092094_510730175.1531294078531 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Please check the units first, the units GAMESS use could be different from= other programs. You probably did that, but there is no harm in rechecking.= Also it should be noted that GAMESS' B3LYP is not equivalent to B3LYP used= in other programs (the equations in original DFT paper's presumably can be= ambiguously interpreted); there is another version B3LYPV1R (called B3LYP1= in earlier GAMESS versions), which is supposed to match *other program's* = versions. On Wednesday, July 11, 2018, 5:19:44 AM GMT+2, Alan Shusterman alan^_^r= eed.edu wrote: =20 =20 Jorge,You might pass along the article, or the procedure part of the artic= le, that you are trying to reproduce. I'm not at all familiar with GAMESS i= nput so I can't help you there, but if I had to guess at what your problem = is I would say, look at the location where the EP is calculated. EP values = change rapidly from extremely (and for most interpretive purposes, uselessl= y) positive near the atomic nucleus to zero at points far from the molecule= . I will bet that you are interested in a location where the value is relat= ively modest (+/- 100-400 kJ/mol), but there can be considerable variation = in these values too so it is necessary to follow the recipe used by the dev= elopers of the QSAR. Good luck,Alan On Tue, Jul 10, 2018 at 5:38 AM, Jorge Iulek iulek]~[uepg.br wrote: Sent to CCL by: Jorge Iulek [iulek(~)uepg.br] Dear all, =C2=A0=C2=A0=C2=A0 I have some molecules structurally similar to ones that = were used in a previous (not by me) QSAR study. To these, a model (equation= to estimate biological activity) was proposed (very good correlation) to e= xplain (reasonably well) activities based on QM descriptors. Therefore I wa= nt to check what the biological activity is estimated for my similar molecu= les. I need to calculate the electrostatic potential (EP) at two specific a= toms and the molecule polarizability, to make the prediction. I should say = that the model was established using other QM program, nevertheless, but I = must go for the free academic program gamess. The article states that the m= olecules were structure optimized (3D gemometry) using=C2=A0 DFT theory, th= e 6-31G(d) basis set at the hybrid functional B3LYP. =C2=A0=C2=A0=C2=A0 My gamess input for optimization contains: ------------------------ =C2=A0$contrl dfttyp=3DB3LYP runtyp=3Doptimize nzvar=3D72 coord=3Dzmt ichar= g=3D0 $end =C2=A0$system timlim=3D525600.0 MWORDS=3D6 $end =C2=A0$statpt hess=3Dguess nstep=3D100 nprt=3D-1 npun=3D0 $end =C2=A0$basis=C2=A0 GBASIS=3DN31 NGAUSS=3D6 NDFUNC=3D1 $end =C2=A0$guess=C2=A0 guess=3Dhuckel $end =C2=A0$SCF DIRSCF=3D.true. CONV=3D1d-7 FDIFF=3D.FALSE. $end =C2=A0$DFT DFTTYP=3DB3LYP $end ... here comes coordinates (as zmatrix), etc... --------------------------- =C2=A0=C2=A0=C2=A0 And then from the resulting .dat file I extract both $DA= TA and $VEC lines to input with lines below to calculate EP... ---------------- =C2=A0$contrl runtyp=3Dprop