R.E.D. I
Frequently Asked Questions


Below, you will find a summary of FAQ gleaned from the net. Most of them come from the AMBER (and GAMESS) mailing lists. Questions are written in green color and the corresponding answers in black...

Please check also the AMBER mailing list archives (http://structbio.vanderbilt.edu/archives/amber-archive/ and http://amber.ch.ic.ac.uk/archive/) and AMBER Web site (http://amber.scripps.edu/Questions/resp.html and http://amber.scripps.edu/Questions/amber-archive.html) for other information about R.E.D. and X R.E.D.



  1) Announcement of the release of program R.E.D. version 1.0
  2) Re: AMBER: Which program to use to visualize the RESP Outputs derived by AMBER 7
  3) Re: AMBER: Which program to use to visualize the RESP Outputs derived by AMBER 7
  4) Re: AMBER: Which program to use to visualize the RESP Outputs derived by AMBER 7
  5) Re: AMBER: contributed parameters
  6) Re: AMBER: contributed parameters
  7) Re: AMBER: contributed parameters
  8) Re: AMBER: R.E.D.
  9) Re: AMBER: Some questions in building a new residue
10) Re: AMBER: Some questions in building a new residue
11) Re: AMBER: About RESP Calculation
12) Re: AMBER: About RESP Calculation
13) Re: AMBER: About RESP Calculation
14) Re: AMBER: About RESP Calculation
15) Re: AMBER: About RESP Calculation
16) Re: AMBER: About RESP Calculation
17) Re: AMBER: About RESP Calculation
18) Re: AMBER: About RESP Calculation
19) Re: AMBER: About RESP Calculation
20) Re: AMBER: esp or resp for all_amino94.lib
21) Re: AMBER: resp demo files
22) Re: AMBER: reproduction of RESP charges for Gly
23) Re: AMBER: reproduction of RESP charges for Gly
24) Re: AMBER: reproduction of RESP charges for Gly
25) Re: AMBER: reorientation in R.E.D
26) Re: AMBER: choosing 3 atoms for reorientaion in R.E.D.
27) Re: AMBER: charges (Cornell et al JACS, 1995) and R.E.D
28) Re: AMBER: charges (Cornell et al JACS, 1995) and R.E.D
29) Re: AMBER: charges (Cornell et al JACS, 1995) and R.E.D
30) Re: AMBER: parametrization partial charges
31) Re: AMBER: Bug in babel conversion
32) Re: AMBER: pdb in RED
33) Re: AMBER: Use of RED
34) Re: AMBER: Loading mol2 Files in XLEaP
35) Re: AMBER: unnatural base - RESP charges in ANTECHAMBER
36) Re: AMBER: Problems with X.R.E.D.
37) Re: AMBER: R.E.D. mol2 file and xleap
38) Re: R.E.D. and network (in french)
39) Re: [Gamess] failure to generate RESP charges
40) Re: AMBER: charge not zero
41) Re: AMBER: RESP partial charge calculation
42) Re: AMBER: RED II and single point
43) Re: AMBER: RESP charge derivation
44) Re: AMBER: Gaussian in RED
45) Re: AMBER: RED fails to read g94 output
46) Re: AMBER: problem with RED II
47) Re: AMBER: RESP calculation with different multiplicities
48) Re: AMBER: problem with RED II
49) Re: AMBER: problem with RED II
50) Re: AMBER: Differences in RESP charges for G in RG and DG residues
51) Re: AMBER: Test Calculation of atomic charges (RESP) for RNA bases
52) Re: AMBER: Test Calculation of atomic charges (RESP) for RNA bases
53) Re: AMBER: Test Calculation of atomic charges (RESP) for RNA bases


1) Announcement of the release of program R.E.D. version 1.0

From: FyD <fyd@u-picardie.fr>
Date: Wed Dec 03 2003 - 19:09:39 GMT


Dear Amber users,

I am pleased to announce the release of program RED version 1.0., http://www.u-picardie.fr/labo/lbpd/RED/.

R.E.D. (Resp Esp charge Derive) calculate automatically 'RESP' and 'ESP' charges starting from an un-optimized PDB structure. R.E.D. sequentially executes (i) either program "GAMESS" or program "Gaussian" to minimize the studied structure and to compute the corresponding Molecular Electrostatic Potential, and (ii) program "RESP" to fit the atom-centered charges to the grid previously determined. Format conversions needed during the procedure and "GAMESS", "Gaussian" and "RESP" inputs are automatically generated. Moreover, a new RESP fitting procedure is proposed allowing to get highly reproducible 'RESP' and 'ESP' charges whatever the QM software is and whatever the starting Cartesian coordinates are. R.E.D. makes the development of the 'RESP' and 'ESP' charges a straightforward, simple and highly reliable procedure. A graphical user-friendly interface, X R.E.D., has been also developed to execute R.E.D. and modify R.E.D. variables.

R.E.D. and X R.E.D. use is described in a downloadable manual. Program R.E.D. has been written with the "Perl" programming language and program X R.E.D. with the "tcl/tk" programming language, making R.E.D. and X R.E.D. highly flexible and portable. These two programs have been tested on PC-LINUX and SGI-IRIX workstations and should work on other UNIX platforms.

R.E.D. and X R.E.D. are "free" (i.e. provided at no cost) for all academic users after signing a license.

Best regards, Francois

F.-Y. Dupradeau
  --
The Scripps Research Institute, San Diego, CA
Faculte de Pharmacie, UPJV, Amiens, France
  --
http://www.u-picardie.fr/labo/lbpd/fyd.htm



2) Re: AMBER: Which program to use to visualize the RESP Outputs derived by AMBER 7

From: FyD <fyd@u-picardie.fr>
Date: Fri Jan 23 2004 - 19:52:28 GMT


I am new in using AMBER 7, and I need to calculate the RESP charges for a specific nucleic acid molecule. After using the G98 to minimize the structure and to find the ESP charges in a HF/6-31G(d) field-basis set, and then using 'antechamber' to get the RESP charges, I got the *.pdb file. (It was automatically created by antechamber) I want to visualize this pdb file such that I can see the RESP Charges next to each corresponding atom. I have used the VMD, but it seems that there is no command to put these RESP charges next to each corresponding atom. Can you give me any suggestions how to do this? It is possible to put the charges by first getting a picture of the molecule, and using ADOBE Photoshop to write down the charges seperately, but it will take some time of me. Any suggestions are welcome and thanks in advance...

I guess you used such command in "antechamber":
"antechamber -i job.log -fi gout -o job.pdb -fo mpdb -c resp"

In my opinion the PDB file generated by "antechamber" (with the RESP charges in it), can NOT be used in charge visualization in some graphic interfaces or even directly used in MM simulation (I do not think that "leap", "insightII", "Sybyl", "VMD" can read charges from the PDB format, at least in amber7 concerning "leap"). To tell the truth, I even do not understand why the RESP/ESP charges can be available in this PDB file (generated by "antechamber") because default PDB format does NOT have charges (see http://www.rcsb.org/pdb/docs/format/pdbguide2.2/guide2.2_frame.html): I guess it is simply for information or this may change in "xleap/amber8"...

It is why in RED, http://www.u-picardie.fr/labo/lbpd/RED/, we decided to use the Tripos format .mol2 where there is indeed a column for these charges in the format (http://www.tripos.com/custResources/mol2Files/). Thus, with RED, after structure optimization, MEP calculation and RESP fit, a Tripos structure 'can' be generated with the optimized Cart. coordinates and charges. We do use such .mol2 file to visualize the charges using insightII for instance:

Molecule/Label/Label_property = Partial_charge....

I guess it should also be possible under Sybyl... "amber7/xleap" can also visualize charges but I do not think "xleap" can load Tripos format (or may-be in the future amber8 version...).


3) Re: AMBER: Which program to use to visualize the RESP Outputs derived by AMBER 7

From: Bill Ross <ross@cgl.ucsf.edu>
Date: Fri Jan 23 2004 - 19:26:15 GMT


Possibly xleap will display charges, which would come from a prepin file generated by antechamber in your case. Pdb-based viewers would presumably require that the charges be included as the optional 'temperature factor' field at the end of the ATOM line:
http://bmerc-www.bu.edu/needle-doc/latest/atom-format.html

Leap can save pdb files with this info, try 'help set' to see how it is done. Of course the charges must be there to be saved, and 'desc' can be used in leap to verify that they are present.



4) Re: AMBER: Which program to use to visualize the RESP Outputs derived by AMBER 7

From: FyD <fyd@u-picardie.fr>
Date: Tue Jan 27 2004 - 16:59:51 GMT


Thanks for your response. I have converted the file to a .mol2 file, and tried to look at the molecule using VMD

I do not think VMD can display charges...

and Insight II, but could not see the charges in these programs.

In InsightII:

Molecule/Label
Label Option = Single property
Label level = Atom
Label_action = on
Molecule Spec = 'Your molecule'
Label property = Partial_charge

=> charges are displayed with 2 digits.

To display charges with four digits:

Session/Environment
Toggle 'Precicion' to 'on'
Precision = '4' instead of '2'

=> charges are displayed with 4 digits i.e. as in the AMBER 'prep' files...


I will try to check the file using some other programs, like Maestro and MacroModel. The main reason why I am writing to you is to ask a specific question about RESP Charge calculation. I am trying to recalculate the results done in Cornell et.all 1995 paper, specifically for the Guanine and Cytosine bases, with different sugar molecules attached to these bases.

You will NOT be able to reproduce the published values! It is why we wrote RED http://www.u-picardie.fr/labo/lbpd/RED/ to be able to get (highly) reproducible RESP/ESP charges whatever is the ab-initio software is ("GAMESS" or "Gaussian" as examples) and whatever is the starting structure representing the target minimum...

I have modeled a Cytosine with a sugar molecule attached to it, similar to the one done in Cornell et all 1995 (For the RNA Case). The molecule is in Fig. 6 on that paper. The next step was to find the RESP Charges for this molecule. So, I am using first g98 to get the ESP fit and then antechamber to get the resp charges. Before using the Gaussian, I changed the first line of the input file with:
#P HF/6-31G* SCF=Tight Pop=MK Iop(6/33=2, 6/41=10, 6/42=17)

From our tests SCF=Tight is NOT usefull. Only the 'accuracy' (Opt=Tight or even VTight) of the optimized structure is important... SCF=Tight in single point calculation does not affect the MEP.

Moreover, it is explained at http://signe.teokem.lu.se/~ulf/Methods/resp.html"
"IOp(6/33=2)" makes Gaussian write out the potential points and potentials (do not change).
"IOp(6/41=10)" specifies that 10 concentric layers of points are used for each atom (do not change).
"IOp(6/42)" give the density of points in each layer. A value of 17 gives about 2500 points/atom. Lower values may be needed for large molecules, since the programs cannot normally handle more than 100 000 potential points. A value of 10 gives about 1000 points/atom."


I ran a small calculation on MeOH using "Iop(6/33=2, 6/41=10, 6/42=17)" or "Iop(6/33=2)" and the number of MEP point is totally different with "Iop(6/33=2, 6/41=10, 6/42=17") (a huge number of MEP points is reported). I do not think that this corresponds to the Connolly surface reported by Kollman and Singh. I would only use what is recommended on the AMBER web site i.e. 'only' "Iop(6/33=2)" (see http://amber.scripps.edu/Questions/resp.html "handy script/program to convert g94 output").

I got this command in one of the pages in internet. A prof. in Sweden, Prof. Ulf Ryde, was using this method while calculating the RESP charges. So, after getting the g98 output file, I used the antechamber with the following command:
antechamber -i <input file> -fi gout -o resp.prep -fo prepi -c resp -s 2

When I used this procedure for a Cytosine with only Hydrogen atom attached to the base, rather a sugar molecule, it worked fine, and I got the RESP charges. But when I have a sugar molecule similar to the one that is used in Cornell et. all 1995, it failed, and the following error message came out while using antechamber: Running:


resp -O -i --------------> The Resp1 calculation
fmt: end of file
apparent state: unit 10 named ANTECHAMBER.ESP
last format: (2i5)
lately reading sequential formatted external IO

resp -O -i ... -----------------> The Resp2 calculation
Unit 3 error on OPEN: gout
Cannot open QOUT, exit


I think Prof. Ulf Ryde compared different ways to get ESP charges (not only RESP charges)...

I checked out the RED.pl file you have written, and there, you are using the following command while doing the MEP Calculation:
#P HF/6-31G* Pop=MK SCF(Conver=6) Iop(6/33=2) NoSymm Test


We simply used what is reported on the AMBER web site. 'NoSymm' can be added to keep the orientation selected by the rigid body re-orientation algorithm available in RED...

