From chemistry-request@server.ccl.net Sun Jul 28 03:28:11 2002 Received: from web12801.mail.yahoo.com ([216.136.174.36]) by server.ccl.net (8.11.6/8.11.0) with SMTP id g6S7SAr20737 for ; Sun, 28 Jul 2002 03:28:10 -0400 Message-ID: <20020728072810.73708.qmail@web12801.mail.yahoo.com> Received: from [203.87.150.242] by web12801.mail.yahoo.com via HTTP; Sun, 28 Jul 2002 00:28:10 PDT Date: Sun, 28 Jul 2002 00:28:10 -0700 (PDT) From: amor san juan Subject: Re: CCL:partial atomic charges for autodock3 To: Matthew Lee Cc: chemistry@ccl.net In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Matthew: It is correct that 3D structure of ligand is needed for QM calculations. To my understanding, Autodock is based on Kollman united atom model which merges non-polar H's(ADT do this) for simplified calculation. Regarding optimization, I have the same dilemma. I used either MM & semiempirical that results to conformation with smaller yz coordinates & negative smaller charges.Take a look to my experiments: 1) Control distributed file ben.pdbq when run by steps mkgpf3===> mkdpf3===>autogrid3===>autodock3 shows data as same as control.However, 2) ben.mol2 created from comm. molecular modeling package used for same steps above showed no grid maps for ACN atoms. Kindly post to CCL the summary of your question's reply(if any). This will give me a hint and you could help me a lot through this. Thanks in advance. ************************** --- Matthew Lee wrote: > The Autodock documentation recommends calculating > the ligand charges > using AMPAC or MOPAC, but to only include *polar* > hydrogens. Isn't it > incorrect to run QM calculations on molecules that > have incomplete > valencies, i.e. a carbon without its hydrogen(s)? > If one were to add all > the hydrogens to a small molecule and do either a > single point calculation > or a geometry optimization for the partial atomic > charges, and then remove > the non-polar hydrogens, the molecule would then be > far too negative. > > How are people using AMPAC/MOPAC in conjunction > with Autodock? Are > you leaving all the non-polar hydrogens there, and > designating an additional > 'h' atom type to be included for the non-polar > hydrogens in the .gpf? Or is > everyone using the rule-based Gasteiger-Marsili > charges? > > Also, I am wondering about the hydrogens on the > macromolecule. From > the documentation, it seems that all hydrogrens, > even the non-polar ones, > present in a protein's .pdbqs file will be > considered H-bond donors (as long > as the atom name starts with H). I deduce this from > 1) the statement that > "When modeling hydrogen bonds explicitly, it is > necessary to add polar > hydrogens to the macromolecule.", which implies to > only add polar hydrogens > to the protein and 2) the fact that the .gpf file > seems to only allow for C, > N, O, S, H, X & M atom types in the macromolecule, > without a separate type > 'h' for non-polar hydrogens. It seems that the > available atom types for the > macromolecule are frozen, in contrast to those for > the ligands, which can be > anything the user wants and can be as many as the > user wants. > Interestingly, the documentation says under the DPF > section of the Appendix: > > ------------------------ > types > Atom names for all atom types present in ligand. > Each must be a single > character, and only one of: C, N, O, S, H, X, or M. > ------------------------ > This seems erroneous, as I tried using an alternate > atom name like P, or > even more vaguely like J, and autodock3 has no > problem as long as 1) an > affinity map was created for that atom name by > autogrid3, 2) the atom name > appears in the "types" line of the .dpf, and 3) the > ATOM_MAPS in > "autocomm.h" is sufficiently large. > > In the AMBER all atom force field, hydrogen bonds > are treated > implicitly by the electrostatics, with no explicit > 12 10 LJ parameters. Are > there any thoughts in the community on doing the > following for setting up > the macromolecule: > 1)including all hydrogens in the macromolecule > 2)using the AMBER partial atomic charges for all > protein atoms > 3)turning off all H-bond acceptors in the ligand by > switching the ligand's > explicit H-bond paramaters for "O-H" and "S-H " in > the .gpf from "12 10" to > "12 6". > > Much thanks in advance for any feedback. > > --Matt Lee > > ******************************************************************* > Matthew Randolph Lee, Ph.D. > > Computational Scientist phone: (858)410-6534 > Lion Bioscience Inc. fax: (858)410-6665 > 9880 Campus Point Drive e-mail: > matthew.lee@lionbioscience.com > San Diego, CA 92121 > ******************************************************************* > > Sincerely, Amor San Juan University of the Philippines-Diliman __________________________________________________ Do You Yahoo!? Yahoo! Health - Feel better, live better http://health.yahoo.com From chemistry-request@server.ccl.net Sun Jul 28 12:57:42 2002 Received: from smtp1.ndsu.nodak.edu ([134.129.111.146]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g6SGvfr32299 for ; Sun, 28 Jul 2002 12:57:41 -0400 Received: from pharm200 (dyn251.sud-118.ndsu.NoDak.edu [134.129.118.251]) by smtp1.ndsu.nodak.edu (8.12.1/8.12.1) with SMTP id g6SGvcHZ025591; Sun, 28 Jul 2002 11:57:42 -0500 Message-ID: <010901c23657$12ffbdc0$fb768186@ndsu.nodak.edu> From: "xin Hu" To: "Matthew Lee" , References: Subject: Re: CCL:partial atomic charges for autodock3 Date: Sun, 28 Jul 2002 11:51:48 -0500 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0106_01C2362D.276A47A0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 This is a multi-part message in MIME format. ------=_NextPart_000_0106_01C2362D.276A47A0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable partial atomic charges for autodock3I had some experience before, = hopefully something I still remember is helpful. 