AMBER Archive (2008)

Subject: Re: AMBER: respgen -a charge groups function ineffective unless

From: Junmei Wang (junmwang_at_gmail.com)
Date: Wed Aug 27 2008 - 15:55:47 CDT


I fixed the problem. You need to download the attached residuegen.c and
prepgen.c and recompile the package.

By the way, I found you try to break a double bond in your molecule (C8-C9).
Although the current version works fine when you breaking a double bond, it
is not recommended. You may get a warning message complaining there is a
unfilled valence for C8. You may safely ignore this warning message if you
don't worry about the planarity of the double bond.

Best

Junmei

On Wed, Aug 27, 2008 at 1:29 AM, Grange Hermitage <ghermitage_at_gmail.com>wrote:

> Junmei,
>
> I find that the new version does not correctly identify the mainchain
> atoms.
> An error report is attached.
> Is it possible that this patch has regressed the earlier residuegen patch
> for incorrect mainchain atoms?
>
> regards,
> John.
>
>
>
>
> On Tue, Aug 26, 2008 at 11:23 PM, Junmei Wang <junmwang_at_gmail.com> wrote:
>
>> Hi,
>> I have improved residuegen to handle different situations of generating
>> residue topologies. Now respgen can generate prepi for normal
>> amino-acid-like or nucleic-acid-like residues, the termainl residues
>> (C-terminal, N-terminal, 3' and 5' nucleic acid residues), the organic
>> molecules (if you want to use multiple conformations and/or cap atoms and/or
>> predefined atomic charges), and amino-acid-like residues with side chains
>> being linked to other residues. Please let me know if you encounter some
>> trouble.
>>
>> Attached are the latest version of residuegen.c and prepgen.c
>>
>> Best
>>
>> Junmei
>>
>>
>>
>> On Sun, Aug 17, 2008 at 10:49 PM, Grange Hermitage <ghermitage_at_gmail.com>wrote:
>>
>>> Hi Junmei,
>>>
>>> How to get a prepi when following steps 1-4?
>>> How to integrate this with residuegen when it has the new sidechain cap
>>> function?
>>>
>>> thanks,
>>> John.
>>>
>>>
>>>
>>>
>>> On Fri, Aug 15, 2008 at 11:19 PM, Junmei Wang <junmwang_at_gmail.com>wrote:
>>>
>>>> Hi, John,
>>>> The reason is that during the gaussain calculations the atom name
>>>> information is lost. So antechamber generates new atom names when reading a
>>>> gaussian output.
>>>>
>>>> You may read the following antechamber command to solve your problem.
>>>>
>>>> antechamber -fi gout -i residue.gout -fo respi -o residue.prepi -c resp
>>>> -a residue.mol2 -fa mol2 -ao name
>>>>
>>>> This command reads in gaussian output (residue.gout) and writes out a
>>>> prepi file. The resp charges are generated and the atom names are from
>>>> residue.mol2.
>>>>
>>>> This problem can also be solved by using a set of antechamber commands.
>>>> The following is one of several possible solutions:
>>>> (1) run antechamber to get gaussian input from residue.mol2
>>>> (2) run antechamber to generate resp charges by reading gaussian output
>>>> or log file, the output file can be an ac or mol2 file.
>>>> (3) run antechamber from the output of the second step to write out a
>>>> charge file (such as antechamber -fi ac -i residue.ac -c wc -cf
>>>> residue.crg)
>>>> (4) run antechamber to read in the charge file from the third step to
>>>> residue.mol2 (antechamber -fi mol2 -fo mol2 -i residue.mol2 -o residue.mol2
>>>> -c rc -cf residue.crg).
>>>>
>>>> Good luck
>>>>
>>>> Junmei
>>>>
>>>>
>>>> On Fri, Aug 15, 2008 at 2:04 AM, Grange Hermitage <ghermitage_at_gmail.com
>>>> > wrote:
>>>>
>>>>> Hi Junmei, this fix is successful in getting gaussian to start.
>>>>>
>>>>> It raises a question, however.
>>>>> The residue.mol2 from which gaussian input is created by antechamber
>>>>> has different atom names than the residue.ac that is extracted from
>>>>> the residue.log created by gaussian.
>>>>> What I have been doing is to change the atom names in the
>>>>> residue.prepin to those that appeared in the residue.mol2 . The atom numbers