What might be the problem that I am encountering while following my procedure? In Cornell's paper, 4 surfaces need to be defined to calculate the RESP charges, and I thought that Iop(...) is giving this definitions.
Then I have changed the first command of the g98 file with the way of yours, and did the same thing: First found the ESP fit points, and then used antechamber. I did not have any error messages in this case, and got the RESP Charges. When I compared with the Cornell et. all 95 results, some of the charges were not close to the results of the paper.


I am not surprised: you can not reproduce published data. It is why we wrote RED...
For the Cornell et al. paper, the charges have been calculated with "Gaussian". As the orientation of the optimized structure obtained from "GAMESS" and "Gaussian" are different, you will not be able to reproduce the published values if you use "GAMESS"... Now if you have the money to use "Gaussian" ;-) I am sure that the orientation of the optimized structure you obtained is different from the orientation of the optimized structures obtained in the Cornell et al paper (even if you do use the default keyword 'Symmetry')... As the orientation are different, the charges are different and the differences can be quite important... See RED manual page 13...


I would love to hear what you think might be the reason why my first method did not work. Thanks in advance.

In the MeOH test I ran with "Iop(6/33=2, 6/41=10, 6/42=17)", there are so much MEP points that their numbers are replaced by "*****" at the end of the g98 output. A reason could be that "Antechamber" does not like such characters, I do not know...


5) Re: AMBER: contributed parameters

From: FyD <fyd@u-picardie.fr>
Date: Wed Feb 18 2004 - 17:12:04 GMT


If one would like to generate his or her own RESP charges for pseudouridine or any modified nucleotide using Amber 7. How would one go about it?

-1- You could apply the strategy detailed in Cieplak et al. J. Comput. Chem 1995, 16, 1357-1377. In particular, see the 'Summary page' 1367. In such startegy, you need to use two different Electrostatic Potential (from dimethyphophate and your modified nucleoside) and you use inter-molecule restraints (see Figure 2).

-2- A different approche would be to use your modified base with a methyl group (instead of the sugar) and use a restraint of .1268 (representing the backbone) for it as it is demonstrated in Spector et al. JACS 1997, 119, 7095-7104.

The second approch might be more simple and will demand really less computer time.



6) Re: AMBER: contributed parameters

From: FyD <fyd@u-picardie.fr>
Date: Fri Feb 20 2004 - 17:08:34 GMT


1) My concern for using the first approach is that C1' atom bears a fairly large negative charge, and the nearby atoms of the modified base (close to C1') are also affected to some noticeable degree.

Yes it seems strange. When I applied this 1st startegy on my unusual nucleotide I got similar charge values that those reported in table VI page 1368. I think deriving RESP charges when you do it for the 1st time is quite tricky and error prone. For instance, in this approach, C1 charges must be free; i.e. without inter-molecular restraints (Are you sure you applied this ?)... We wrote RED http://www.u-picardie.fr/labo/lbpd/RED/ to help people to use "RESP". You will find tutorials but it is only for simple cases; i.e. there is no intermolecule restraints. Anyway, you might found some information useful in the manual also.

Secondly, if I were to use the RESP charges generated using the first approach, do I define a new residue in xleap for the modified nucleotide?

I would say you load the PDB file (it is obviously not recognized). Then, you create the bonds (bondbydistance command). You enter 'manually' the RESP charges you calculated and the atom types you selected (i.e you can "edit" the unit) and you save the OFF library (.off file), coordinates (prmcrd) and topology/FF information (prmtop). Try for instance this:

source leaprc.ff98
M = loadpdb Mol.pdb
# M is the 'unit' name, Mol.pdb is your PDB fike
bondbydistance M
edit M
# You enter manually the charges values
# And atom type in the table displayed

saveoff M M.off
# M.off can be loaded with 'loadoff M.off'
saveamberparm M prmtop prmcrd


It is done for 'in vacuo' simulation. But you might want to add ion and solvate your system in a box of solvant...

source leaprc.ff98
M = loadpdb Mol.pdb
bondbydistance M
edit M
# => Add charges and atom types...
solvateoct M WATBOX256 10
addions Na+ 0
saveoff M M.off
saveamberparm M prmtop prmcrd


Ready for, periodic simulations...

2) If I were to use the second approach as you mentioned with the restraint of .126

First I would use '0.1268' and not '0.126' that will help you to get a 'more neutral system' and this might be problematic if you want to neutralize with ions in "leap"...

(during the second stage, I assume),

No, it is in the 1st stage and after you freeze the charges in the second stage...

I would then assign these RESP charges to the base atoms in the PREP module as suggested in JACS, 119, 7095-7104). My question is that what charges I would use for C1' and H1' atoms in this case.

The strategy is different, in the Cieplak's paper the C1 atom is free because inter-molecular restraints are used (allows the system to be more free, allows to get a better fit between the electrostaic potential and the charges). In the second approche, you do not use inter-molecular restraints because your unusual base is rigid...


7) Re: AMBER: contributed parameters

From: FyD <fyd@u-picardie.fr>
Date: Sat Feb 21 2004 - 00:05:48 GMT


If I were to reproduce the results in table VI page 1368 applying the summary page 1367,

First, you will NOT be able to reproduce the published values... It is why we developed RED. With RED you can control the orientation of your minimized molecule and thus the charges are reproducible... We found that for a same minimum, RESP charge differences up to ~ .045 e for an atomic charge can be observed if the molecular orientation is 'controled' (or better to say 'not'- controlled ;-) by the ab-initio software ("Gaussian" or "GAMESS")...

1) Do I apply the intermolecular restraint for phosphate-joints to each nucleoside (G,C,A,T) and at the same time (in the first stage) the intermolecular sugar equivalencing restrain (except C1' and H1') is applied to the four nucleosides?

Well, I think it is clear in the Cieplak et al JCC paper, in the Summary page 1367. The listing of all the restraints is available this section: 5 different types of restraints in the 1st RESP stage and only methyl and methylene groups are re-fit in the second stage... This means you need 5 molecular electrostatic potential (4 nucleosides and the DMP)...

2) If the approach in 1) is correct, would adding modified nucleotides on top of four regular nucleosides (G,C,A,T) and using the similar strategy work

Ouf!!! It means huge computer time because you have to get ab-initio data for the 4 'normal' nucleoside minima... My understanding is that the goal of the JCC paper was to develop charges for the REGULAR nucleotides. In your case, why not simply using different conformations of your unusual nucleoside (you have to use inter-conformational restraints; you can see examples in the RED manual where we used multi-orientation fit; orientation and conformation are treated in the same way in the RESP inputs...) with DMP (with inter-molecular restraints between DMP and the different conformations used).

or why not using the Spector et al strategy as I alread told you ?


- I mean if I should still expect similar charges for the four bases and their C1' and H1' atoms as generated in 1)?

We applied Cieplak et al. strategy using ONLY different conformations of an unusual nucleoside (with DMP also) and the C1' and H1' charges were very similar that those reported in the JCC paper for the 4 regular DNA nucleosides...

Is that how you applied your unusual nucleotide?

Unusual means to me not the regular 4 DNA nucleosides...


8) Re: AMBER: R.E.D.

From: FyD <fyd@u-picardie.fr>
Date: Tue Feb 24 2004 - 01:42:13 GMT


Just wonder if there is any publication out about MEP calculations used in R.E.D.

See section VII of the RED manual:
A. Pigache, P. Cieplak & F.-Y. Dupradeau, Automatic and highly reproducible RESP and ESP charge derivation: Application to the development of programs RED and X RED, Submitted/(accepted now) to the 227 th ACS National Meeting, Anaheim, CA, March 28 - April 1, 2004 (and manuscript in preparation).



9) Re: AMBER: Some questions in building a new residue

From: Chung-Chien Wei <hiro@tsailun.chem.ntnu.edu.tw>
Date: Wed Mar 10 2004 - 19:49:33 GMT


In Prof. Duan's work "A Point-Charge Force Field for Molecular Mechanics Simulations of Protein Based on Condensed-Phase Quantum Mechanical Calculations", on J. Comput. Chem., the section "Charge Derivation" wrote:

"Effective charges were obtained by fitting the electrostatic potential of peptides, using RESP method. A two-stage fitting procedure was used. (1)In the first stage of fitting, the two conformers (alpha & beta) of each dipeptide (ex:Ace-Ala-Nme) were combined. (2)In the second stage, the chemically equivalent atoms were set to hove the same charges, while the charges of the terminal blocking groups and those of heavy atoms were fixed. Because the charge-marching process may introduce small errors, we purposely limited the errors to the matched atoms by fixing the charges of the blocking groups and heavy atoms. (3)Finally, the charges of the blocking group were fit by combining electrostatic potentials of all amino acids."

Now I have some questions:
Question 1 is about step(1) that how to "combine" the two conformers ? (just averaging phi & psi angles ?) Is it just to get single conformation ?
Question 2 is about step(3) that if you use dipeptide (Ace-Ala-Nme) to run, how do you remove charge of Ace & Nme and retain the charge of Ala part zero ? (Because charge of Ala in AMBER library without Ace & Nme is zero, and not close-shell.)

So I hope that someone can instruct me explicit steps for building a new residue well.


I have come to the same problem too. I am trying to recalculate the RESP charges for Cytosine molecule which was originally done by Cornell et all 95 paper. They have the RESP charges for Cytosine for different sugars attached to the base. As far as I understood from the scripps' website, they are using the '#P mp2/6-31g*' basis set. And the results are done for this space. But in the RESP Help page of scripps, they suggest to use hf/6-31g* with Iop(6/33=2) command function in the gaussian. The main problem for all of us, I think, is the initial unoptimized structure of the molecule. There can be a lot of conformations for a specific molecule, and depending on which conformation you choose, the resultant RESP charges will be different. For cytosine molecule, it is easy to find the initial structure, but if someone wants to do the calculations for a structure which does not have an explicit crystal structure in the databases, this will be a problem.


10) Re: AMBER: Some questions in building a new residue

From: FyD <fyd@u-picardie.fr>
Date: Wed Mar 10 2004 - 20:31:25 GMT


As said Ilyas Yildirim, it is very unlikely you will be able to reproduce the published charge values. It is obviously true that for a same molecule, different conformations can be selected and then different sets of ESP/RESP charges produced. A more tricky problem is that even for a SAME optimized conformation, the molecular orientation also affects charge values... (And I do not talk about the optimization thresholds used by the ab-initio software that can induce also charge differences...) Even for small molecules like EtOH, DMSO or DMP (which seems to be very simple cases for deriving RESP charges) the charge differences obtained can be highly disturbing and tricky...

Now concerning the charge differences you reported (in comparison to the work of Duan et al.), in my opinion, you have errors in your inputs (the differences are too big!). To learn how to derive RESP charges, a very good starting point is the work of Cieplak et al. in J. Comput. Chem. 1995, 16, 1357-1377. We also wrote the program RED http://www.u-picardie.fr/labo/lbpd/RED/ to help people to derive RESP charges. A multi-orientation RESP/ESP fit can be automatically performed (by analogy to the multi-orientation RESP fit). The new version of RED will incorporate automatic multi-orientation, multi-conformation and extra-point handling...

In the Cieplak et al. paper, there are details about how to use blocking groups (NMe & ACE, as you are interested in...) in the RESP charge derivation and more generally how to use multi-conformational RESP fit with or without charge group restraints.



11) Re: AMBER: About RESP Calculation

From: David A. Case <case@scripps.edu>
Date: Fri Mar 19 2004 - 16:14:29 GMT


4. I tried to follow Cornell's paper to recalculate the results published in 1995. I read the RESP documents that are in the scripps website, and in those info files it is saying that in order the results to be compatible with Cornell et al. 95 force field,
#P MP2/6-31G* pop=mk iop(6/33=2)
needs to be used in the gaussian input file.

Can you point to the exact URL where this is? I am not doubting you, by my Google search failed to find this... The first statement in the FAQ entry on RESP states: "MP2 is not used for obtaining the electrostatic potential to fit." We will try to scrub out or explain better any contradictory statements, if we can find them.


12) Re: AMBER: About RESP Calculation

From: FyD <fyd@u-picardie.fr>
Date: Fri Mar 19 2004 - 20:10:19 GMT


I am trying to recalculate the Cornell et al.95 results, but as far as I understood, it is not possible because of the orientation problem of the molecule at interest.

Well you might be able to reproduce the charge values published if you the same basis set, the same ab initio software, the same thresholds criteria to minimize the structure and if you are lucky concerning the generation of the molecular orientation generated by the ab initio software. Moreover, you might be able to reproduce the charge values for one model (however at which accuracy: two??, three?? or four digits?? after the decimal point) and not for an other one.


13) Re: AMBER: About RESP Calculation

From: David A. Case <case@scripps.edu>
Date: Fri Mar 19 2004 - 20:15:39 GMT


We got the AMBER 7 a couple of months ago to our group, and I read the document 0README which is under the resp directory. (Specifically /amber7/src/resp) It was described in Mr. Caldwell's email which is part of this 0README file.