dfttyp=3DB3LYP nzvar=3D0 coord=3Dunique icharg= =3D0 $end =C2=A0$system timlim=3D525600.0 MWORDS=3D6 $end =C2=A0$basis=C2=A0 GBASIS=3DN31 NGAUSS=3D6 NDFUNC=3D1 $end =C2=A0$guess guess=3Dmoread $end =C2=A0$DFT DFTTYP=3DB3LYP $end =C2=A0$ELPOT IEPOT=3D1 WHERE=3DNUCLEI OUTPUT=3DBOTH $end =C2=A0$FFCALC OFFDIA=3D.TRUE. $end ... $DATA and $VEC ... ------- =C2=A0=C2=A0=C2=A0 And molecular polarizability with... -------------- =C2=A0$contrl runtyp=3Denergy local=3Dboys $end =C2=A0$local=C2=A0 poldcm=3D.true. polnum=3D.true.=C2=A0 =C2=A0fcore=3D.fal= se.=C2=A0 $end =C2=A0$contrl dfttyp=3DB3LYP nzvar=3D0 coord=3Dunique icharg=3D0 $end =C2=A0$SYSTEM MWORDS=3D6 $END =C2=A0$BASIS=C2=A0 GBASIS=3DN31 NGAUSS=3D6 NDFUNC=3D1 $END =C2=A0$GUESS=C2=A0 GUESS=3Dmoread $END =C2=A0$SCF DIRSCF=3D.true. CONV=3D1d-7 FDIFF=3D.FALSE. $end =C2=A0$DFT DFTTYP=3DB3LYP $end ... $DATA and $VEC ... ------------ =C2=A0=C2=A0=C2=A0 The matter is that with these instructions, even for the= original molecules (of the QSAR study), I cannot reproduce author's result= s. In fact, the QSAR model (equation) seems to be quite sensitive to the re= sulting EP (their coefficients in the equation are large, so minor differen= ces have large impact). Therefore, possibly I miss something here at my inp= ut files for gamess. I would much thank if someone would criticize my input= s and point to some error. Should I expect, in any case, to get very approx= imate results, once the calculation implies the same theory, basis set, fun= ctional, etc. (I guess), at using different QM programs? =C2=A0=C2=A0=C2=A0 Yours, Jorge -=3D This is automatically added to each message by the mailing script =3D- E-mail to subscribers: CHEMISTRY.(_).ccl.net or use: =C2=A0 =C2=A0 =C2=A0http://www.ccl.net/cgi-bin/cc l/send_ccl_message E-mail to administrators: CHEMISTRY-REQUEST.(_).ccl.net or use =C2=A0 =C2=A0 =C2=A0http://www.ccl.net/cgi-bin/cc l/send_ccl_message Subscribe/Unsubscribe:=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/chemistry/s u= b_unsub.shtmlConferences: http://server.ccl.net/chemistr y/= announcements/conferences/ Search Messages: http://www.ccl.net/chemistry/s earchccl/index.shtml =C2=A0 =C2=A0 =C2=A0http://www.ccl.net/spammers.t xt RTFI: http://www.ccl.net/chemistry/a boutccl/instructions/ --=20 Alan Shusterman Chemistry Department Reed College 3203 SE Woodstock Blvd Portland, OR 97202-8199 503-517-7699 http://blogs.reed.edu/alan/ "Patience, persistence, and a sense of humor." Dave Barrett =20 ------=_Part_2092094_510730175.1531294078531 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Please check the units first, the units GAMESS use could be di= fferent from other programs. You probably did that, but there is no harm in= rechecking. Also it should be noted that GAMESS' B3LYP is not equivalent t= o B3LYP used in other programs (the equations in original DFT paper's presu= mably can be ambiguously interpreted); there is another version B3LYPV1R (c= alled B3LYP1 in earlier GAMESS versions), which is supposed to match *other= program's* versions.