1. AutoDock3 uses polar-H for protein indeed, like the ligand. The = polar-H are added using "Protonate" module of AutoDock. If you uses = Sybyl, you should select " united atom H". Check the pdbqs file, If = there are non-polar H's, you should re-prepare the file. 2. You can modify the source code(the head file of AutoGrid&AutoDock " = autocomm.h") to add any atome typy for the protein, such as "h" for the = non-polarH. (or P if it is really needed for you in protein.). While all = H included in the protein, surely you need the all-atom partial charge = instead of kollman united atom charges.=20 3. It seems you can simply modify the "gpfgen.awk" or "dpfgen.awk" to = control the 12-6 and 12-10 parameters. 4. I am not familar with MOPAC, AutoDock3 recommand the Gesteger charge = models for ligand (from Sybyl or Babel assigned it), and using Autotor = to merger the non-polarH. If you prefer to QM charges, I think it should = be the same, get the charges with all H, and then apply AUtotors. good luck Xin *************************************** Xin Hu Department of Pharmaceutical Sciences College of Pharmacy North Dakota State University Fargo, ND 58105 Tel: 701-231-4307 (home) 701-231-7159 (Office) Fax: 701-231-8333 E-mail: xin.hu@ndsu.nodak.edu xxh0541@hotmail.com *************************************** ----- Original Message -----=20 From: Matthew Lee=20 To: 'chemistry@ccl.net'=20 Sent: Friday, July 26, 2002 6:21 PM Subject: CCL:partial atomic charges for autodock3 The Autodock documentation recommends calculating the ligand = charges using AMPAC or MOPAC, but to only include *polar* hydrogens. = Isn't it incorrect to run QM calculations on molecules that have = incomplete valencies, i.e. a carbon without its hydrogen(s)? If one = were to add all the hydrogens to a small molecule and do either a single = point calculation or a geometry optimization for the partial atomic = charges, and then remove the non-polar hydrogens, the molecule would = then be far too negative. =20 How are people using AMPAC/MOPAC in conjunction with Autodock? = Are you leaving all the non-polar hydrogens there, and designating an = additional 'h' atom type to be included for the non-polar hydrogens in = the .gpf? Or is everyone using the rule-based Gasteiger-Marsili charges? Also, I am wondering about the hydrogens on the macromolecule. = From the documentation, it seems that all hydrogrens, even the = non-polar ones, present in a protein's .pdbqs file will be considered = H-bond donors (as long as the atom name starts with H). I deduce this = > from 1) the statement that "When modeling hydrogen bonds explicitly, it = is necessary to add polar hydrogens to the macromolecule.", which = implies to only add polar hydrogens to the protein and 2) the fact that = the .gpf file seems to only allow for C, N, O, S, H, X & M atom types in = the macromolecule, without a separate type 'h' for non-polar hydrogens. = It seems that the available atom types for the macromolecule are frozen, = in contrast to those for the ligands, which can be anything the user = wants and can be as many as the user wants. Interestingly, the = documentation says under the DPF section of the Appendix: =20 ------------------------=20 types =20 Atom names for all atom types present in ligand. Each must be a = single character, and only one of: C, N, O, S, H, X, or M. ------------------------=20 This seems erroneous, as I tried using an alternate atom name like P, = or even more vaguely like J, and autodock3 has no problem as long as 1) = an affinity map was created for that atom name by autogrid3, 2) the atom = name appears in the "types" line of the .dpf, and 3) the ATOM_MAPS in = "autocomm.h" is sufficiently large. In the AMBER all atom force field, hydrogen bonds are treated = implicitly by the electrostatics, with no explicit 12 10 LJ parameters. = Are there any thoughts in the community on doing the following for = setting up the macromolecule: 1)including all hydrogens in the macromolecule=20 2)using the AMBER partial atomic charges for all protein atoms=20 3)turning off all H-bond acceptors in the ligand by switching the = ligand's explicit H-bond paramaters for "O-H" and "S-H " in the .gpf = > from "12 10" to "12 6". Much thanks in advance for any feedback.=20 --Matt Lee=20 *******************************************************************=20 Matthew Randolph Lee, Ph.D.=20 Computational Scientist phone: (858)410-6534=20 Lion Bioscience Inc. fax: (858)410-6665=20 9880 Campus Point Drive e-mail: matthew.lee@lionbioscience.com=20 San Diego, CA 92121=20 *******************************************************************=20 ------=_NextPart_000_0106_01C2362D.276A47A0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable partial atomic charges for autodock3
I had some experience before, hopefully = something I=20 still remember is helpful.
 