>>>>> are the same in the .mol2 and .ac, and from that correspondence I have
>>>>> changing the name in the prepin to that which was in the mol2. Now that the
>>>>> atom numbers don't correspond between the .mol2 and .ac, how shall I
>>>>> establish the correspondence? I am looking for a way that can be automated,
>>>>> for this name patch is time consuming.
>>>>>
>>>>> cheers,
>>>>> John.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> On Wed, Aug 13, 2008 at 2:16 PM, Junmei Wang <junmwang_at_gmail.com>wrote:
>>>>>
>>>>>> Hi, John
>>>>>> To generate internal coordinates, there is a requirement on the
>>>>>> sequence order of the atoms: the current atom i must be bonded to at least
>>>>>> one foregoing atoms j (j<i). Otherwise, error happens. You molecule has
>>>>>> this problem, for example, the No 11 atom is bonded to Nos 24,28,33,35, but
>>>>>> none atom from No 1 to No 10. You may rearrange the sequence order of atoms
>>>>>> to meet the requirement. To make life easier, I wrote a new function to
>>>>>> arrange the sequence order of atoms automatically. This function is called
>>>>>> when the output format of antechamber is gzmat, mopint and jzmat. Please
>>>>>> copy the attached two files to your antechamber directory (AmberTools) and
>>>>>> recompile the package.
>>>>>>
>>>>>> Best
>>>>>>
>>>>>> Junmei
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Mon, Aug 11, 2008 at 9:53 PM, Grange Hermitage <
>>>>>> ghermitage_at_gmail.com> wrote:
>>>>>>
>>>>>>> Hi Junmei,
>>>>>>>
>>>>>>> Attached is a report for the condition described in the second
>>>>>>> paragraph below.
>>>>>>>
>>>>>>> cheers,
>>>>>>> John.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On Mon, Aug 11, 2008 at 5:29 PM, Grange Hermitage <
>>>>>>> ghermitage_at_gmail.com> wrote:
>>>>>>>
>>>>>>>> Hi Junmei,
>>>>>>>>
>>>>>>>> Relevant files attached.
>>>>>>>> In tleap, the charge of the residue is zero if the blank lines
>>>>>>>> before the charge groups are removed from the generated files, and not zero
>>>>>>>> if not removed.
>>>>>>>>
>>>>>>>> There is another problem, this one with antechamber generating
>>>>>>>> gaussian input from a mol2 file. The gaussian input is incorrect for the
>>>>>>>> mol2s I am using unless the atoms are renumbered according to a lightest
>>>>>>>> sidegroup first approach, otherwise the generated input uses the same atom
>>>>>>>> twice in gaussian input lines, and generates unacceptable angles and divide
>>>>>>>> by zeros. is this problem known to you? If not, I will generate an error
>>>>>>>> report.
>>>>>>>>
>>>>>>>> cheers,
>>>>>>>> John.
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Wed, Aug 6, 2008 at 1:38 AM, Junmei Wang <junmwang_at_gmail.com>wrote:
>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>> Could you send us an example? I think a blank line is needed
>>>>>>>>> between group charge constraint and inter-molecular charge constraints. I
>>>>>>>>> checked the output of respgen.c (RESIDUE_GEN_RESP.INPUT1 and
>>>>>>>>> RESIDUE_GEN_RESP.INPUT2) in $AMBERHOME/test/antechamber/residuegen, they
>>>>>>>>> work fine.
>>>>>>>>>
>>>>>>>>> Best
>>>>>>>>>
>>>>>>>>> Junmei
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Sat, Aug 2, 2008 at 12:00 AM, Grange Hermitage <
>>>>>>>>> ghermitage_at_gmail.com> wrote:
>>>>>>>>>
>>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> Using the -a input option for respgen in AmberTools to define
>>>>>>>>>> charge groups:
>>>>>>>>>> you must remove the blank line before the charge groups in
>>>>>>>>>> the generated .respin1 or .respin2 file or the charge group definitions are
>>>>>>>>>> ineffective.
>>>>>>>>>>
>>>>>>>>>> Could the prepgen matter I mentioned yesterday and this respgen
>>>>>>>>>> one be placed on a bug list?
>>>>>>>>>>
>>>>>>>>>> Ta,
>>>>>>>>>> John.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>





-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu