From owner-chemistry@ccl.net Wed Aug 22 13:42:00 2018
From: "Jorge Iulek iulek---uepg.br" <owner-chemistry _ server.ccl.net>
To: CCL
Subject: CCL:G: get electrostatic potential at atoms and molecular polarizability
Message-Id: <-53440-180822100246-15948-fYEJvhoEopEz9+fqEjLz7A _ server.ccl.net>
X-Original-From: Jorge Iulek <iulek ~ uepg.br>
Content-Language: en-US
Content-Type: multipart/alternative;
 boundary="------------231173384B0937C788161F9B"
Date: Wed, 22 Aug 2018 11:02:24 -0300
MIME-Version: 1.0


Sent to CCL by: Jorge Iulek [iulek(-)uepg.br]
This is a multi-part message in MIME format.
--------------231173384B0937C788161F9B
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit

Dear all,


     It is been a long time I posted my question, sorry I could not 
provide a comprehensive feedback earlier.

     First, I thank all that have responded. In summary, my question 
refers to my disability to reproduce calculated electrostatic potential 
and polarizability values I found (better saying, the predict biological 
activity values they were used to predict, at a QSAR model) in the 
literature, primarily using gamess.

      I will comment upon suggestions I received and results. I advance 
that, no, I cannot reproduce results in any case, still.

a)   "here is the issue of 5d (spherical) v. 6d (Cartesian)"

"Two obvious suspects are Cartesian vs spherical basis set, as already 
suggested"

"

it is my impressiom that your problems may come from small details like
a cartesian or spherical reperesentation of spherical harmonics ..

"

I tried all possibilities within gamess, card ISPHERE (-1,0,+1). -1 is 
the default, that corresponds to cartesian basis. 0 gives the same 
results, but I get different results if I use +1.Nevertheless, no result 
reproduction.

b) "find out what the original researchers used".

" pass along the article, or the procedure part of the article"

"it is necessary to follow the recipe used by the developers of the QSAR"

I read thoroughly the paper, there is no supplementary material and I 
even wrote the authors, no answer.

c) _"the location where the EP is calculated"
_

The article states : "Electrostatic Potential (EP) at each atom" and "EP 
at atom numbers", which, to my understanding corresponds to gamess 
"ELPOT IEPOT=1 WHERE=NUCLEI"

As a matter of fact, the QSAR equation is very sensitive (high 
coefficients) to the  calculated EP values.

"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."

With the input above, gamess states "..., If there is a nucleus at the 
evaluation point, that nucleus is ignored, avoiding a singularity...", 
therefore my understanding that it would do as the article states....

d) "different variants of B3LYP (VWN3 vs VWN5 based)"

I learned that "dfttyp=B3LYP" is the same of "dfttyp=B3LYPV5". But 
curiously, If I input "dfttyp=B3LYPV3", I get the same results.

"it is my impressiom that your problems may come from small details like
... or different implementations of the B3LYP functional. Both have already
been exposed ..."

"noted that GAMESS' B3LYP is not equivalent to B3LYP used in
       other programs"

And I see there is a good number of possibilities.

"there is another version B3LYPV1R (called B3LYP1 in earlier GAMESS 
versions), which is supposed to match *other program's* versions"

I also tried this , but results did not match, again.

e) " the units GAMESS use could be different from other programs"

yes, I checked and, in principle, units are corrected. if I had used 
other units, values should be much more different.

f) " 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."

     I should learn how to do this in gamess. Nevertheless, there is no 
clue that such procedure was adopted in the original article.

g) "

In gaussian, spherical harmonics is the default. You must tell the code:
use 5d 7f 9g if you pretend using QTAIM programs to analyze your
results, for instance. Otherwise the code uses 6d, 9f, ...
"

     I could not find how to do this in gamess, in fact, there just a 
brief mention on 5d, but concerning the "Sapporo basis set family"

   I should also mention that I entered the lowest energy conformer 
(estimated with rd_kit); but even if I use other conformer, minimum 
impact is observed in the calculated values (within my calculations), 
so, in principle, I do not think it is a matter of conformer used.