Ouch. OK, it looks like this file is out-of-date and somewhat misleading. My suggestion is that Amber users basically ignore the $AMBERHOME/src/resp/0README file (in both amber7 and amber8). There are several options to run RESP, in (my) recommended order:

1. Use the RED interface; there is a pointer on the Amber web site;
or
2. Use antechamber to create the required input files.
or
3. Follow the instructions in the above 0README file, but use HF, not MP2 for doing electrostatic potential fitting.



14) Re: AMBER: About RESP Calculation

From: FyD <fyd@u-picardie.fr>
Date: Fri Mar 19 2004 - 20:30:00 GMT


but as far as I understood, it is not possible because of the orientation problem of the molecule at interest.

The question is not "to be possible" or "not be possible to"... the question is the reproducibility of the published charges, that's it... F.


15) Re: AMBER: About RESP Calculation

From: Chris Moth <Chris.Moth@vanderbilt.edu>
Date: Fri, 19 Mar 2004 15:38:40 -0600


I have a couple of question about RESP calculation. I found the .pdb file of a Cytosine molecule in Cambridge Crystalographic Database, CYTIDI10. I want to find the RESP charges of this particular molecule and have a couple of questions to ask:

I have been performing RESP parameterizations of small molecule enzyme inhibitors. As a relative newcomer (lowly grad student) to RESP parameterization myself, I thought it might help if I mention a few of my personal experiences, observations, and conclusions.

I have found there to be a gulf between the (excellent) treatment of RESP in the literature, and specific steps needed to parameterize a molecule prior to starting an MD simulation. My comments reflect _my_ navigation of this gulf, and share some lessons I have learned the hard way.

If any of these ideas are misguided, I encourage correction from the more experienced folks on the list!


1. I have tried to use the program RED written by Mr. Dupradeau. I entered the .pdb file and it took 3 days and still didnt end up with the calculation. I killed the job. Are the gaussian calculations taking that long?

I have seen g98 geometry optimization calculations take this long when the start structures are far from a reasonable local minimum in energy.

BTW, here is the third line of my g98 geometry optimization input file: #RHF/3-21G* OPT Test.
Then, for MEP calculation prior to RESP we use: #RHF/6-31G* SP Pop=ESP iop(6/33=2) TEST.
The other folks on the list who have already responded know far more about this however...

By first coarsely minimizing the molecular structure using a quick-and-dirty electrostatic assignments in a package like MOE first, I've found that g98 runs to completion within 12 hours or so. (50 to 60 atom molecules - 300MHZ R12000 SGI Uniprocessor machine) You can probably get the same kind of results with INSIGHTII. If there is a MMF94 force field option, you might give that a try.


I did not change the ATOM NAMES, and later I realized that in the sugar molecule, there is a CH2 term. Is it possible that this is the reason for this problem? (If it is a problem for the gaussian to calculate a particular job more than 3 days)

Gaussian does not care about atom types or names or bonds - just atomic numbers (element symbols) and their coordinates.

2. Do we have to optimize the structure of a particular molecule in order to calculate the RESP charges? For instance, can I just use the crystal structure I got from the CCD and just do the MEP calculation and then the RESP calculations? Is this legitimate?

Here is exactly the rub. My read of the literature is that the ideal electrostatic parameterization would involve Bolzmann weighted averaging of charges derived from RESP fits of all reasonable conformers. However, that is not computationally tractable.

So - there is a real human factor here. Things are going to depend on your particular project, and your chemical intuition.

Charges do vary considerably with conformation, and slightly with orientation, to be sure. (I've played around with this myself - and have seen variations of up to 3/10ths of a q through conformational change) I get through the day by taking the Cornell et. al. charges as they come - since they have been extensively put to the test, and have been shown to exquisitely mimic natural phenomena (see papers predicting A vs B form DNA duplex transitions, oligo/ligand binding, and on and on etc).

For parameterization of my own molecules, I have had to factor in the downstream analysis needs of my work. (I.e. - I goofed things up the first time around, and now am doing a better job). For example - a couple of rules of thumb I work with now are:

- Ligand coformers with intramolecular hydrogen bonds should not considered (this was mentioned in the original Cornell et al paper). RESP fitting of these conformers places far too much electronegativity on the hydrogen bonding heavy atoms. And, this conformation is not going to be "right" for ligand bound to protein

- Stereoisomers should be parameterized identically. This seems obvious on the surface... and, indeed, in my experience, RESP fitting mirror image molecules did result in differences in only the 1/100ths of a q. By bypassing RESP fitting of the sceond enantiomer, and manually setting identical charges, you honor the MM model. And, at the end of the day, this is the only way you can make clean comparisons of binding energetics between R vs S enantiomers at the MM level of theory, anyway.

So, I have concluded that is more important for electrostatics to be consistently derived throughout a project than be "correct". But yes, at some level, it must be the case that including more conformers in a RESP fit will improve "correctness" So, for your work, perhaps averaging a putative global minimum (use coarse MOE systematic conformational search) conformation, and a crystal structure conformation might be a way to go.

And of course, if you see a positive charge on an oxygen, you've goofed somewhere :)


3. In the RED program, it is talking about re-orientation. I followed the manual, and did not understand exactly what it is meant by reorientation. In the tutorials there are examples for re-orientation, but after getting an optimized structure, why do we have to reorient the structure? The structure has a clear structure when viewed in InsightII or Molden. So, isnt it legitimate to do the MEP calculation in Gaussian and then the RESP calculations without reorient the structure?

I think others have addressed this far better than I can. I have not yet used RED - but Gaussian does build a rectangular box of electrostatic points around the ligand, and the orientation of the box vs the ligand will affect final charges assignments a bit.

Perhaps in a perfect world, g98 would have an option to build a perfect unvarying sphere of points around the ligand center-of-mass :) But, I'm sure that would defeat other requirements of the MEP which involve building ESP points at fixed multiples of VdW radii......

I hope some of this has provided useful "food for thought" without being too informal or too much in error.



16) Re: AMBER: About RESP Calculation

From: Qing Zhang <qingzhang_nyu@yahoo.com>
Date: Sat Mar 20 2004 - 06:43:03 GMT


but as far as I understood, it is not possible because of the orientation problem of the molecule at interest.

The published charges can be reproduced well, although not exactly, if one strictly follows the paper:
Piotr Cieplak, Wendy D. Cornell, Christopher Bayly, & Peter A. Kollman, "Application of the multimolecule and multiconformational RESP methodology to biopolymers: charge derivation for DNA, RNA, and proteins", J. Comp. Chem., 16: 1357-1377, 1995.


Based on this paper and previous discussion with Dr. Cieplak, I have formulated a protocol that produces AMBER-compatible partial charges of carcinogen-modified B-deoxynucleotides. The protocol has been successfully applied to several carcinogen-modified B-deoxynucleotides. One can also easily apply it to other unusual deoxynucleotides. The protocol is at: http://monod.biomath.nyu.edu/~qzhang/BP-adenine/protocol.txt

The protocol includes how to modify a molecule structure in insightII, how to optimize the structure and produce ESP in Gaussian98, and how to produce RESP charges in AMBER 6. The crystal structure should be modified not only to reproduce AMBER partial charges, but also to remove bumps between non-bonded atoms because Gaussian does not know where the bonds are given only atom coordinates. Optimization in Gaussian (HF/6-31G*) can take several to ten days depending on the molecule size and optimization convergence.

The protocol also comes with a few Perl programs that make the procedure easy: http://monod.biomath.nyu.edu/~qzhang/Research.htm



17) Re: AMBER: About RESP Calculation

From: FyD <fyd@u-picardie.fr>
Date: Sun Mar 21 2004 - 07:30:41 GMT


The published charges can be reproduced well, although not exactly, if one strictly follows the paper: Piotr Cieplak, Wendy D. Cornell, Christopher Bayly, & Peter A. Kollman, "Application of the multimolecule and multiconformational RESP methodology to biopolymers: charge derivation for DNA, RNA, and proteins", J. Comp. Chem., 16: 1357-1377, 1995.

- Using program Gaussian, the charges "can be" reproduced "although not exactly" as you said. ;-)

- Using program GAMESS the differences in charges "might be" slightly more important...

Now, what is the meaning of "can be", "although not exactly" and "might be" ? Reproducibility after 2, 3 or 4 digits after the decimal point ?

As I said before, in a previous email, from our tests, we found a charge difference up to 0.043 e. Such differences strongly depend on the model studied (It is why RESP charge derivation might be tricky in 'some' cases). Good examples are models EtOH or DMSO which present 2 different orientations (after minimization using the re-orientation scheme available in Gaussian) and a maximum charge difference of 0.043 e for one heavy atom. On the contrary, for some other models we were only able to generate a unique orientation using Gaussian. In this case, the charge differences were obviously smaller and the charge reproducibility really better (directly related in this case to the minimization threshold criteria used in Gaussian).


One can also easily apply it to other unusual deoxynucleotides. The protocol is at:
http://monod.biomath.nyu.edu/~qzhang/BP-adenine/protocol.txt


Let's follow your link: dimethylphosphate (conformation gauche gauche) is another good example showing that RESP charge derivation can be tricky... Why deciding that Methyl II is 'the' group II and not the group IV and why deciding that O3' is O3' and not O5'? Indeed, the charge differences between the group II and group IV is exactly the ones induced by the orientation of the 2 methyls in the conformation gauche gauche. This means that there is an incertainty concerning the charge of atom O3' and O5'. What is the difference in charge obtained in this case for the 2 methyl carbons ? 0.016 This is smaller than 0.043 but however means that molecular orientation represents an incertainty of 20 % of the charge value. Such difference may not alter MM simulation but the orientation is anyway responsible of an charge incertainty of 20 %.

Thus our goals in developping RED were:
- To solve "inaccuracy problems" related to the RESP ESP charge derivation protocole as those described above.

- To derive highly reproducible RESP and ESP charges using "Gaussian" BUT ALSO "GAMESS", i.e. WHATEVER the ab intio software is.

- To provide an automatic way to derive RESP (and ESP charges).

- To define a set of charges for a particular orientation allowing to reproduce published values and checking errors.



18) Re: AMBER: About RESP Calculation

From: FyD <fyd@u-picardie.fr>
Date: Fri Mar 19 2004 - 17:45:22 GMT


1. I have tried to use the program RED written by Mr. Dupradeau.

Program RED was written by A. Pigache and the corresponding study was carried out in collaboration by: P. Cieplak and F.-Y. Dupradeau

I entered the .pdb file and it took 3 days and still didnt end up with the calculation. I killed the job. Are the gaussian calculations taking that long?

You should check how "works" your optimization: Does it oscillate ?
A good way to check this is to apply the following shell command:
egrep "SCF Done|Step number|Maximum Force|RMS Force|Maximum Displacement|RMS Displacement" *.log

Then, you check at which step number the four threshold criteria are the smallest and you use the corresponding "Standard orientation" as input for a second optimization run using this time Opt=(Tight,CalcFC) (see RED manual page 12). Once it is done you run RED as it is describe at the end of page 7 of the RED manual in order to read the the "Gaussian" minimization output as input for MEP computation. (In the future RED version "GAMESS" ouputs will be read as inputs for "Gaussian" calculation...)


I did not change the ATOM NAMES, and later I realized that in the sugar molecule, there is a CH2 term. Is it possible that this is the reason for this problem? (If it is a problem for the gaussian to calculate a particular job more than 3 days)

The atom name are modified to create automatically RESP inputs (see RED manual page 13). That has nothing to do with Gaussian/GAMESS optimization.

Finally, it is impossible to say that 3 days are a long computation time as you do not report which machine type you are using.


2. Do we have to optimize the structure of a particular molecule in order to calculate the RESP charges? For instance, can I just use the crystal structure I got from the CCD and just do the MEP calculation and then the RESP calculations? Is this legitimate?

I am not an expert in that... I would say that ab initio and X-ray structures can present structural differences. Moreover, the ab initio minimization is performed in gaz phase and I think X-ray structures are more "condensed phase" structures. So I would apply the ab initio minimization step in the derivation of RESP /ESP charges...

3. In the RED program, it is talking about re-orientation. I followed the manual, and did not understand exactly what it is meant by reorientation.

There are at least 4 papers that describe the molecular orientation effect on the charge values: Two of them are referenced in the RED manual.
Spackman JCC 1996, 17, 1-18
Merz Jr JCC 1992, 13, 749-767
Woods et al. JCC 1990, 11, 297-310
Breneman & Wiberg JCC 1990, 11, 361-373

The ab initio software ("Gaussian" and "GAMESS") can/does re-oriente the structure also: You should check the orientation of the "Standard orientation" generated by "Gaussian": you will see that it changes during the minimization... It means we simply added in RED a new/another re-orientation method allowing to get highly reproducible RESP charges whatever the ab initio software is.