=20
=20
On Wednesday, July 11, 2018, 5:19:44 AM GMT+2, Alan Shu= sterman alan^_^reed.edu <owner-chemistry(_)ccl.net> wrote:


=
Jorge,
You might pass along the article, or the procedure pa= rt of the article, that you are trying to reproduce. I'm not at all familia= r with GAMESS input so I can't help you there, but if I had to guess at wha= t your problem is I would say, look at the location where the EP is calc= ulated. EP values change rapidly from extremely (and for most interpret= ive purposes, uselessly) positive near the atomic nucleus to zero at points= far from the molecule. I will bet that you are interested in a location wh= ere the value is relatively modest (+/- 100-400 kJ/mol), but there can be c= onsiderable variation in these values too so it is necessary to follow the = recipe used by the developers of the QSAR. Good luck,
Alan

On Tue, Jul 10, 2018 at 5:38 AM, J= orge Iulek iulek]~[uepg.br <owner-chemistry.(_).ccl.net&= gt; wrote:
uepg.br]
Dear all,


    I have some molecules structurally similar to ones that = were used in a previous (not by me) QSAR study. To these, a model (equation= to estimate biological activity) was proposed (very good correlation) to e= xplain (reasonably well) activities based on QM descriptors. Therefore I wa= nt to check what the biological activity is estimated for my similar molecu= les. I need to calculate the electrostatic potential (EP) at two specific a= toms and the molecule polarizability, to make the prediction. I should say = that the model was established using other QM program, nevertheless, but I = must go for the free academic program gamess. The article states that the m= olecules were structure optimized (3D gemometry) using  DFT theory, th= e 6-31G(d) basis set at the hybrid functional B3LYP.

    My gamess input for optimization contains:
------------------------
 $contrl dfttyp=3DB3LYP runtyp=3Doptimize nzvar=3D72 coord=3Dzmt ichar= g=3D0 $end
 $system timlim=3D525600.0 MWORDS=3D6 $end
 $statpt hess=3Dguess nstep=3D100 nprt=3D-1 npun=3D0 $end
 $basis  GBASIS=3DN31 NGAUSS=3D6 NDFUNC=3D1 $end
 $guess  guess=3Dhuckel $end
 $SCF DIRSCF=3D.true. CONV=3D1d-7 FDIFF=3D.FALSE. $end
 $DFT DFTTYP=3DB3LYP $end
... here comes coordinates (as zmatrix), etc...
---------------------------

    And then from the resulting .dat file I extract both $DA= TA and $VEC lines to input with lines below to calculate EP...
----------------
 $contrl runtyp=3Dprop dfttyp=3DB3LYP nzvar=3D0 coord=3Dunique icharg= =3D0 $end
 $system timlim=3D525600.0 MWORDS=3D6 $end
 $basis  GBASIS=3DN31 NGAUSS=3D6 NDFUNC=3D1 $end
 $guess guess=3Dmoread $end
 $DFT DFTTYP=3DB3LYP $end
 $ELPOT IEPOT=3D1 WHERE=3DNUCLEI OUTPUT=3DBOTH $end
 $FFCALC OFFDIA=3D.TRUE. $end
... $DATA and $VEC ...
-------

    And molecular polarizability with...
--------------
 $contrl runtyp=3Denergy local=3Dboys $end
 $local  poldcm=3D.true. polnum=3D.true.   fcore=3D.fal= se.  $end
 $contrl dfttyp=3DB3LYP nzvar=3D0 coord=3Dunique icharg=3D0 $end
 $SYSTEM MWORDS=3D6 $END
 $BASIS  GBASIS=3DN31 NGAUSS=3D6 NDFUNC=3D1 $END
 $GUESS  GUESS=3Dmoread $END
 $SCF DIRSCF=3D.true. CONV=3D1d-7 FDIFF=3D.FALSE. $end
 $DFT DFTTYP=3DB3LYP $end
... $DATA and $VEC ...
------------

    The matter is that with these instructions, even for the= original molecules (of the QSAR study), I cannot reproduce author's result= s. In fact, the QSAR model (equation) seems to be quite sensitive to the re= sulting EP (their coefficients in the equation are large, so minor differen= ces have large impact). Therefore, possibly I miss something here at my inp= ut files for gamess. I would much thank if someone would criticize my input= s and point to some error. Should I expect, in any case, to get very approx= imate results, once the calculation implies the same theory, basis set, fun= ctional, etc. (I guess), at using different QM programs?