     At the several possible combinations, I wrote authors, yet four 
weeks ago, no answer. Therefore, I do not think I can get into the 
details how those values were calculated.

     Should there be any other ideas, I would be happy to try. In any 
case, thank you very much for suggestions already stated.


Jorge


--------------231173384B0937C788161F9B
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: 8bit

<html>
  <head>
    <meta http-equiv="content-type" content="text/html; charset=utf-8">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <p>Dear all,</p>
    <p><br>
    </p>
    <p>    It is been a long time I posted my question, sorry I could
      not provide a comprehensive feedback earlier.</p>
    <p>    First, I thank all that have responded. In summary, my
      question refers to my disability to reproduce calculated
      electrostatic potential and polarizability values I found (better
      saying, the predict biological activity values they were used to
      predict, at a QSAR model) in the literature, primarily using
      gamess.</p>
    <p>     I will comment upon suggestions I received and results. I
      advance that, no, I cannot reproduce results in any case, still.<br>
    </p>
    <p>a)   "here is the issue of 5d (spherical) v. 6d (Cartesian)"</p>
    <p>"Two obvious suspects are Cartesian vs spherical basis set, as
      already suggested"</p>
    <p>"</p>
    <pre wrap="">it is my impressiom that your problems may come from small details like
a cartesian or spherical reperesentation of spherical harmonics ..</pre>
    <p>"<br>
    </p>
    <p>I tried all possibilities within gamess, card ISPHERE (-1,0,+1).
      -1 is the default, that corresponds to cartesian basis. 0 gives
      the same results, but I get different results if I use
      +1.Nevertheless, no result reproduction. <br>
    </p>
    <p>b) "find out what the original researchers used". <br>
    </p>
    <p>" pass along the article, or the procedure part of the article"</p>
    <p>"it is necessary to follow the recipe used by the developers of
      the QSAR"</p>
    <p>I read thoroughly the paper, there is no supplementary material
      and I even wrote the authors, no answer.<br>
    </p>
    <p>c) <u>"the location where the EP is calculated"<br>
      </u></p>
    <p>The article states : "Electrostatic Potential (EP) at each atom" 
      and "EP at atom numbers", which, to my understanding corresponds
      to gamess "ELPOT IEPOT=1 WHERE=NUCLEI"</p>
    <p>As a matter of fact, the QSAR equation is very sensitive (high
      coefficients) to the  calculated EP values.</p>
    <p>"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.</p>
    <div dir="auto">That being said, I have no idea where the actual EP
      is measured."<br>
      <br>
      With the input above, gamess states "..., If there is a nucleus at
      the evaluation point, that nucleus is ignored, avoiding a
      singularity...", therefore my understanding that it would do as
      the article states....<br>
      <br>
    </div>
    <p> </p>
    <p>d) "different variants of B3LYP (VWN3 vs VWN5 based)"</p>
    <p>I learned that "dfttyp=B3LYP" is the same of "dfttyp=B3LYPV5".
      But curiously, If I input "dfttyp=B3LYPV3", I get the same
      results.</p>
    <pre wrap="">"it is my impressiom that your problems may come from small details like
... or different implementations of the B3LYP functional. Both have already
been exposed ..."

"noted that GAMESS' B3LYP is not equivalent to B3LYP used in
      other programs"