In the tutorials there are examples for re-orientation, but after getting an optimized structure, why do we have to reorient the structure?

It means that using your minimized structure, if you cannot get a constant orientation (generated by the ab initio software you use) for the Cartesian coordinates on which is computed the MEP, you will get different set of charges for each orientation possible. And that is THE point: GAMESS and Gaussian canNOT generate a constant orientation for a minimum: Did you try to optimize EtOH using GAMESS and Gaussian (see RED tutorial in the manual) ? Using both software we got 4 sets of different charges. Moreover, we were NOT able to reproduce the charges published by Fox & Kollman JPC B 1998, 102, 8079-8079. The maximum diff. observed was .043 for one carbon atom, same problem for the DMSO... Thus, in some cases the ab initio software can generate several orientations for the same minimum. Moreover, the orientation generated by GAMESS and Gaussian are different. This means that different sets of RESP/ESP charges can be potentially generated for the same minimum.

To solve this problem, we decided to add a "general?" "common?" re-orientation procedure for both ab initio software. It is the rigid body re-orientation that is applied on the minimized structure just before the MEP calculation. Thus controlling the orientation selected using this algo., we can get reproducible RESP charges. A multi-orientation RESP fit (by analogy to the multi-conformational RESP fit) can also be executed in RED which allows to average the charge differences observed for one orientation over several orientations. I think all this is more or less explained in the RED manual pages 13-14. However, it is clear it is not easy understandable without the corresponding paper...


The structure has a clear structure when viewed in InsightII or Molden. So, isnt it legitimate to do the MEP calculation in Gaussian and then the RESP calculations without reorient the structure?

It is up to you. The answer is the following: Do you want that people that will read your work will be able to reproduce the charge values you will publish ? If you care about this, apply the re-orientation algo. available in RED otherwise do not apply it in RED (and it was the way to derive RESP/ESP charge before writting RED).

4. I tried to follow Cornell's paper to recalculate the results published in 1995. I read the RESP documents that are in the scripps website, and in those info files it is saying that in order the results to be compatible with Cornell et al. 95 force field, #P MP2/6-31G* pop=mk iop(6/33=2) needs to be used in the gaussian input file. I use one of the cytosine molecule used in Cornell et al. 95 paper. The gaussian gives an error while doing the calculations. I do not exactly remember what the error sentence is, but it needs more space in the harddrive, more than 20 Gbyte space. Is this normal? It happened to me twice.

What is your cpu ? Is it a 32 bit cpu ? Did you check the size of your temporary files. If one is bigger than 2GB, your calculation will be killed. In this case, you have to split the temporary file into several ones, smaller. See gaussian 98 Reference page 17-18.

I used #P HF/6-31G* pop=mk iop(6/33=2) SCF(Conver=6) Nosymm Test which I found in Mr. Dubradeau's RED program manual, and it gives me a result without any error. PS: The above Gaussian command was described in an email sent by Jim Caldwell to amber@cg1.ucsf.edu in Feb 21, 1996 with a subject "esp to resp".

Well if I remember well you also reported the use of other Gaussian keywords to derive RESP charges available at: http://signe.teokem.lu.se/%7Eulf/Methods/resp.html
and it did not work in antechamber...

To conclude, the derivation of RESP charges can be really tricky because numerous options are possible and different charges for the same minimum can be obtained. It is a reason also why we work on RED, to allow people to publish reproducible charge. By this way, we can really judge were are the errors in the protocole. I hope RED will provide a 'standard' protocole for RESP and ESP charge derivation... The new version will incorporate multi-conformation and multi-orientation RESP/ESP fit, extra points (lone paires) handling.



19) Re: AMBER: About RESP Calculation

From: FyD <fyd@u-picardie.fr>
Mon Mar 22 2004 - 17:09:00 GMT


Thanks a lot for your detailed explanation on the inaccuracy caused by "orientation". I also visited your website. If I understand correctly, the "orientation" refers to both molecule transformation (whole molecule rotation and/or translation) and internal conformation change (such as the two methyl groups in dimethylphosphate). So I have three questions as following:

Orientation and conformation are different... If you take the same conformation, you can align (by translations and rotations) this conformation on the X axis or on the Y axis (for instance) and this conformation will have two different orientations (i.e 2 different set of Cartesian coordinates).

About conformation and ESP charges: See Stouch et al JCC 1992, 13, 622-632.

About orientation and ESP charges: See ref. 14, 15 in RED manual and previous emails.


What is your method to solve the inaccuracy caused by orientation? I see a "new RESP fitting procedure" mentioned on RED website. Since the RESP fitting procedure is after Gaussian or GAMESS calculation on a specific orientation, how does your new procedure eliminate the orientation effects from Gaussian or GAMESS results?

A rigid-body re-reorientation algo. applied on the minimized structure (from Gaussian or GAMESS). See RED manual pages 13-14.

Do you average RESP results on many orientations?

Yes, we do average RESP results on 'different' orientations: we called it multi-orientation-RESP fit by analogy to the multi-conformational RESP fit, see below.

2. Clearly, AMBER 1995 partial charges were produced on a specific orientation. Do you think which choice is better for MD simulations: a "new RESP fitting procedure" only on a part of a molecule OR consistency with current AMBER 1995 partial charges for the whole molecule? We know that AMBER and CHARMM have quite different partial charges, but both perform well.

As I already said, the question is: Do you want that people that will read your works can REPRODUCE your published charges ? If "yes", use RED with the rigid body algo. if "not" you can still use RED however without using the rigid-body algo. (keeping the orientation selected by the ab initio software).

On the contrary, we do NOT question the VALIDITY of such charges in AMBER and CHARMM. As you said, it is known they perform well. Reproducibility and validity are not the same thing...


3. Does the orientation "inaccuracy" really matter in MD simulations? How much is the "inaccuracy" caused by software updates and machine differences?

See RED manual pages 3-4.

As example, if you take the dimethylphosphate example the charge incertainty (because of the orientation of the 2 methyl groups) is 0.016. It is unlikely that such charge incertainty will affect MD simulation. Now, if you build a box of solvent using EtOH or DMSO and you have a charge incertainty of .046, this starts to be more important. I do not have the answer for this particular case. Thus, using RED you can apply a multi-re-orientation RESP fit that allows to average the differences in charge observed for one orientation over several orientations.

With RED version 2.0, the user will be able to apply multi-orientation and/or multiconformation RESP fit. This will allow to associate the 'validity' of the RESP charges (using multi-conformational fit) and the 'reproducibility' of the RESP charges (using multi-orientational fit).



20) Re: AMBER: esp or resp for all_amino94.lib

From: FyD <fyd@u-picardie.fr>
Date: at May 01 2004 - 00:28:57 BST


I am wondering that what kind of charges are for all_amino94.lib in Amber7.0. Are they Esp or Resp?

RESP charges

I first want to reproduce the charges for Arg with same geometry in all_amino94.lib and calculate the charges for neutral Arg.
When I calculate charges using same geometry and method, am I supposed to have same charges for each residue as those in all_amino94.lib, right?


See:
http://amber.scripps.edu/Questions/mail/316.html
http://amber.scripps.edu/Questions/mail/317.html
http://amber.scripps.edu/Questions/mail/318.html
http://amber.scripps.edu/Questions/mail/319.html



21) Re: AMBER: resp demo files

From: FyD <fyd@u-picardie.fr>
Date: Mon May 03 2004 - 22:40:15 BST


I am attempting to perform a multi-conformation charge fit using resp. The resp manual says to see /amber5/demo/resp_charge_fit for examples.

If you are interested in RESP tutorials only for "multi-conformational" RESP fit, you can use RED _as you have it_ with "multi-orientational" RESP fit. Multi-orientational or multi-conformational use the same types of RESP inputs...

In your case, the RESP tutorial will be useful only for inter-molecule restraints.

However, I don't have access to amber and have not been able to find files on the amber webpage. Does anyone have access to these files or could point me to their location?

I think they are only available within AMBER 5/6/7/8.

PS: We have a beta-version of RED version 2.0 with multi-conformational and/or multi-orientational RESP fit. I can send it to you if you are interested...



22) Re: AMBER: reproduction of RESP charges for Gly

From: FyD <fyd@u-picardie.fr>
Date: Tue May 04 2004 - 21:16:00 BST


I found that charges for Gly in all_amino94.lib (Amber7.0) are same as charges in Ref. P.Cieplak et al, JCP, 16, 1357 (1995). I tried to reproduce RESP charges for Gly using R.E.D and antechamber program (in version 7). First of all, I used the coordinates for Gly from all_amino94.lib and did single point energy calculation using Gaussian98. (I think all geometries in all_amino94.lib are already optimized)

What I would do is to use the starting coordinates you got from ... as a starting point for minimization and NOT for single point (SP). Indeed, I see at least two reasons why you could observed charge differences with the values available in the xleap OFF libraries...

I wonder the number of digits after the decimal point you used in your SP input: if you started from a PDB file (that has only three digits) some small charge diff. may be observed; It is why the PDB files generated by R.E.D. have four digits (The Tripos format generated by RED version 2.0 will present 6 digits)...

I guess you generated a "Standard orientation" within your SP (I mean in the Gaussian output). However, it is unlikely (depending on your model) that you will get the same Standard orientation than the one generated when RESP charges where calculated in 1994-1995. Thus, if 'your' minimum molecular orientation used to compute the MEP is different from the one calculated in 1994-1995, the RESP charges will be different. The charge differences induced by 2 different molecular orientations are un-predictable and can display some noticeable (and highly disturbing) differences... It is the main reason why we wrote R.E.D.


After this I got charges for Gly but they are pretty different from charges in all_amino94.lib.
Why are they so different? They used two Gly conformations for RESP charge-fitting in reference. What does this mean?


See above

That means I can't use the coordinates for Gly in all_amino94.lib to reproduce charges?

This simply means that it is very unlikely you will be able to reproduce the same RESP charge values than those published in 1995. The differences observed might or might not be disturbing (depending on the model and/or orientation).

Second problem is that I also tried to use R.E.D to get resp charges. When I used gaussian output after single point energy calculation it gives "Invalid Minimization OUTPUT !" error message.

You need to use a GAMESS or Gaussian MINIMIZATION output as input in R.E.D. If you load a SP output, you will get such messages... It is explained in the R.E.D. manual.

Can't R.E.D read gaussian output from single point energy calculation? How can I modify the code in this case? Is there any clear example to reproduce RESP charges? Are there any source of geometry and method to reprocude RESP charges?

So far, the ONLY way to derive reproducible RESP charges in to use the rigid-body re-orientation algo. available in R.E.D. and to publish the RESP charges with the orientation(s) selected. You can also apply multi-orientation RESP fit to average the charge diff. observed for one orientation over several orientations... In this case, the charges are reproducible whatever the ab initio software used, i.e. using GAMESS or Gaussian...

PS: I think you can found similar explanations from the links I provided you in your previous email.
http://structbio.vanderbilt.edu/archives/amber-archive/2004/1359.phtml
See also the R.E.D. manual...



23) Re: AMBER: reproduction of RESP charges for Gly

From: Bill Ross <ross@cgl.ucsf.edu>
Date: Tue May 04 2004 - 21:35:49 BST


They used two Gly conformations for RESP charge-fitting in reference. What does this mean?

Best to read the papers involved. Also, conformation is not the same as orientation.

my understanding of which is that charges were fitted on ESP's derived from 2 representative conformations to make them more general.



24) Re: AMBER: reproduction of RESP charges for Gly

From: FyD <fyd@u-picardie.fr>
Date: Tue May 04 2004 - 21:48:43 BST


They used two Gly conformations for RESP charge-fitting in reference. What does this mean?

ups, sorry... I did not read carefully this part of your email... If different conformations are used, this makes the problem even more complex because different orientations can be selected for different conformations... (With R.E.D. version 2.0, this will be automatically handled...)

Best to read the papers involved.
my understanding of which is that charges were fitted on ESP's derived from 2 representative conformations to make them more general.


See Reynolds et al. JACS 1992, 114, 9075.

Now what are the charge differences you observed ? If they are above 0.1, the problem could come from the RESP inputs themself...



25) Re: AMBER: reorientation in R.E.D

From: FyD <fyd@u-picardie.fr>
Date: Wed May 05 2004 - 18:55:51 BST


There are 6 different orientations for N-Methyl-Acetamide example in R.E.D(1.0) tutorial. How can I get this reorientation information (6 different number sets)?

You are right that is a first key point.

If you decide to choose the ab-initio re-orientation algo. [i.e. by placing the center of nuclear charge at the origin (available in Gaussian) or using the molecular principal axes (available in GAMESS)] you have at least two problems:

- You can get different minimum molecular orientations generated by the 2 ab-initio software => the charges derived using GAMESS and Gaussian (geometry optimization & MEP computation) will be different (moreover, the values of such differences are un-predictable). That is the first problem.