1.  AutoDock3 uses polar-H for = protein indeed,=20 like the ligand. The polar-H are added using "Protonate" module of = AutoDock. If=20 you uses Sybyl, you should select " united atom H". Check the pdbqs = file, =20 If there are non-polar H's, you should re-prepare the file.
2. You can modify the source code(the = head file of=20 AutoGrid&AutoDock " autocomm.h") to add any atome typy for the = protein, such=20 as "h" for the non-polarH. (or P if it is really needed for = you in=20 protein.). While all H included in the protein, surely you need the = all-atom=20 partial charge instead of kollman united atom charges.
3. It seems you can simply modify the = "gpfgen.awk"=20 or "dpfgen.awk" to control the 12-6 and 12-10 parameters.
4. I am not familar with MOPAC, = AutoDock3 recommand=20 the Gesteger charge models for ligand (from Sybyl or Babel assigned it), = and=20 using Autotor to merger the non-polarH. If you prefer to QM charges, I = think it=20 should be the same, get the charges with all H, and then apply=20 AUtotors.
 
good luck
 
Xin
***************************************
Xin Hu
Department of=20 Pharmaceutical Sciences
College of Pharmacy
North Dakota State=20 University
Fargo, ND 58105
Tel: 701-231-4307=20 (home)
       701-231-7159 = (Office)
Fax:=20 701-231-8333
E-mail: xin.hu@ndsu.nodak.edu
 =            =20 xxh0541@hotmail.com
**********= *****************************
----- Original Message -----
From:=20 Matthew Lee
Sent: Friday, July 26, 2002 = 6:21 PM
Subject: CCL:partial atomic = charges for=20 autodock3

        The = Autodock=20 documentation recommends calculating the ligand charges using AMPAC or = MOPAC,=20 but to only include *polar* hydrogens.  Isn't it incorrect to run = QM=20 calculations on molecules that have incomplete valencies, i.e. a = carbon=20 without its hydrogen(s)?  If one were to add all the hydrogens to = a small=20 molecule and do either a single point calculation or a geometry = optimization=20 for the partial atomic charges, and then remove the non-polar = hydrogens, the=20 molecule would then be far too negative. 

        How are = people=20 using AMPAC/MOPAC in conjunction with Autodock?  Are you leaving = all the=20 non-polar hydrogens there, and designating an additional 'h' atom type = to be=20 included for the non-polar hydrogens in the .gpf? Or is everyone using = the=20 rule-based Gasteiger-Marsili charges?

        Also, I = am=20 wondering about the hydrogens on the macromolecule.  From the=20 documentation, it seems that all hydrogrens, even the non-polar ones, = present=20 in a protein's .pdbqs file will be considered H-bond donors (as long = as the=20 atom name starts with H).  I deduce this from 1) the statement = that "When=20 modeling hydrogen bonds explicitly, it is necessary to add polar = hydrogens to=20 the macromolecule.", which implies to only add polar hydrogens to the = protein=20 and 2) the fact that the .gpf file seems to only allow for C, N, O, S, = H, X=20 & M atom types in the macromolecule, without a separate type 'h' = for=20 non-polar hydrogens.  It seems that the available atom types for = the=20 macromolecule are frozen, in contrast to those for the ligands, which = can be=20 anything the user wants and can be as many as the user wants. =20 Interestingly, the documentation says under the DPF section of the=20 Appendix: 

------------------------
types=20 <string>
Atom names for all atom types = present=20 in ligand.  Each must be a single character, and only one of: C, = N, O, S,=20 H, X, or M.

------------------------
This seems=20 erroneous, as I tried using an alternate atom name like P, or even = more=20 vaguely like J, and autodock3 has no problem as long as 1) an affinity = map was=20 created for that atom name by autogrid3, 2) the atom name appears in = the=20 "types" line of the .dpf, and 3) the ATOM_MAPS in "autocomm.h" is = sufficiently=20 large.

        In the = AMBER all=20 atom force field, hydrogen bonds are treated implicitly by the = electrostatics,=20 with no explicit 12 10 LJ parameters.  Are there any thoughts in = the=20 community on doing the following for setting up the = macromolecule:

1)including all hydrogens in the macromolecule=20
2)using the AMBER partial atomic charges for = all=20 protein atoms
3)turning off all H-bond = acceptors in=20 the ligand by switching the ligand's explicit H-bond paramaters for = "O-H" and=20 "S-H " in the .gpf from "12 10" to "12 6".