</pre>
    <p>And I see there is a good number of possibilities.</p>
    <p>"there is another version B3LYPV1R (called B3LYP1 in earlier
      GAMESS versions), which is supposed to match *other program's*
      versions" <br>
    </p>
    <p>I also tried this , but results did not match, again.</p>
    <p> </p>
    <p>e) " the units GAMESS use could be different from other programs"</p>
    <p>yes, I checked and, in principle, units are corrected. if I had
      used other units, values should be much more different.</p>
    f) " Small details, however may do a real difference....¨
    <pre wrap="">"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."</pre>
    <p>    I should learn how to do this in gamess. Nevertheless, there
      is no clue that such procedure was adopted in the original
      article.  <br>
    </p>
    <p>g) "<br>
    </p>
    <pre wrap="">In gaussian, spherical harmonics is the default. You must tell the code:
use 5d 7f 9g if you pretend using QTAIM programs to analyze your
results, for instance. Otherwise the code uses 6d, 9f, ...
"
<font size="+1">
</font></pre>
    <p>     I could not find how to do this in gamess, in fact, there
      just a brief mention on 5d, but concerning the "Sapporo basis set
      family"</p>
    <pre wrap="">
</pre>
    <p>  I should also mention that I entered the lowest energy
      conformer (estimated with rd_kit); but even if I use other
      conformer, minimum impact is observed in the calculated values
      (within my calculations), so, in principle, I do not think it is a
      matter of conformer used.<br>
    </p>
    <p>    At the several possible combinations, I wrote authors, yet
      four weeks ago, no answer. Therefore, I do not think I can get
      into the details how those values were calculated.</p>
    <p>    Should there be any other ideas, I would be happy to try. In
      any case, thank you very much for suggestions already stated.</p>
    <p><br>
    </p>
    <p>Jorge<br>
    </p>
  </body>
</html>

--------------231173384B0937C788161F9B--


From owner-chemistry@ccl.net Wed Aug 22 14:17:00 2018
From: "Jorge Iulek iulek###uepg.br" <owner-chemistry~~server.ccl.net>
To: CCL
Subject: CCL:G: get electrostatic potential at atoms and molecular polarizability
Message-Id: <-53441-180822100247-15959-hEB99NJTILi2NMW/jy0j9Q~~server.ccl.net>
X-Original-From: Jorge Iulek <iulek/./uepg.br>
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset=utf-8; format=flowed
Date: Wed, 22 Aug 2018 11:01:58 -0300
MIME-Version: 1.0


Sent to CCL by: Jorge Iulek [iulek^_^uepg.br]
Dear Victor,


     Besides your help on the computational chemistry issue, thank you 
(muchas gracias) for explaining me further name usages and some history 
facts!