- Even using one particular software (i.e. GAMESS 'OR' Gaussian, this time), a constant molecular orientation for the same minimum cannot be generated whatever the model studied is. If we take the example of Gaussian (because it is mainly used by the AMBER community), different "Standard orientations" can be generated by Gaussian for the same minimum: Thus, different set of charges will be generated for the same minimum. Once again, the charge diff. are not predictable. That is the second problem.

- Thus, we implemented in prog. R.E.D. a rigid body re-orientation algo. based on three atoms that is applied after geometry optimization. If this re-orientation algo. scheme is selected (by providing 3 'well defined' atoms in the starting PDB file), the charges will be reproducible because the minimum molecular orientation is known whatever the ab-initio software is (accuracy reached = 0.0001 e; i.e. the format of the charge values reported in a ".prep" or "OFF" files for the Cornell et al. FF). That is the main idea.


How can I know which molecule needs reorientation or not?

Once again, that is an important point, you are right.

- The first idea is that charges which are published must be reproducible whatever the ab-initio prog. is. This is achieved when the rigid-body re-orientation algo. is applied. Thus, people have to publish the molecular orientation used in the RESP or ESP fit WITH the set of RESP charges.

- The second idea is the multi-orientational RESP fit. The question you asked is more or less why selecting one orientation and not another one ? If quite different atomic charges are observed for two different minimum molecular orientation, we decided to incorporate a multi-orientational RESP (by analogy to the multi-conformational RESP fit) fit allowing to average the charge difference observed for one orientation over several orientations.

Thus, in the R.E.D. tutorial, you mentionned that a 6 re-orientation RESP fit is performed on NMA. It was just an example of multi-orientational RESP fit. Once again, the charges are reproducible because the 6 orientations used in the RESP fit have to be provided.

Thus, concerning which orientation(s) to select: it is up to the user which has to report/explain the molecular orientation(s) used in the RESP fit.



26) Re: AMBER: choosing 3 atoms for reorientaion in R.E.D

From: FyD <fyd@u-picardie.fr>
Date: Thu May 06 2004 - 23:00:44 BST


I tested charge reproducibility using different sets of Cartesian coordinates with your EtOH example (with 4 reorientations for EtOH). different sets of Cart. coord. as I expected.

So it works, good for us ;-)

I have questions as following:
1. How did you select 3 atoms and 4 reorientations (ex.EtOH) for the rigid-body re-orientation algo. in R.E.D?


We got the charges for each orientation. As they were different we decided to apply 4-orientation RESP fit. That's it...

What is criteria to choose these 3 'well-defined' atoms?

- Scientifically ? No-one !

- Convenient way: use only heavy atoms selected more or less randomly...

=> Shocking ? well yes and no ;-) because with Gaussian/GAMESS one believes that the minimum orientation is unique and it is wrong !!! Do not forget that RED has been written to allow the derivation of "reproducible" charges.


I think that choosing orientations is really important to reproduce charges.

It is why we wrote RED to control the minimm molecular orientation...

2. If I chose 3 well-defined atoms and reorientations for already published molecules (ex. Cornell et al, 1995) and run using the rigid-body reorientation algo. in R.E.D, would I have reproducible charges for published > molecules?

Your charges will be different from the ones available in the Cornell et al. paper, but they will be reproducible (if you provide the orientation(s) you used) and the Cornell et al ones are not ....

3. Based on your last reply, multi-orientational RESP fit would give more general RESP values....

Yes, at least it make sens. However, it has not been tested.

That means that it is better to use as many orientations as possible? How many orientations?

Now in your two last questions you mixed up two thinks: "reproducibility" and "effectiveness". It simply makes sens to use multi-orientation RESP fit if charge differences are observed for different orientations. We have examples using up to 12 orientations. It is up to the user... Do not mix up "reproducibility" and "effectiveness". "Effectiveness" is obained with multi-conformational RESP fit... In the new RED version we mix multi-orientational and multiconformation RESP fit to associate charge "reproducibility" and charge "effectiveness".


27) Re: AMBER: charges (Cornell et al JACS, 1995) and R.E.D

From: FyD <fyd@u-picardie.fr>
Date: Wed May 12 2004 - 17:42:31 BST


I am trying to calculate the charges for neutral Arg. In Cornell et al (JACS, 1995) and Cieplak et al (J.Comp.Chem.1995), the scheme for fitting the non-terminal amino acides was used with blocking groups (ACE, NME). The following are my questions according to these two articles.

1. Do I need to add two blocking groups to optimize neutral Arg using Gaussian(6-31G*)? i.e. ACE-AA-NME in paper

If you want to follow the strategy of Cieplak et al., I guess the answer is 'YES' (after reading the pages 1370-1373)... However, it is ONLY true if your residue is NOT N-term or C-term, i.e CENTRAL. See below.

2. For charge fit, the charges of C=O (C:0.5973, O:-0.5679) or N-H (N:-0.4157, H: 0.2719) groups in the blocking groups have identical C=O and N-H charges for adjacent residues in the protein or peptide on pg.1371, Cieplak et al (J.Comp.Chem. 1995).

Yes, all the residues in the file "all_amino94.in" (CENTRAL AMINO-ACIDS) have the rules you said (just above). It is cleary explained why in the Cieplak et al paper: because restraints (with ACE and NME blocking group) were used...

However, C=O and N-H charges are not all the above charges for all amino acids in Cornel et al (JACS, 1995). Why do some amino acids have identical N-H and C=O charges to the ones in blocking groups and some don't?

Yes, the charges of N, H, C and O are different in the the 'all_aminont94.in' and 'all_aminoct94.in' because the way to get the charges for terminal residues (N-term and C-term) is different than the one used for CENTRAL amino-acids. Read pages 1375-1376.

3. In R.E.D, can we apply specific charge constranits? i.e. accorindg to paper, it seems to put specific C=O and N-H charges and do charge fitting for rest atoms.

Directly ? No. Here is what I do to apply intra-molecular and/or inter-molecular restraints:

- Run R.E.D. controlling (or not) the minimum molecular orientation using one or several conformations (for multi-conformational RESP fit, I can provide you the beta version of R.E.D. version 2) without restraints.

=> You get two RESP inputs for the 2 RESP stages: input1 & input2

- In a new directory, modify these inputs manually adding the restraints (intra-molecular and or inter-molecular) manually (see AMBER manual)

- Re-run RESP without R.E.D. this time using the command provided in the R.E.D. manual see page: (end of page 12). You can create the two following alias in your $home/.cshrc:

alias resp1 'resp -O -i input1 -e espot -o output1 -p punch1 -q qout -t qout1 -w qwts -s esout'
alias resp2 'resp -O -i input2 -e espot -o output2 -p punch2 -q qout1 -t qout2 -w qwts -s esout'


Thus, each time you modify your RESP input1 & 2 to introduce new restraints, you can execute the 'resp1' and 'resp2' commands... This allows to do a lot of tests to see if the restraints provided in the input are 'really' taken into account.

I hope my questions are not complicated. If anybody understand how to calculate AMBER charges (Cornell et al,JACS, 1995) for amino acids, could you help me out?

I would do:

- Select different conformations for your ARG (see the table page 1362 for which one you may need)

- Run R.E.D. version II to apply multi-conformation RESP fit (with the X conformations you selected) controlling the minimum molecular orientation (if you want to publish the charges: re-producibility)

- Add the restraints you need (i.e. it depends if your ARG is N-term/C-term or CENTRAL)

- and re-run RESP after modification of their inputs with the 2 'alias' provided.

There are a looooot of informations in the Cieplak et al paper. The difficulty is to not mix them...



28) RE: AMBER: charges (Cornell et al JACS, 1995) and R.E.D

From: FyD <fyd@u-picardie.fr>
Date: Wed May 12 2004 - 19:01:31 BST


I understood the charges of N, H, C and O are different for terminal and central amino acids. Look at Cornell et al (JACS, 1995) pg.5190-5191. All amino acis in Figure 4 are central(non-terminal) but they don't have same charges as the ones in blocking groups (ACE, NME) for ARG, PRO, LYS. Could you explain why they are different?

It looks like there are three types of CENTRAL amino-acids (AA): the neutral AA, the acidic AA & the basic AA. I guess, it means different strategies were used for these 3 types of AA: It makes sense.

See also amberX/examples/resp_charge_fit/peptoid the example described by Piotr Cieplak: As you can see the restaints are in agreement with the NEUTRAL (& CENTRAL) AA...


If I well understood, you want to derive RESP charges for neutral ARG. You are lucky, the way to derive the RESP charges of neutral AA is the way described in the Cieplak et al. paper ;-) See page 1374-1375, only neutral amino-acids are reported.


29) Re: AMBER: charges (Cornell et al JACS, 1995) and R.E.D

From: Piotr Cieplak <cieplak@cgl.ucsf.edu>
Date: Wed May 12 2004 - 19:57:20 BST


Yes, Francois is right. During resp fitting for AA we had three groups of AA - neutral, positively and negatively charged. Each of them was fitted separately. Piotr


30) Re: AMBER: parametrization partial charges

From: FyD <fyd@u-picardie.fr>
Date: Mon May 17 2004 - 16:37:02 BST


I am trying to reassign partial charges to a non-standar TRP. My model is ACE-TRP-NMA. Once I have the results I would change only the charges of the non-backbone atoms. I suppose that the group I want to change should have a total charge equal(but with opposite sign) that the rest of the TRP. What happend if this group that not satisfy this condition? What should I do to make this new aminoacid be neutral?

Highly similar questions (about neutral Arg) & answer can be found at:
http://structbio.vanderbilt.edu/archives/amber-archive/2004/1516.phtml
http://structbio.vanderbilt.edu/archives/amber-archive/2004/1517.phtml
http://structbio.vanderbilt.edu/archives/amber-archive/2004/1518.phtml

See also: Cieplak et al. J. Comput. Chem. 1995 16 1357-1377

And also:
http://amber.scripps.edu/Questions/mail/316.html
http://amber.scripps.edu/Questions/mail/317.html
http://amber.scripps.edu/Questions/mail/318.html
http://amber.scripps.edu/Questions/mail/319.html



31) Re: AMBER: Bug in babel conversion

From: FyD <fyd@u-picardie.fr>
Date: Tue May 25 2004 - 19:14:55 BST


While using the conversion command described in RED manual, pp 15, I found a bug. The following command converts a pdb file not having connection information to a pdb file that will have the connection informations.

There are a looot of bugs in Babel, in particular if you are interested in studying "sp" atoms... So I would always check the conections and Cart. coordinates generated to see if they match what you wanted...


32) Re: AMBER: pdb in RED

From: FyD <fyd@u-picardie.fr>
Date: Thu May 27 2004 - 20:16:59 BST


I am reading the manual and there are two examples with different definitions for hydrogen atoms. My questions is : should and hydrogen have the same number as the atom that it is linked, if that atom is an heteroatom or a carbon (non methyl, or methylene)?

Yes, if you want to follow the published way to derive RESP charges:
- Equivalencing of CH2 & CH3 group are carried out in the 2nd RESP stage
- All other types of equivalence implying hydrogens are carried out in the 1st RESP stage.

The automatic generation of RESP inputs in R.E.D. is based on the PDB atom names.

Example: CH3-NH2
CH3 equivalenced in the 2nd RESP input
NH2 equivalenced in the 1st RESP input

=> Atom names for methylamine:
CH3 = CT1 H1 H1 H1, "T" allows equivalencing in the 2nd RESP input.
NH2 = N2 H2 H2, no "T" => equivalencing in the 1st RESP input.
or
CH3 = CT999 H999 H999 H999
NH2 = N2 H2 H2

In the second R.E.D. version 2.0, RESP input generation has been updated to make this more convenient and in particular to allow the generation of non-standard RESP inputs...



33) Re: AMBER: Use of RED

From: FyD <fyd@u-picardie.fr>
Date: Fri May 28 2004 - 21:20:42 BST


I write because I am trying to reproduce the charges of the aminoacids that were calculated with RESP using ACE and NMA as blokcing groups. When using the RED program I need the following constraints:

- ACE, NMA and the aminoacid must be neutral
- The C=O of the aminoacid should have the same charge as that of ACE.

Moreover that charge must be set to 0.5973 for C, and -0.5679 for O. The same criteria should be applied to the N-H of the aminoacid and that of NMA.

I have read the RED manual but it does not have information about doing this. I think that I could solve the problem using directly RESP, but this is probable more difficult. I am reading the file resp.doc but it has a lot of variables I don´t understand. Please if you can tell me what to do, I will be very grateful.