    Yours,


Jorge



-=3D This is automatically added to each message by the mailing script =3D-=

E-mail to subscribers: CHEMISTRY.(_).ccl.net or use:
     http://www.ccl.net/cgi-bin/cc l/sen= d_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST.(_).ccl.net or use
     http://www.ccl.net/cgi-bin/cc l/sen= d_ccl_message

Subscribe/Unsubscribe:      http://www.ccl.= net/chemistry/s ub_unsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs Conferences: ht= tp://server.ccl.net/chemistr y/announcements/conferences/

Search Messages: http://www.ccl.net/chemistry/s earc= hccl/index.shtml

     http://www.ccl.net/spammers.t xt

RTFI: http://www.ccl.net/chemistry/a boutccl/inst= ructions/





--
Alan= Shusterman
Chemistry Department
Reed College
3203 SE Woodstock Bl= vd
Portland, OR 97202-8199
503-517-7699
http://blogs.reed.edu/alan= /
"Patience, persistence, and a sense of humor." Dave Barrett
------=_Part_2092094_510730175.1531294078531-- From owner-chemistry@ccl.net Wed Jul 11 10:37:01 2018 From: "=?iso-8859-1?Q?V=EDctor_Lua=F1a?= Cabal victor^fluor.quimica.uniovi.es" To: CCL Subject: CCL:G: get electrostatic potential at atoms and molecular polarizability Message-Id: <-53391-180711055625-27263-uxLo9BYL0a9Amrfiao5mQA]|[server.ccl.net> X-Original-From: =?iso-8859-1?Q?V=EDctor_Lua=F1a?= Cabal Content-Disposition: inline Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=utf-8 Date: Wed, 11 Jul 2018 11:56:17 +0200 MIME-Version: 1.0 Sent to CCL by: =?iso-8859-1?Q?V=EDctor_Lua=F1a?= Cabal [victor:-:fluor.quimica.uniovi.es] * Alan Shusterman alan^_^reed.edu [2018-07-10 10:11:32 -0700]: > Jorge, > You might pass along the article, or the procedure part of the article, > that you are trying to reproduce. I'm not at all familiar with GAMESS input > so I can't help you there, but if I had to guess at what your problem is I > would say, look at *the location where the EP is calculated*. EP values > change rapidly from extremely (and for most interpretive purposes, > uselessly) positive near the atomic nucleus to zero at points far from the > molecule. I will bet that you are interested in a location where the value > is relatively modest (+/- 100-400 kJ/mol), but there can be considerable > variation in these values too so it is necessary to follow the recipe used > by the developers of the QSAR. Good luck, Jorge, Let me add my 0.02€ to the exchange. Many published works on the EP employ rrough approximations, particularly those that try to approach the EP from "atomic charges" rather than integrating the real equation. Most that approaches have no relation with the real EP. However, I don't think your approach suffers from that. it is my impressiom that your problems may come from small details like a cartesian or spherical reperesentation of spherical harmonics or different implementations of the B3LYP functional. Both have already been exposed ... As a general rule the results from two ab initio codes (gamess, gaussian, nwchem, columbus, dalton, pyscf, ...) should be equivalent. Gaussian, as a commercial house, tends to avoid that their cpu times can be compared with those of other codes, but that refers to computational efficiency, not physical properties. There are so many comparisons between gaussian, gamess, ... that any meaningful difference can be discarded, imho. Small details, however may do a real difference. If you accept a suggerence I would forget any approximation on the EP and use a real integration of the Poison equation or similar method: the formal integration of the EP, in any way. Good luck, Víctor Luaña From owner-chemistry@ccl.net Wed Jul 11 21:16:00 2018 From: "Jorge Iulek iulek++uepg.br" To: CCL Subject: CCL:G: get electrostatic potential at atoms and molecular polarizability Message-Id: <-53392-180711210153-3525-ftJKOUwlFLBFyDKJ/RP3bg++server.ccl.net> X-Original-From: Jorge Iulek Content-Language: en-US Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=utf-8; format=flowed Date: Wed, 11 Jul 2018 22:01:39 -0300 MIME-Version: 1.0 Sent to CCL by: Jorge Iulek [iulek[a]uepg.br] Dear all,     This discussion list is really great. Many experts taking their time to deliver their warnings, to solve my problem. I add to my acknowledgments Dr. Cabal, Visvaldas, Shusterman and Huebner.     