Jorge


On 07/12/2018 05:07 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]
> * Jorge Iulek iulek++uepg.br <owner-chemistry*|*ccl.net> [2018-07-11 22:01:39 -0300]:
>> 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.
> Jorge,
>
> Being from Brasil probablly you know already, but let me use your excuse
> to let others ñnow which is the spaniards tradition relative toour
> names:
>
> I am Víctor Luaña Cabal, that means both Luaña and Cabal are my family
> names and Víctor my surname. I am from northern spain, a place where
> people love treating each other on quite plain terms: similarly to the
> swedish is their crime literature describes rightly their uses.
>
> If someone I know calls me Dr. Luaña I expect trouble. I always teach my
> students on the very first day: please, call me Víctor. If someone
> calls me Dr. Cabal that is an almost sure sign it comes from a predatory
> journal or meeting. I tend to hate such kind of "invitations". I guess
> we all have received many lately. I even answered positively to one of
> such calls out of curiosity, so I know real details on how it
> works this modern curse.
>
> So, coming back to the spaniard names: Dr. Luaña would be a correct
> albeit quite formal treatment. Dr. Luaña Cabal would be too. Víctor is
> most right for me, and any other formula is extrange or directly wrong.
>
> Many spaniards use a hyphen to join their family names: Luaña-Cabal,
> to remove trouble from foreigners, but my Luaña is quite unusual (even
> in spain) so I decided to keep only my first family name.
>
> The road that goes from Miguel Portillo padre (Sr) to Michel Porteeloh
> jr. (I tried to imitate the british accent) is a relevant one. Let me
> remember that the spanish civil war was the first step in WWII and it
> splitted spaniards into two halfs those loyal to the constitutional
> republic and the militars taking control by the arms: the francoist band
> supported by Hitler and Mussolini.
>
> Don't worry, this is not a post on history. Extraordinary british
> professors (hispahistas) have written excellent books you may read if
> the subject interests you. However, huge ammounts of scientists were
> forced to a migration to save their lives. Mexico, among other
> latinamerican countries received most, but quite a few discples of Ramón
> y Cajal, for instance, found their place at Universities in Canada, USA
> or UK. Rhey were most bright peopel, after all.
>
> On the science content I agree with you. most CCL mails tend to "why
> gaussian makes my calculation fail?" and most times the answer would be
> read carefully the manual, please, or use your tuition fees wisely.
>
> Your question is not like that and refers to something that truly worries
> many people: computational research *must* be reproducible. In other
> ways it would not qualify for real science. As a general rule it uses to
> be but, in the practice, a non trivial calculation may hide details that
> are important and the pressure on journals to reduce papers size means
> that sometimes details can be eliminated and not always included on the
> additional material that most serious journals accept.
>
> I guess many people you mention is intrigued to known finally which is the
> cause of your trouble, but all the information exchange is also related
> to the fact that summertime is coming and teaching pressure is mostly
> over on the northern hemisphere.
>
>>      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
> In gaussian, spherical harmonics is the default. You must tell the code:
> use 5d 7f 9g if you pretend using QTAIM programs to analyze your
> results, for instance. Otherwise the code uses 6d, 9f, ...
>
> Best luck,
>            Víctor Luaña
> --
>      .  .     One should not treat others in ways that one would not like
>     / `' \    to be treated
>    /(o)(o)\   (Golden rule of ethics)
>   /`. \/ .'\
> /   '`'`   \ "Freedom!, freedom!, freedom! After that put whatever
> |  \'`'`/  | term you like"
> |  |'`'`|  |  --Largo Caballero (socialist, spanish president in the exile,
>   \/`'`'`'\/   shortly before dying)
> ==(((==)))===================================+ A person is slave of his words
> ! Dr.Víctor Luaña, in silico chemist & prof. ! and owner of his silences.
> ! Departamento de Química Física y Analítica !
> ! Universidad de Oviedo, 33006-Oviedo, Spain ! "Never let your sense of morals
> ! e-mail:  <victor*|*fluor.quimica.uniovi.es>  ! prevent you from doing what is
> ! phone: +34-984080927    fax: +34-985103125 ! right."
> +--------------------------------------------+ -- Salvor Hardin, "Foundation"
>   GroupPage: <http://azufre.quimica.uniovi.es/>
>   Articles:  <http://scholar.google.com/citations?user=Ibl1BWAAAAAJ&hl=es>
>   git-hub:   <https://github.com/aoterodelaroza>
>   ORCID: 0000-0003-4585-4627; RID: H-2045-2015>
>
>


From owner-chemistry@ccl.net Wed Aug 22 15:06:01 2018
From: "Gilson, Michael mgilson::ucsd.edu" <owner-chemistry/a\server.ccl.net>
To: CCL
Subject: CCL: Pre-announcement of D3R Grand Challenge 4, a new blinded prediction challenge for protein-ligand poses and affinities
Message-Id: <-53442-180822145637-21326-UeiMiEo8EMuEkxIrnAwoQA/a\server.ccl.net>
X-Original-From: "Gilson, Michael" <mgilson^_^ucsd.edu>
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset="us-ascii"
Date: Wed, 22 Aug 2018 18:56:26 +0000
MIME-Version: 1.0


Sent to CCL by: "Gilson, Michael" [mgilson===ucsd.edu]
Dear Colleagues,

We are pleased to announce that Grand Challenge 4 (GC4) will open on
September 4, with the following submission deadlines:
    Stage 1a, cross-docking challenge: October 4
    Stage 1b, self-docking challenge: October 19
    Stage 2. affinity ranking and free energies: December 4

Challenge components will include:
    Affinity ranking of ~450 Cathepsin S inhibitors from the same large
dataset drawn from in GC3
    Affinity ranking of ~150 beta secretase 1 (BACE) inhibitors
    Pose prediction of 20 BACE inhibitors
    Free energy prediction challenges suitable for alchemical free
energy methods, for both Cathepsin and BACE

Details will be posted soon at drugdesigndata.org.

Many thanks to Janssen and Novartis for the Cathepsin and BACE data,
respectively.

Happy docking!
D3R Team