You cannot do it directly, you have to manually modify the inputs generated by R.E.D. (in a second time) after running R.E.D. without restraint:
See
http://structbio.vanderbilt.edu/archives/amber-archive/2004/1516.phtml

PS: I will summarize the FAQ related to R.E.D. on the R.E.D. web site



34) Re: AMBER: Loading mol2 Files in XLEaP

From: FyD <fyd@u-picardie.fr>
Date: Thu Jun 03 2004 - 17:26:38 BST


Hi! I'm using Amber 8 on an SGI machine running Irix 6.5, and I've been having a bit of trouble with xLEaP regarding the "loadmol2" command mentioned in the manual. I've tried to load a file by entering the following commands:
variable = loadmol2 "test.mol2"
variable = loadmol2 test.mol2
loadmol2 test.mol2


With each of them I get "syntax error" back from the program, and the file fails to load. The file is located in the same directory I'm running the program from. Also, if it makes any difference, the file is a canonical B-conformation T:A 15-residue oligonucleotide (DNA). Has anyone else had this problem? If so, how did you fix it? Any help would be appreciated.

It worked for me:
M = loadmol2 M.mol2

However, the section "@<TRIPOS>SUBSTRUCTURE" must be present in the Tripos format. See http://www.tripos.com/custResources/mol2Files/


For instance Babel does generate .mol2 file without this section. If a Tripos file is loaded without this section the unit created by leap stays empty...


35) Re: AMBER: unnatural base - RESP charges in ANTECHAMBER

From: FyD <fyd@u-picardie.fr>
Date: Tue Jun 08 2004 - 17:23:46 BST


We'd like to compute RESP charges for an unnatural nucleoside. We know how it works in principle -- the joint between the nucleoside and the phosphate -- but we don't know how to setup the input for ANTECHAMBER and there does not seem to be an example. Reading the code to figure out the input would be an option; but it would take a while and I'd rather not go that way. So, if there is somebody with such an input, we would like to have that template. That would really help us.

There were already some discussions about this in the AMBER mailing list. See:
http://amber.scripps.edu/Questions/mail/318.html
http://amber.scripps.edu/Questions/mail/319.html

You have examples of RESP inputs (with restraints) in: amberX/examples/resp_charge_fit


36) Re: AMBER: Problems with X.R.E.D.

From: FyD <fyd@u-picardie.fr>
Date: Fri Jun 25 2004 - 18:20:30 BST


I am new to working with R.E.D. and I am looking for some assistance from the group. I recently set up a simulation to run the sequence ACE LEU NME using six orientations on R.E.D (using the GAMESS package). I was looking for resp charges and used the pdb minimization options. Everything seemed to go well however it never finished the minimization of the pdb file. I finally terminated the project after seven days after giving up hope that it would finish minimization (the log file was appended to every few minutes and grew to the size of 5 Mb at the end of the seven days). Has anyone had any experience like this before? If this is normal, how can I tell when the minimization should be completing?

If I understand you, your problem is not with using X R.E.D. but using GAMESS... Several points:

- When you have a big optimization job, I think it is better to run minimization without R.E.D. and once you got the minimum (using different minimization input tests), to run R.E.D. with the following variables:
$OPT_Calc = "Off";
$MEPCHR_Calc = "On";
$MOL_START = "./Your_PDBFile.pdb";
$JOB_OPT = "./Your_Path/Your_GAMESSminimization_out.log";


- I am not surprised it takes so long because we strongly decreased the minimization thresholds in GAMESS minimization in order to get similar results than Gaussian i.e. we used the following keywords:

$CONTRL INTTYP=HONDO QMTTOL=1.0E-08 ITOL=30 ICUT=20 $END
$SCF CONV=1.0E-08 $END
$STATPT OPTTOL=1.0E-06 $END


Thus it may not be very 'convenient' in particular if you do not use a cluster of computers.

- Now if I look at your GAMESS ouput, you get:
MAXIMUM GRADIENT = 0.0000827 RMS GRADIENT = 0.0000318
This means your structure is ~ correct but its accuracy is not enough to insure reproducible charges.

- What can you do? I would try 2 things:

* 1st starts from the structure which presents the MAXIMUM GRADIENT = 0.0000827 and continue the minimization but calculating its hessians. See manual RED.pdf page 12 using: the $STATPT and $FORCE groups calculating the hessian every 10 structures (for instance). The Hessian calculation will take time but should allows you to reach the minimum in less minimization steps.

* If you still cannot get the minimum, you could start to decrease the GAMESS minimization thresholds i.e. remove from the minimization input:

INTTYP=HONDO QMTTOL=1.0E-08 ITOL=30 ICUT=20 and if it is still not enough use: OPTTOL=1.0E-05

However, I would not recommand to decrease the minimization thresholds (in particular I would keep OPTTOL=1.0E-06) because your charges might not be reproducible with this new input (OPTTOL=1.0E-05)...



37) Re: AMBER: R.E.D. mol2 file and xleap

From: FyD <fyd@u-picardie.fr>
Date: Wed Jul 14 2004 - 20:30:04 BST


I am going through the plastocynin tutorial. Can you tell me how I can use GAFF of antechamber for this purpose, since i could not figure out how to use the resp charges generated by RED as there is no relevent file type which can be used.

R.E.D. generates a Tripos "Mol_o.mol2" file with the RESP/ESP charges in it.

To load this file into leap/amber8 (only), run:
xleap -f leaprc.ff99
M = loadmol2 Mol_o.mol2
saveamberparm M M.top M.crd


Before be sure to manually add the atom types needed according to the Cornell et al. paper, (JACS 117 5179. See Table 1 "List of atom types"), i.e. modify the 6th column of this .mol2 file: cf http://www.tripos.com/custResources/mol2Files/

Personnally, I prefer to "manually" add the atom types because this step is crucial and determines your force field accuracy. As said Ross, you have to be aware of what you do...

Using "saveamberparm M M.top M.crd" leap (1st run), it is going to tell you, which force field parameters are missing. Then, you can create your personnal FRCMOD file (using GAFF or by fitting ab-initio data) which contains 'your' NEW FF params that can be load into leap (in a 2nd run) i.e. using:

xleap -f leaprc.ff99
M = loadmol2 Mol_o.mol2
FRCMOD = loadamberparam frcmod.Anshul
saveamberparm M M.top M.crd



38) Re: R.E.D. and network (in French)

From: FyD <fyd@u-picardie.fr>
Date: Wed, 21 Jul 2004 19:30:40 +0200


Mais je n'ai pas accès à gaussian sur une machine locale et la seule manière de lancer gaussian est de passer par le CINES. c'est pourquoi je ne peux utiliser XRED ou RED

- Si vous pouvez vous connecter par 'ftp' et 'telnet' vous pouvez mettre une version de R.E.D. sur votre machine distante et l'executer par telnet, c.a.d. executer qqchose du genre:
nohup perl RED.pl > RED.log &

- Si vous pouvez vous connecter par 'ssh' et 'sftp' vous pouvez alors executer R.E.D. et X RED. Cependant X RED est utile principalement pour des calcules sur une machine locale....

- Si la machine du CINES est un cluster avec gestionnaire de queue: vous pouvez inclure la command "perl RED.pl > RED.log" dans un script pour PBS; c'est ce que je fais ici tres souvent...

- Si vous voulez lancer un calcul avec Gaussian sur pls processeurs, utilisez RED-SMP.pl au lieu de RED.pl en indiquant le nbre de proc. a utiliser. Cela veut dire que RESP et Gaussian doivent etre compiles sur la machine du CINES. On peut tres bien imaginer que vous utilisiez un 'Gaussian' compile pour tout le monde dans /usr/local/gaussian sur cette machine distante et un 'RESP' que vous compilez dans 'votre' login pour vous. RESP est fourni separement vous n'avez pas besoin d'avoir AMBER... cf http://amber.scripps.edu/Questions/resp.html

J'aimerais convertir mes fichiers output de Gaussian en fichiers input de RESP:

Pour des convertions manuelles, cf:
http://amber.scripps.edu/Questions/resp.html et http://amber.scripps.edu/Questions/resp.txt



39) Re: [Gamess] failure to generate RESP charges

From: FyD <fyd@u-picardie.fr>
Date: Mon, 12 Jul 2004 19:20:45 +0200


Meanwhile I tried to run RED with semiimperical basis set in GAMESS.

My personal opinion is that it is a bad idea...

It worked out preaty fast but it took 725 steps to converge.

It means your starting structure was bad. It might be the reason why your 6-31G* optimization takes so much time...

I would like to know if such a large calculation is normal or is there something wrong with my system and how can I check it. Next MEP was calculated successfully but it failed to calculate the final RESP charges. the punch file which was generated is as follows: Statistics of the fitting:

The initial sum of squares (ssvpot) 0.000
The residual sum of squares (chipot) 0.000
The std err of estimate (sqrt(chipot/N)) NaN
ESP relative RMS (SQRT(chipot/ssvpot)) NaN

Dipole Moment (Debye)= 0.00000


"NaN" means 'Not a Number'. Something did not work... I think R.E.D. did not find the GAMESS '.dat' file to transform the GAMESS MEP ouput format into RESP format. R.E.D. get this MEP .dat from the GAMESS scratch directory. See the 'rungms' GAMESS script:
set SCR=/disk2/QM_SCR # ~ line 33 of GAMESS rungms '/disk2/QM_SCR' being my scratch directory.

We modified 'rungms' in a way that GAMESS save its MEP .dat file (or punch file) in his scratch $SCR directory.
Thus, R.E.D. finds, converts and moves it in the directory where you decided to save your data.
See rungms ~ lines 73-80:

# file assignments.
# all binary files should be put on a node's local disk ($SCR directory),
# both ASCII punch files might be written by NFS to user's permanent disk
set echo
setenv EXTBAS /dev/null
setenv IRCDATA $SCR/$JOB.irc
setenv INPUT $SCR/$JOB.F05
setenv PUNCH $SCR/$JOB.dat
setenv AOINTS $SCR/$JOB.F08
setenv MOINTS $SCR/$JOB.F09
setenv DICTNRY $SCR/$JOB.F10
setenv DRTFILE $SCR/$JOB.F11



40) Re: AMBER: charge not zero

From: FyD (fyd_at_u-picardie.fr)
Date: Wed Aug 04 2004 - 20:57:51 BST


1. What i want to do is that, I have to replace some of the aminoacids from a protein with another aminoacid which is attached with an organic molecule (lets name this unit as x). Now I prepare the prepin file for this unit (x). the end group in this unit (end of the aminoacid part) is protectd with NHMe and ACE. now when i replace the original amino acid with this unit, what should i do to the end group. Should it be removed and if yes then do i need to make some adjustments in the charges after its removal.

If I understand you, you have a new residue 'AAX'. In your protein, it should have the following form -NHCH"X"CO-, if it is a 'central' AA i. e. AAn-NHCH"X"CO-AAn+2. To get the charges for such systems, you could use the capped model below as it is described in the Ciepak et al. J. Comput. Chem. 1995, 16, 1357-1377:
CH3CO-NHCH"X"CO-NHMe.
It is better to select phi, psi dihedrals describing 2 or 3 different conformations. For instance, one in a canonical alpha helix, and another one in a canonical beta-sheet. See once again the Cieplak et al. paper and also http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#27

I guess it is better to apply multi-conformational and multi-orientation RESP fit in this case. I can provide you R.E.D.-II to do this if you want...
- You select 2/3 conformations based on the Cieplak et al. paper, or based on some ideas you have about your system, and let's say you select 4 orientations. You run R.E.D.-II and you get the charges for this CH3CO-NHCH"X"CO-NHMe system...
- You copy the 'espot' (contains the espot$i*$j, $i = conformations, $j = orientations), 'input1' and 'input2' files obtained in a new directory.
- You add restaints to put to total charge of your CH3CO- and NHMe- groups = zero in the 1st RESP input and you freeze them in the 2nd RESP input (i. e. = -1).
See http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#28 and
amberX/examples/resp_charge_fit/peptoid i. e.
- You modify manually the 'input1' and 'input2' and you run the 2 aliases provided at: http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#27
- You have your charges for your central AAX residue i. e. -NHCH"X"CO- in the new punch2 file generated after running RESP manually... The total charge for -NHCH"X"CO- is then obviously = zero...
- You create a Sybyl file for -NHCH"X"CO- (using 'insightII' for instance). The conformation and orientation of this AAX are NOT important and do NOT matter at this stage. ONLY the RESP charges describing different conformations & orientations matter...
- You add manually (text editor or in insightII) the RESP charges (and atom types describing your FF) in this file selecting only the RESP charges of -NHCH"X"CO- from the new punch2 (i. e. the charges of CH3CO- and -NHMe are not used anymore). See http://www.tripos.com/custResources/mol2Files/ to known which columns to modify...
With R.E.D.-III, the FF atom types will be automatically recognized and added in the Tripos format...
- You load this .mol2 file in amber8/xleap

M = loadmol2 AAX.mol2
saveoff AAX.off


You have a new 'AAX.off' library which describes your new central AAX residue ready to be loaded with a 'loadoff AAX.off' command in a leaprc.ff99 file when you will need to load your whole protein-PDB file containing this new AAX....