I excuse myself I cannot verify and test all suggestions so soon (and some of them I must study further), but I will take all of them in consideration and let you know of the results.     Concerning the warning below, I might say that the article does not have a supplementary material session (I read all text and hope to have gathered all details possible), but that the program used was Gaussian98. Authors say "EP at atoms ..." which I suppose is the EP at the exact location of the nucleus, which I suppose would be given by the gamess instruction  "$ELPOT IEPOT=1 WHERE=NUCLEI OUTPUT=BOTH $end". But your comment prompts me to double check what are the atom coordinates I input for EP calculation, I exported it form the .dat file ($DATA group) coming out of the optimization step (as well as the $VEC group).     Yours, Jorge On 07/11/2018 06:56 AM, Víctor Luaña Cabal victor^fluor.quimica.uniovi.es wrote: > Sent to CCL by: =?iso-8859-1?Q?V=EDctor_Lua=F1a?= Cabal [victor:-:fluor.quimica.uniovi.es] > * Alan Shusterman alan^_^reed.edu [2018-07-10 10:11:32 -0700]: >> Jorge, >> You might pass along the article, or the procedure part of the article, >> that you are trying to reproduce. I'm not at all familiar with GAMESS input >> so I can't help you there, but if I had to guess at what your problem is I >> would say, look at *the location where the EP is calculated*. EP values >> change rapidly from extremely (and for most interpretive purposes, >> uselessly) positive near the atomic nucleus to zero at points far from the >> molecule. I will bet that you are interested in a location where the value >> is relatively modest (+/- 100-400 kJ/mol), but there can be considerable >> variation in these values too so it is necessary to follow the recipe used >> by the developers of the QSAR. Good luck, > Jorge, > > Let me add my 0.02€ to the exchange. Many published works on the EP > employ rrough approximations, particularly those that try to approach the > EP from "atomic charges" rather than integrating the real equation. > Most that approaches have no relation with the real EP. However, I > don't think your approach suffers from that. > > it is my impressiom that your problems may come from small details like > a cartesian or spherical reperesentation of spherical harmonics or > different implementations of the B3LYP functional. Both have already > been exposed ... > > As a general rule the results from two ab initio codes (gamess, gaussian, > nwchem, columbus, dalton, pyscf, ...) should be equivalent. Gaussian, > as a commercial house, tends to avoid that their cpu times can be compared > with those of other codes, but that refers to computational efficiency, > not physical properties. There are so many comparisons between gaussian, > gamess, ... that any meaningful difference can be discarded, imho. Small > details, however may do a real difference. > > If you accept a suggerence I would forget any approximation on the EP and > use a real integration of the Poison equation or similar method: the > formal integration of the EP, in any way. > > Good luck, > Víctor Luaña> > > From owner-chemistry@ccl.net Wed Jul 11 22:33:00 2018 From: "David Shobe shobedavid+*+gmail.com" To: CCL Subject: CCL:G: get electrostatic potential at atoms and molecular polarizability Message-Id: <-53393-180711223206-1786-uUisTt9BVjBRr6nQmwBD1g]-[server.ccl.net> X-Original-From: David Shobe Content-Type: multipart/alternative; boundary="000000000000fa41060570c42a5a" Date: Wed, 11 Jul 2018 21:31:50 -0500 MIME-Version: 1.0 Sent to CCL by: David Shobe [shobedavid##gmail.com] --000000000000fa41060570c42a5a Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Jorge-- I seriously doubt that the EP (I assume you mean electrostatic potential) is to be measured