        Much = thanks in=20 advance for any feedback.

--Matt Lee

****************************************************************= ***=20
Matthew Randolph Lee, Ph.D.

Computational Scientist=20         phone:  = (858)410-6534=20
Lion Bioscience Inc.   =20         fax:   =20 (858)410-6665
9880 Campus Point Drive=20       e-mail: =20 matthew.lee@lionbioscience.com
San Diego, = CA =20 92121
****************************************************************= ***=20

------=_NextPart_000_0106_01C2362D.276A47A0-- From chemistry-request@server.ccl.net Sun Jul 28 15:13:05 2002 Received: from travelers.mail.cornell.edu ([132.236.56.13]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g6SJD5r01164 for ; Sun, 28 Jul 2002 15:13:05 -0400 Received: from travelers.mail.cornell.edu (travelers.mail.cornell.edu [132.236.56.13]) by travelers.mail.cornell.edu (8.9.3/8.9.3) with SMTP id PAA22219; Sun, 28 Jul 2002 15:12:53 -0400 (EDT) Date: Sun, 28 Jul 2002 15:12:53 -0400 (EDT) From: rlw28@cornell.edu X-Sender: rlw28@travelers.mail.cornell.edu To: chemistry@ccl.net cc: rlw28@cornell.edu, rwood005@twcny.rr.com Subject: CHARMM Verlet dynamics Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi all I've managed to run 1 nanosecond of Verlet dynamics on my solvated peptide. The temperature output was around 300 K, which is what I wanted; unfortunately, the waters flew away! If I watched a movie of the dynamics, it looked like my system blew up. This is not good. Here were the settings that I used: DYNA VERLET NSTEP 70000 TIMESTEP 0.001 IHBFRQ 10 - IPRFRQ 1000 IHTFRQ 0 IEQFRQ 50 NTRFRQ 0 NSAVC 100 ISEED 314159 - NSAVV 0 FIRSTT 300 FINALT 300 IUNCRD 21 CDIE EPS 1.0 GROUP SWITCH - IUNWRI 20 ISVFRQ 1000 INBFRQ 20 CUTNB 13.0 CTOFNB 12.0 CTONNB 10.0 - TWINDL -10.0 TWINDH 10.0 IASORS 0 ISCVEL 1 IASVEL 1 VSWITCH ICHECW 0 IPRFRQ 1000 I noticed that I had ICHECW set to 0. I thought this might be the problem, so I set it to 1 (in other words, to check to see if the temp. falls into the window of 290 to 310 K). But this didn't help. So I went to the CHARMM documentation and I found the following example: "For late equilibration and analysis runs: DYNAMICS LEAP VERLET RESTART NSTEP 20000 TIMESTEP 0.001 - IPRFRQ 1000 IHTFRQ 2000 IEQFRQ 5000(*) NTRFRQ 5000 - IUNREA 30 IUNWRI 31 IUNCRD 50 IUNVEL -1 KUNIT 70 - NPRINT 100 NSAVC 100 NSAVV 0 IHBFRQ 0 INBFRQ 25 - hbond-spec nonbond-spec - FIRSTT 100.0 FINALT 300.0 TEMINC 100.0 - IASORS 0 IASVEL 1 ISCVEL 0 ICHECW 1 TWINDH 10.0 TWINDL -10.0 (*) Window checking should be disabled for the analysis run (i.e. IEQFRQ=0) if you want a real microcanonical ensemble." Now, this has some heating in it, which I don't want to do. So I reset my IEQFRQ from 50 to 0, changed ICHECW from 0 to 1, changed ISCVEL to 0, and changed NTRFRQ. This is what my settings are now: DYNA VERLET NSTEP 70000 TIMESTEP 0.001 IHBFRQ 10 - IPRFRQ 1000 IHTFRQ 0 IEQFRQ 0 NTRFRQ 500 NSAVC 100 ISEED 314159 - NSAVV 0 TSTRUCT 300 FIRSTT 300 FINALT 300 IUNCRD 21 CDIE EPS 1.0 GROUP SWITCH - IUNWRI 20 ISVFRQ 1000 INBFRQ 20 CUTNB 13.0 CTOFNB 12.0 CTONNB 10.0 - TWINDL -10.0 TWINDH 10.0 IASORS 0 ISCVEL 0 IASVEL 1 VSWITCH ICHECW 1 IPRFRQ 1000 The temperature after a few ps is around 145 K or so. Will it go up? Are my settings "correct"? Does the fact that I have harmonically constrained the ends of my peptide cause the waters to fly away? I appreciate any help you can offer. Richard