2. Should my end group contain the end group or not?
3. Presently I am replacing the aminoacid in InsightII. I manually remove the orignal amino acid and then form a bond with this unit. I would like to know if there is any other simpler method to do so.


I use also InsightII to build starting structures and switch to xleap/sander to recognize/run simulations...
In InsightII, 2 advice:
-1 Save your PDB/mol2 files without the hydrogens (if possible), 'xleap' will add them automatically. The atom name of the hydrogens generated by InsightII are often not recognized by 'xleap'...
-2 Be sure that the 'atom name' and 'residue name' of each AA (not only AAX) in the file generated by InsightII match the corresponding 'names' in the xleap .off libraries...
And it will work...



41) Re: AMBER: RESP partial charge calculation

From: FyD (fyd_at_u-picardie.fr)
Date: Tue Aug 10 2004 - 18:46:16 BST


I am trying to calculate the RESP charge for an heme group with Fe+3. I would like to know which is the number I have to consider for Fe+3 in the RESP input file. I have used a complete basis set in the ESP calculation, so I have considered all the electrons for Fe, but I do not know if the number in RESP input must be 26, that is the Fe atomic number or 23, that is the actual total number of electrons in Fe+3. The same happends with other ions, like the O in a negative charged carboxilate group -C=O. Have I put 9 or 8 corresponding to the O atom in the RESP input file?

See below

RESP assumes the ESP file is from an ALL electron calculation. I don't remember any 6-31G* all electron basis set for Fe so maybe you are using a pseudo potential for the Fe? If so, you must use the actual number of electrons on the Fe atom instead of the atomic number in the RESP input.

The 6-31G* basis set is available for up to Z = 36 in GAMESS/Gaussian
See V.A.Rassolov, J.A.Pople, M.A.Ratner, T.L.Windus J. Chem. Phys. 1998, 109, 1223-1229
or at least up to Z = 25... I was also surprised to find such basis set available in GAMESS/Gaussian for iron when we wrote R.E.D.-II... I guess it is because this work was published in 1998...


I am trying to calculate RESP partial charges for a heme-Fe+3 system. I have previous calculate de electrostatic potencial using gaussian: BLYP/6-31G* scf=(tight,maxcyc=500) pop=(mk,readRadii) # IOP(6/33=2)
The readRadii item was added to include the Fe radii of 0.49 since it doesn't appear in the default library.


Yes, you have to do by this way for Gaussian. On the contrary, GAMESS seems to use 1.8 as default value for such elements. See below... But I am surprised by your value of 0.49... Is it not too small ? I would use at least 1.8 as in GAMESS... and see also http://signe.teokem.lu.se/%7Eulf/Methods/resp.html

After that I have built the correspondings input and espot files to run the RESP program. I have otained very high charges for certain C and H! Charges like -15 aren't reasonalby..., and in atoms far away from the Fe! Could anybody help me to know if there is something wrong in using the RESP program with a system that has an Fe atom?

Here are some information for your problem:
- GAMESS handles atom radii as follows:
Z = 1-17 the radii of Kollman/Singh/Merz
Z = 19-35 the radii seem to be fixed to 1.8 and from what I understand if you want to change this '1.8' value you have to modify the file 'prplib.src' in the GAMESS source code and recompile.
- Gaussian handles atom radii as follows:
Z = 1-17 the radii of Kollman/Singh/Merz
Z = 19-34 the radii are unknown and you have to provide the atom radii as you did...
Z = 35 radii = 2.3, where does it come from ?

In R.E.D.-II, RESP charges for elements up to Z = 35 are now automatically handled. It is coded in a simple way: We decided to apply the GAMESS way for the elements Z = 19-35 for Gaussian by default. However, it has been written in a way that can be simply modified by replacing 1.8 by your own value...

If you want I can provide you R.E.D. II, otherwise it is going to be released beginning of September... With R.E.D. II, you will be also able to associate multi-conformation and/or multi-orientations for your heme molecule, to use Gaussian optimization in GAMESS MEP computation (and vis-versa) and you can also automatically generates 'non-standard' RESP inputs and then performed non-standard RESP fits...



42) Re: AMBER: RED II and single point

From: FyD (fyd_at_u-picardie.fr)
Date: Tue Aug 10 2004 - 19:35:05 BST


thanks for your answer. I will follow the steps that are indicated in the page you sent me (http://signe.teokem.lu.se/%7Eulf/Methods/resp.html). We have minimized our system with another program because it was too high for Gaussian. I have done the single point calculation in Gaussian as it is indicated.

OK If you have a single point, you can 'force' R.E.D.-I or R.E.D.-II to read GAMESS/Gaussian SP as an OPT output. Here is what to do for Gaussian:
You simply add the words "Stationary point found" just before the "Standard orientation" available in your single point calculation, and R.E.D. will consider this SP as an OPT... Then, you can load this 'cheated' OPT output in R.E.D. as an optimization output using the following keywords:
$OPT_Calc = "off";
$MEPCHR_Calc = "on";
$MOL_START = "./Heme.pdb"; # Starting PDB File (un-optimized structure)
$JOB_OPT = "./JOB1-gau.log"; # Minimization OUTPUT if $OPT_Calc = "OFF"



43) Re: AMBER: RESP charge derivation

From: FyD (fyd_at_u-picardie.fr)
Date: Wed Aug 11 2004 - 17:55:57 BST


I have derived RESP charges for some base analogues. I used a methyl instead of sugar for the Gaussian calculations (# HF/6-31G* OPT POP=MK IOP(6/33=2) FChk=All Iop(6/33=2)). As a control a derived charges for methyl-G and compare them with the Amber charges for G. I got the results below. There are some differences. The question is, given the differences that I observed between the Amber charges of G and the calculated charges of m-G can I rely that the my charges are good?? The RESP was performed in 2 stages as described for Amber.

From our tests, RESP atomic charges can present charge difference up to 0.07 e if you generate the appropriate inputs. Now, I would not be surprised to see charge differences up to 0.1 e depending on... so much parameters...
I would say that charge differences bigger than 0.1 e results from bad inputs, or the use of a different minimum to compute the MEP.
I would obviously suggest you to use R.E.D. with several orientations as I guess only a unique conformation is possible with your methyl-base...
You can find a lot of information at http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm

If you want to use R.E.D. for your methyl derivative, run first R.E.D. Then copy the input1 & 2 and espot file in a new directory and add the restaint charges for your methyl group in the input1. See http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#28 and http://structbio.vanderbilt.edu/archives/amber-archive/2004/2696.phtml.



44) Re: AMBER: Gaussian in RED

From: FyD (fyd_at_u-picardie.fr)
Date: Sat Aug 14 2004 - 06:16:30 BST


I have two questions about RED:
1. How amine hydrogen should be named? Amber names are H21 and H22 for instance...shall I keep this names ar do I have to do the same as for methyls?


They should be equivalenced in the first RESP stage. So they should have the same name without a "T".
Please the FAQ there is an example for methylamine http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#32


2. The gaussian optimization with a RG residue with a methyl instead of sugar takes ages when using Opt=Tight and SCF(Conver=8). Does one really need these tight convergence criteria?

If you want to derive reproducible RESP charges, the answer is 'YES' otherwise 'NO'. Please read the manual in the section related to the QM input generated all is explained...


45) Re: AMBER: RED fails to read g94 output

From: FyD (fyd_at_u-picardie.fr)
Date: Sat Aug 14 2004 - 06:34:21 BST


but after the optimization RED exited with the message below. I figured out that this might be the same as the bug no.1 report so I changed the code accordingly. Then i tried to run RED but only for the charge calculation with optimizing again. I used the "JOB1-gau.log" file as minimization output but RED exited saying that the minimization output is not valid.

Several things:
- The atom order in your starting PDB file and QM output must be THE SAME
- Your QM ouput must contains three set of words:
1) 'Normal termination of Gaussian' at the end
2) 'Stationary point found' located BEFORE a
3) 'Standard orientation'

I know sometimes the 'Standard orientation' is not printed in the Gaussian ouput... In this case, use the .chk file to run a single point to re-generate an optimized 'Standard orientation'. Add the words 'Stationary point found' located before this 'Standard orientation' and it should be OK...

Check also carefully that each atom name in the starting PDB name is composed of the 'chemical symbol' + 'T' (if needed) + 'a number'. This affects the minimisation ouput recognition i. e. avoid atom names such as O4', 1H1 or HO1
O4' should be replaced by O4 (for instance)
1H2 by H2 (for instance)
HO1 by H1 (for instance)
This has been really improved in the R.E.D. version II which will be released beginning of september.



46) Re: AMBER: problem with RED II

From: FyD (fyd_at_u-picardie.fr)
Date: Thu Aug 26 2004 - 16:44:16 BST


I am using RED II for the multiorientation multiconformation RESP charge calculation. I used 2 conformations and 5 orientations for this purpose. After the minimization of 2 conformations, I concatenated the 2 .log files in single file. Then I ran the RED for the charge calculation.

This means you used:
$OPT_Calc = "Off";
$MEPCHR_Calc = "On";
$JOB_OPT = "your_path/your_concatenated_file.log";


the problem I faced was that after the calculations for the first molecule the process stopped as the scratch directory of GAMESS was not empty (contained .dat file).

This means a former GAMESS job was killed by someone and some temporary files remained in this scratch directory...
To define this scratch directory, I created the following variable in my ".cshrc" file:
setenv SCR /disk2/QM_SCR
Then, each time I get such ERROR message from R.E.D. "ERROR: Scratch directory is NOT empty", I execute "rm $SCR" and then I re-run R.E.D.


to rectify this problem I renamed the *.dat file in the scratch directory to *.pun; after the gamess.00.x execution is over.
mv /usr/people/anshul/misc/GAMESS/gamess/temp/JOB2-gam1.dat /usr/people/anshul/misc/GAMESS/gamess/temp/JOB2-gam1.pun
and also made changes in the RED-vII.pl


You should NOT modify the R.E.D.-II source code for this but simply read the FAQ-39 http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#39

After this the RED process was running nicely and completed the MEP calculations.
------------------------------------------------------------------------
The MEP is/are being computed ...
Conformation 1 ... [ OK ]
See the file(s) "JOB2-gam1-(X).pun"
Conformation 2 ... [ OK ]
See the file(s) "JOB2-gam2-(X).pun"
------------------------------------------------------------------------


but after this when the programme tried to derive the RESP charges, it failed:
---------------------------------------------------------------------
The RESP charges are being derived ... [ FAILED ]
See the "punch2" and "output(1|2)" files
---------------------------------------------------------------------


The output and punch files showed the following error:
---------------------------------------------------------------------
Statistics of the fitting:
The initial sum of squares (ssvpot) 0.000
The residual sum of squares (chipot) 0.000
The std err of estimate (sqrt(chipot/N)) NaN
ESP relative RMS (SQRT(chipot/ssvpot)) NaN
----------------------------------------------------------------------


I red in the FAQ of RED that this error is because RED did not find .dat files in the scratch directory of GAMESS. I changes the .dat files in the RED-vII.pl file to .pun but still it ended with the same error. Can anyone tell me how to deal with this. if i keep te .dat files then it does not run and if i rename it, it still does not run to completion. Is there anyother way by which we can handel these problems?


See http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#39. Then, if you still have this problem look at the end of the RESP ouputs (i. e. "output1" and/or "output2" files) you might find an error message coming from the RESP program itself. Indeed, with R.E.D.-II, multi-orientation and multi-conformation can be performed in same time. This means that the number of centers increases very quickly. And from what we checked there is a limitation in RESP to 500 centers...
- Read the bug-5 for a discussion about this problem http://www.u-picardie.fr/labo/lbpd/RED/bugs-I
- Check the number of centers you have in your molecular systems i. e.: number of centers = $atoms * $i conformations * $j orientations
If you are above 500, look at the RESP source code and try to increase this number and recompile RESP, or simply use 2 conformations and only 4, 3 or 2 orientations...



47) Re: AMBER: RESP calculation with different multiplicities

From: FyD (fyd_at_u-picardie.fr)
Date: Thu Aug 26 2004 - 19:09:18 BST


I am trying to calculate RESP partial charges using RED-vII program. This program run Gaussian calculations considering the system multiplicity=1 by default. Can anybody tell me what I have to change in the program RED.pl or RED-vII.pl to consider a different multiplicity?