at the nucleus. It's likely to be infinity insofar as the nucleus is modeled as a point charge. That being said, I have no idea where the actual EP is measured. --David On Wed, Jul 11, 2018, 9:21 PM Jorge Iulek iulek++uepg.br < owner-chemistry-,-ccl.net> wrote: > > Sent to CCL by: Jorge Iulek [iulek[a]uepg.br] > Dear all, > > > This discussion list is really great. Many experts taking their > time to deliver their warnings, to solve my problem. I add to my > acknowledgments Dr. Cabal, Visvaldas, Shusterman and Huebner. > > I excuse myself I cannot verify and test all suggestions so soon > (and some of them I must study further), but I will take all of them in > consideration and let you know of the results. > > Concerning the warning below, I might say that the article does not > have a supplementary material session (I read all text and hope to have > gathered all details possible), but that the program used was > Gaussian98. Authors say "EP at atoms ..." which I suppose is the EP at > the exact location of the nucleus, which I suppose would be given by the > gamess instruction "$ELPOT IEPOT=3D1 WHERE=3DNUCLEI OUTPUT=3DBOTH $end".= But > your comment prompts me to double check what are the atom coordinates I > input for EP calculation, I exported it form the .dat file ($DATA group) > coming out of the optimization step (as well as the $VEC group). > > Yours, > > > Jorge > > > On 07/11/2018 06:56 AM, V=C3=ADctor Lua=C3=B1a Cabal > victor^fluor.quimica.uniovi.es wrote: > > Sent to CCL by: =3D?iso-8859-1?Q?V=3DEDctor_Lua=3DF1a?=3D Cabal [victor= :-: > fluor.quimica.uniovi.es] > > * Alan Shusterman alan^_^reed.edu > [2018-07-10 10:11:32 -0700]: > >> Jorge, > >> You might pass along the article, or the procedure part of the article= , > >> that you are trying to reproduce. I'm not at all familiar with GAMESS > input > >> so I can't help you there, but if I had to guess at what your problem > is I > >> would say, look at *the location where the EP is calculated*. EP value= s > >> change rapidly from extremely (and for most interpretive purposes, > >> uselessly) positive near the atomic nucleus to zero at points far from > the > >> molecule. I will bet that you are interested in a location where the > value > >> is relatively modest (+/- 100-400 kJ/mol), but there can be considerab= le > >> variation in these values too so it is necessary to follow the recipe > used > >> by the developers of the QSAR. Good luck, > > Jorge, > > > > Let me add my 0.02=E2=82=AC to the exchange. Many published works on th= e EP > > employ rrough approximations, particularly those that try to approach t= he > > EP from "atomic charges" rather than integrating the real equation. > > Most that approaches have no relation with the real EP. However, I > > don't think your approach suffers from that. > > > > it is my impressiom that your problems may come from small details like > > a cartesian or spherical reperesentation of spherical harmonics or > > different implementations of the B3LYP functional. Both have already > > been exposed ... > > > > As a general rule the results from two ab initio codes (gamess, gaussia= n, > > nwchem, columbus, dalton, pyscf, ...) should be equivalent. Gaussian, > > as a commercial house, tends to avoid that their cpu times can be > compared > > with those of other codes, but that refers to computational efficiency, > > not physical properties. There are so many comparisons between gaussian= , > > gamess, ... that any meaningful difference can be discarded, imho. Sma= ll > > details, however may do a real difference. > > > > If you accept a suggerence I would forget any approximation on the EP a= nd > > use a real integration of the Poison equation or similar method: the > > formal integration of the EP, in any way. > > > > Good luck, > > V=C3=ADctor Lua=C3=B1a> > > > > > > > > -=3D This is automatically added to each message by the mailing script = =3D-> > > --000000000000fa41060570c42a5a Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
Jorge--