Just create a new variable $MULT_VAL in the MAIN PROGRAM section (at the end) similarly to the $CHR_VAL one
$CHR_VAL = "3";
$MULT_VAL= "Your value";


and introduce this new R.E.D. variable in the GAMESS/Gaussian minimization/MEP inputs i. e. something like for GAMESS
ICHARG=$CHR_VAL MULT=$MULT_VAL
instead of
ICHARG=$CHR_VAL MULT=1

and something like for Gaussian:
printf JOB1_FILE ("Optimization %s \n\n%s %s \n",$TITLE,$CHR_VAL,$MULT_VAL);
instead of
printf JOB1_FILE ("Optimization %s \n\n%s 1 \n",$TITLE,$CHR_VAL);

Same idea for QM MEP input... I think it should work.


48) Re: AMBER: problem with RED II

From: FyD (fyd_at_u-picardie.fr)
Date: Fri Aug 27 2004 - 11:02:37 BST


Thanks for the reply. I was able to rectify the problem without modification to the RED source code. I have calculated the RESP charges and then created the frcmod and prepin files for the new residue. But now I face another problem. 1. I made the modifications to the protein structure by attaching an organic molecule to an aminoacid residue in INSIGHT II. In the saved pdb file the attached organic molecule is present in the end of the file as the last residue. Now I load my prepin and frcmod files followed by this molecule in leap and make proper bonds between the organic molecule and the aminoacid. Now when I check this structure it shows a few warnings of close contacts and an error:

If I understand you, you have two parts in your residue... The aminoacid and the org. fragment. This means you have two residues and you have a branched molecule...

Created a new atom named: OXT within residue: .R ALA 582

Did you check the atom names of this ALA 582? It must match the ones in the ALA OFF library of xleap.

How to overcome this problem?

Check carefully the atom name AND residue name in your PDB file and in your xleap OFF libraries... They have to be the same...

2. Next I tried to make modifications at 2 aminoacids. Now when I tried to load this structurte in leap, leap aborts with this error:

ERROR: Comparing atoms C15, H161, H162, H163 to atoms C15, O11, H161, H162
!FATAL ERROR----------------------------------------
!FATAL: In file [chirality.c], line 120 !FATAL: Message: Atom O11 is not
in the first list
!ABORTING.


First see http://amber.scripps.edu/bugfixes/8.0/bugfix.15. I wonder if your problem does not come from the fact that your have two residues in this new aminoacid and consequently that you have a branched molecule. I had this problem once and was able to solve it checking carefully the 'head' and 'tail' of each residue... See the xleap documentation for this. Moreover, I wonder if you should not had a 'TER' character between the two parts of your new aminoacids in your PDB file as your molecule is branched...


49) Re: AMBER: problem with RED II

From: FyD (fyd_at_u-picardie.fr)
Date: Sat Aug 28 2004 - 10:52:20 BST


As you had advised me earlier, I splited my structure into 2 parts to calculate the RESP charges. For this first I used AM1 method for convergence and then I shifted over to 6 31G* method. Meanwhile ZI proceeded from the AM 1 minimized structure, just to check the prpotocl It worked fine. But the process I ran using 6 31 G* has still not converged. the nstep value is 213 and the gradients are as follows:
MAXIMUM GRADIENT = 0.0000873 RMS GRADIENT = 0.0000339
Why is the structure still taking so long to comverge?


This means your AM1 structure is quite different from your 6-31G* one. I guess.
- You could try to use your AM1 structure as input for 3-21G and then the 3-21G structure as input for 6-31G* ?
- You could try to calculate the Hessian every 10 steps as I already told you.
- I am not surprised by your problem because if you do not have a powerfull computer (i. e. a cluster) it might take a while to get a 6-31G* opt. structure...


Another question which I want to ask is as follows: As you know my system consists of 2 parts an amino acid (lys) and an organic molecule. Now when I calculate the RSP charges for this, should the charges calculated for the lys portion be similar to the ones that are used in amber and if yes then how much is the permicible difference?

I would say yes... However what does it mean 'similar' ? I would say, you should not observe a max. difference of 0.07 e or max. 0.1 e (because of additional restraints in your case)... If you get something higher than that you have a problem somewhere (i. e. RESP inputs)...

the organic molecule is bound to lys and i have splitted the system into lys and organic molecule both protected by end groups.

Such branched molecule seems OK to me.


50) Re: AMBER: Differences in RESP charges for G in RG and DG residues

From: FyD (fyd_at_u-picardie.fr)
Date: Sat Sep 04 2004 - 11:12:33 BAST


The RESP atomic charges on guanine atoms (and the other bases also) are different in RG and DG residues. From your experience, are these differences significant? I mean, could one use the guanine charges from RG into DG residues (or the other way arround) without affecting the MM calculations?

Personnally, I prepared unusual nucleotide RESP charges using only B-DNA type nucleotide for B- and A-DNA simulation...

I am asking this question because I am wondering if we derive charges for base analogues using N-methyl-base (which will obviously be different than both in the RG and DG) could these charges be used for both R and D residues? Or, we have to derive the charges using the whole nucleotide in order to be precise? Obviously the latter will take much more time.

I think you wonder the right questions... I would say you derive the RESP charge using one strategy (using multiorientation and multiconformation) and you see what you get during the MD. If it is what you wanted, it is OK otherwise try another method for RESP charge derivation.

Another question: was the chi angle ever taken into consideration when deriving the RESP charges for the nucleic acids residues, or it is assumed that C2'endo pucker means high-anti chi and C3'endo pucker means anti chi?.

- If I remember, the chi dihedral is different in a canonical oligonucleotide B-DNA, and when we optimize the nucleoside alone for RESP charge derivation. This is another problem...
- I think the RESP charges for the whole Cornell et al. database have been derived based on B-DNA sugar pucker and are used also in A-DNA type simulation...


I have done some tests and it seems that the orientation sugar-base affects the value of the charges on several base atoms. Also, in the original paper (Cieplak et al. 1995) is mentioned that there were calculations for the N-methyl bases but this is not detailed. I have obtained charges for the N-methyl-G and it would be interested to compare them with the original values.

- I would try different RESP charge models (i. e. A- and/or B-nucleoside type; methyl derivative or with DMP + nucleoside) and see if they affect your MD simulations...
- Finally, the fact that your unusual base is possibly implied in a strong hydrogen bond is another source of important charge differences which can strongly affect your MD simulation...



51) Re: AMBER: Test Calculation of atomic charges (RESP) for RNA bases

From: Piotr Cieplak (cieplak_at_cgl.ucsf.edu)
Date: Fri Sep 10 2004 - 18:28:25 BST


Yes, you are right there are some discrepancies in the charge fitting using the very old (pre 1994) and current RED approach. Previously it was quite an effort to run geometry optimization on the whole nucleotides now it is much easier to do so. Anyway, now we trust more the multi reorientational approach to get charges. I don't know what is RN-6r category in your tables but apparently one can see some convergence when going from single RN type charge fit to RN-6r and Multi-RN approaches. And this would be probably a way to go. And charges in AMBER - yes with current technology we can think about changing/improving them a little more, but it would also require retesting of other properties in MM and most probably adjusting dihedral parameters.


52) Re: AMBER: Test Calculation of atomic charges (RESP) for RNA bases

From: FyD (fyd_at_u-picardie.fr)
Date: Fri Sep 10 2004 - 01:35:09 BST


I have performed a test calculation using RED (Multi molecule - multi reorientation RESP fit) for the 4 canonical RNA bases: RG, RA, RC and RU. The starting point were canonical A-form nucleotides with C3'-endo sugar pucker. I used 3 reorientations. The fit was performed as described in the original paper of RESP, with all the sugar atoms but C1' and H1' intermolecular fitted.

I used a very similar approach but using C2'endo sugar pucker ;-)

Presumably using this protool I hoped to reproduce the atomic charges on RN nucleotides that are currently used in amber.

Something wrong about the word 'reproducibility'. The RESP charges in the AMBER topology database are NOT 'reproducible' because of mainly 2 reasons:
-1 The molecular orientation of the optimized structure is not unique for a target minimum using GAMESS and/or Gaussian. Indeed, the internal re-rorientation algorithm available in the 2 QM soft does not generate a constant molecular orientation for a target minimum. This is the most important problem. The charges are not reproducible because the molecular orientation is not perfectly controlled in the QM soft (GAMESS or Gaussian) and different when we compare it in the two QM software (GAMESS and Gaussian).
-2 The minimum accuracy slightly affect the charge values also (small effect).

Now if you decide to control the molecular orientation of the optimized structure using the rigid-body re-orientation algorithm implemented in R.E.D. the charges will be reproducible because the molecular orientation based on the three atoms you selected are known. This means that your set of charges will be reproducible by others IF you provide the molecular orientation information based on three atoms you used. That's it.

Finally, you can use multi-orientation RESP fit to average the differences in charge observed for one orientation over several orientations.


To my surprise there are striking differences sometimes (especially on purines). Some time ago Francois told me that acceptable differences are in the range of 0.07. I checked the input and I calculated the charges for single nucleotides and I concluded that there is nothing wrong with my inputs. I will list below the results on the base atoms using wonder marks where I found differences (I tolarated a difference up to 0.03-0.04).

Yes, 0.040 is a VERY common difference... We observed up to 0.07 in one particular case...

Now, the question that I have is: Should we worry about these differences??? Has somebody of you ever tried to do this calculation? If yes ... are your data fitting mine or the Amber better?

See first http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#17. I do not have the answer. I can only answer to you three things:
- The charge incertainty induced by the molecular orientation is very similar to the charge difference between the 2 RESP stages
- Such differences are observed for C atom in CH3/CH2 groups but also for N and O atoms.
- An hydrogen bond can be responsible of a charge difference of 0.2: 0.04 represent ~ 20% of this value...


I am fighting in getting charges for some base analogues. I tried -CH3 instead of sugar but it is clear that this is not a good approach because the differences between the charges from entire nucleodies differ significantly.

I think the limitation of the Me- approach is that a unique conformation can be selected, in this case multi-orientation is an answer. But in same time it is quicker. So it is up to the user... My opinion is that if you can explain what you need with a set of charges used during MD simulations, it is OK...

My idea was now to fit all these new bases with the existing RNA bases and that's why I performed this test calculations.

So you are going to mixte non-reproducible RESP charges with reproducible RESP charges...

We started to generate a set of reproducible RESP charges applying the Cieplak et al. approach and the multi-orientation/multi-conformation RESP fit strategy with R.E.D.-II. The R.E.D. II manual is more or less done. R.E.D. II should be released very soon now...


The problem with my analogues is that there is very litte experimental information them and their behaviour in RNA. They are fluorescent so they might be interesting for the future. I also dont have too much time to run MD simulations using these bases. I am wondering if there is answer to this questions and also I am wondering if these differences are observed...did someone ever thought of changing the charges in Amber??

We started to do it... Not finished yet...


53) Re: AMBER: Test Calculation of atomic charges (RESP) for RNA bases

From: FyD (fyd_at_u-picardie.fr)
Date: Mon Sep 13 2004 - 22:59:11 BST


Thank you very much for your advices and answers. Just a few remarks: RN-6r column in my previous email means that only one molecule was used with 6 reorientations. And indeed I got convergence (~0.02) between those charges and the ones calculated with multi-RN fit. Second, regarding what Francois said...no I am not going to mix reproducible charges and non-reproducible charges. I meant that I will calculate that charges by a multi-RN fit using both the new nucleotides and the old ones.
I generally agree that if one can properly explain MD behaviour with a aset of charges they should be ok....but what i am actually wondering is .. if I dont plan to run MD calculations with these analogues but I am just interested in having them parameterized. One last question: Are the differences observed especially on the atoms of the RG residue acceptable ???


I am a little surprised by the differences you got between RG and Multi-RN for the N1 atom for instance i. e. diff = 0.12. My guess is that such differences do not originate ONLY from an orientation difference (although you never know). Indeed, as I told you the max. difference induced by different molecular orientation we found was .07, and .04 was a difference often observed. So something else might be responsible of such a higher difference.

- In R.E.D. we use Opt=Tight in the geometry optimization to 'force' the accuracy of the target minimum. I think it is always possible that the minima you got might be different from the ones calculated in 1994-1995... As said Piotr, at that time it was more difficult to get a whole nucleotide optimized.

A good example of this problem is the methyl rotamers closed to a carbonyl which are responsible of quite strange charge difference. Did you try to compare the charges of methyl rotamers of NMA ? The minima present almost the same energy values but the RESP charges are really different...

- The orientation of DMP can induce tricky results. See http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm#17
The question is: which CH3O group is defined as the number X or the number Y in the inter-molecular restraints ? This is ambigous.

There are so much parameters that it is difficult to say, the most random being that everybody is human and that errors can be very easily done in particular in tables of numbers and/or RESP inputs... It is also a reason why we decided to wrote R.E.D. to provide an automatic/rigorous way to derive RESP and ESP charges to limit the potential number of errors.



backR.E.D. II FAQback