I seriously doubt that the EP (I assume you mean electrostatic poten= tial) is to be measured at the nucleus. It's likely to be infinity inso= far as the nucleus is modeled as a point charge.
That being said, I have no idea where the actual E= P is measured.

--David
On Wed, Jul = 11, 2018, 9:21 PM Jorge Iulek iulek++uepg.br= <owner-chemistry-,-ccl.net= > wrote:

Sent to CCL by: Jorge Iulek [iulek[a]uepg.br]
Dear all,


=C2=A0=C2=A0=C2=A0=C2=A0 This discussion list is really great. Many experts= taking their
time to deliver their warnings, to solve my problem. I add to my
acknowledgments Dr. Cabal, Visvaldas, Shusterman and Huebner.

=C2=A0=C2=A0=C2=A0=C2=A0 I excuse myself I cannot verify and test all sugge= stions so soon
(and some of them I must study further), but I will take all of them in consideration and let you know of the results.

=C2=A0=C2=A0=C2=A0=C2=A0 Concerning the warning below, I might say that the= article does not
have a supplementary material session (I read all text and hope to have gathered all details possible), but that the program used was
Gaussian98. Authors say "EP at atoms ..." which I suppose is the = EP at
the exact location of the nucleus, which I suppose would be given by the gamess instruction=C2=A0 "$ELPOT IEPOT=3D1 WHERE=3DNUCLEI OUTPUT=3DBOT= H $end". But
your comment prompts me to double check what are the atom coordinates I input for EP calculation, I exported it form the .dat file ($DATA group) coming out of the optimization step (as well as the $VEC group).

=C2=A0=C2=A0=C2=A0=C2=A0 Yours,


Jorge


On 07/11/2018 06:56 AM, V=C3=ADctor Lua=C3=B1a Cabal
victor^fluor.quimica.uniovi.es wrote:
> Sent to CCL by: =3D?iso-8859-1?Q?V=3DEDctor_Lua=3DF1a?=3D Cabal [victo= r:-:fluor.quimica.uniovi.es]
> * Alan Shusterman alan^_^reed.edu <owner-chemistry/a\ccl.net= > [2018-07-10 10:11:32 -0700]:
>> Jorge,
>> You might pass along the article, or the procedure part of the art= icle,
>> that you are trying to reproduce. I'm not at all familiar with= GAMESS input
>> so I can't help you there, but if I had to guess at what your = problem is I
>> would say, look at *the location where the EP is calculated*. EP v= alues
>> change rapidly from extremely (and for most interpretive purposes,=
>> uselessly) positive near the atomic nucleus to zero at points far = > from the
>> molecule. I will bet that you are interested in a location where t= he value
>> is relatively modest (+/- 100-400 kJ/mol), but there can be consid= erable
>> variation in these values too so it is necessary to follow the rec= ipe used
>> by the developers of the QSAR. Good luck,
> Jorge,
>
> Let me add my 0.02=E2=82=AC to the exchange. Many published works on t= he EP
> employ rrough approximations, particularly those that try to approach = the
> EP from "atomic charges" rather than integrating the real eq= uation.
> Most that approaches have no relation with the real EP. However, I
> don't think your approach suffers from that.
>
> it is my impressiom that your problems may come from small details lik= e
> a cartesian or spherical reperesentation of spherical harmonics or
> different implementations of the B3LYP functional. Both have already > been exposed ...
>
> As a general rule the results from two ab initio codes (gamess, gaussi= an,
> nwchem, columbus, dalton, pyscf, ...) should be equivalent. Gaussian,<= br> > as a commercial house, tends to avoid that their cpu times can be comp= ared
> with those of other codes, but that refers to computational efficiency= ,
> not physical properties. There are so many comparisons between gaussia= n,
> gamess, ...=C2=A0 that any meaningful difference can be discarded, imh= o. Small
> details, however may do a real difference.
>
> If you accept a suggerence I would forget any approximation on the EP = and
> use a real integration of the Poison equation or similar method: the > formal integration of the EP, in any way.
>
> Good luck,
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0V=C3=ADctor Lua=C3=B1a>
>
>



-=3D This is automatically added to each message by the mailing script =3D-=
E-mail to subscribers: CHEMISTRY-,-ccl.net or use:
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/cg= i-bin/ccl/send_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST-,-ccl.net or use
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/cg= i-bin/ccl/send_ccl_message
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/chemi= stry/sub_unsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs
Conferences: http://server.ccl.= net/chemistry/announcements/conferences/

Search Messages: http://www.ccl.net/che= mistry/searchccl/index.shtml
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/spammers.txt
RTFI: http://www.ccl.net/chemistry/a= boutccl/instructions/


--000000000000fa41060570c42a5a--