Re: [AMBER] Methylated histidine lib and frcomd files

From: FyD <>
Date: Thu, 11 Jul 2013 10:29:39 +0200

Dear George,

> 1) I looked at the conformation generated by the QM program. It is almost
> identical with the input conformation (taken from a crystal structure)
> and therefore looks very reasonable to me.

This is just perfect... One should particularly check the optimized
geometry if the molecule has a non-null total charge (your case).

> 2) Concerning the molecular orientation procedure I am really unsure about
> its effect on the calculation. I haven't touched any of the Mol_red*.p2n
> files that were all generated automatically generated by the server.

See the P2N file I generated using the non-automatic mode (to keep the
3 residues you defined) of R.E.D. Server/Ante_R.E.D. 2.0:
as you can see there is a 'REORIENT' line...

> Do you think I should include the same REMARK REORIENT statement:
> REMARK REORIENT 1 5 7 | 7 5 1,

ok - why not - or simply use what Ante_R.E.D. 2.0 has generated...

> in all Mol_red*.p2n files for consistency? So to have the same orientation
> procedure applied in all molecules?

the idea is that the charge values become reproducible if the QMRA
procedure is applied for all the molecules involved in charge

> If yes, I can edit the Mol_red*.p2n files accordingly but then I don't
> know how to resubmit the job to the R.E.D server. Can you tell me how to
> do this?

Just resubmit your P2N file as you did it the first time:

if you want to save some cpu time, you could skip the geometry
optimization step and provide the geometry optimization output
generated by Gaussian with the P2N file and choose the mode 2 of
R.E.D. Server/R.E.D. IV:

regards, Francois

>> Dear George,
>>> The RED job P5749 with the methyl-histidine dipeptide has finished.
>> I looked at your P5749 data:
>> 1) about the conformation of the dipeptide that R.E.D.
>> Server/Gaussian09 has generated:
>> See for instance: P5749/Data-R.E.D.Server/Mol_m1/Mol_m1-o1-qmra.pdb
>> Do you 'like' this conformation? i.e. is this conformation in
>> agreement with that _you want_ to get involved in charge derivation?
>> as the conformation strongly affect charge values many papers suggest
>> that one should always have a look at the optimized geometry generated
>> by a QM program. In short - one selects a conformation for charge
>> derivation that is close to that/these observed experimentally or See
>> Cieplak et al.:
>> for
>> more information.
>> re-run R.E.D. Server until you 'like' the optimized geometry generated
>> for your dipeptide by changing the input geometry or by forcing the
>> use of dihedral (in your case) constraint(s) during geometry
>> optimization; in particular have a look at the phi, psi and chi1
>> dihedral angles...
>> 2) about the molecular orientation procedure applied after geometry
>> optimization:
>> ls -lv P5749/*.p2n
>> -rw-r--r-- 1 user user 3797 juil. 9 16:26 Mol_red1.p2n
>> -rw-r--r-- 1 user user 1261 juil. 9 16:26 Mol_red2.p2n
>> -rw-r--r-- 1 user user 3617 juil. 9 16:26 Mol_red3.p2n
>> -rw-r--r-- 1 user user 1097 juil. 9 16:26 Mol_red4.p2n
>> -rw-r--r-- 1 user user 3611 juil. 9 16:26 Mol_red5.p2n
>> -rw-r--r-- 1 user user 3501 juil. 9 16:26 Mol_red6.p2n
>> egrep "REORIENT|TRANSLATE|ROTATE" P5749/*.p2n
>> P5749/Mol_red2.p2n:REMARK REORIENT 1 5 7 | 7 5 1
>> P5749/Mol_red4.p2n:REMARK REORIENT 1 5 7 | 7 5 1
>> as you can see only the data that is stored internally (P2N files 2 &
>> 4) by R.E.D. Server follow the 'RBRA' procedure, while your dipeptide
>> (molecules 1, 3, 5 and 6) does follow the 'QMRA' one; in other words
>> the re-orientation procedure (RBRA) of each optimized geometry is not
>> performed for the dipeptide you generated with Ante_R.E.D. 2.0; My
>> feeling is that Ante_R.E.D. 2.0 should have generated a keyword such
>> as 'REMARK REORIENT blabla (numbers)' in the p2N file; but it looks
>> like you remove this keyword. Could re-run the R.E.D. Server job
>> keeping this 'REMARK REORIENT blabla (numbers)' keyword.
>> Let me know what you decided...
>>> Would it be possible to post-process the R.E.D. Server/R.E.D. IV data to
>>> generate atom types, residue connections and frcmod files?
>> ok - no problem.
>> regards, Francois
>>>> Dear George,
>>>> You first need to create a _correct_ dipeptide molecule, and then save
>>>> it to the PDB file format; ACE means CH3CO & NME means NHCH3 (at that
>>>> time you will be able to run Ante_R.E.D. (better using ante_R.E.D. 2.0
>>>> vs 1.x; in your case 1.x should be ok).
>>>> in your case you need to create:
>>>> R = side chain of this methylated residue
>>>> (it looks like yours has a total charge = 0)
>>>> pay attention to:
>>>> - create two trans peptide bonds
>>>> - define the phi, psi & chi dihedral angles...
>>>> then carefully read:
>>>> vs
>>>> and it should be OK ;-)
>>>> we can post-process your R.E.D. Server/R.E.D. IV data using R.E.D.
>>>> Python so that the atom types, residue connections and frcmod files
>>>> are automatically generated; just ask in the q4md-fft or Amber list
>>>> and provide the PXXXX R.E.D. Server job name... R.E.D. Python handles
>>>> by now all the Amber XX force fields (XX = year).
>>>> regards, Francois
>>>>> I created a pdb with a methylated His together with ACE and NMA caps
>>>>> (attached). Then I run (from the RED-III.52-Tools) on
>>>>> this pdb to create a p2n file (attached).
>>>>> The R.E.D. Server/Ante_R.E.D. 2.0 generates a p2n file with only one
>>>>> residue
>>>>> which I guess is what is needed in this case.
>>>>> Finally, I went to the R.E.D server, I selected the "Use RED IV for
>>>>> automatically generating amino acid fragments" option and I uploaded
>>>>> the attached p2n file. Thie job crashed with the error in the log
>>>>> file:
>>>>> ERROR: Wrong inter-molecular charge constraint or equivalencing
>>>>>> Dear George,
>>>>>>> It sounds like it is a lot easier if I use the R.E.D server where
>>>>>>> the
>>>>>>> work-flow has been automated, right?
>>>>>> you first run R.E.D. Server/Ante_R.E.D. 2.0 & then re-run R.E.D.
>>>>>> Server/R.E.D. IV after having checked/modified (if needed) the p2n
>>>>>> file generated by Ante_R.E.D.
>>>>>> PDB --> P2N ---> mol2
>>>>>> please see:
>>>>>> We are aware these 2 steps are a limiting factor by now; the main
>>>>>> advantage is that the user can modify the P2N file(s) after its
>>>>>> generation and this makes the system quite flexible and allows
>>>>>> handling complex cases of charge derivation
>>>>>> With R.E.D. python all is combined in one step. But in this case the
>>>>>> code is far more 'sophisticated'.
>>>>>>> I saw that the server interfaces either with Ante_R.E.D. 2.0 or
>>>>>>> R.E.D.
>>>>>>> IV
>>>>>>> program. Can you tell me what is the difference and which one should
>>>>>>> I
>>>>>>> use?
>>>>>> Please read the tutorials; in short you first execute Ante_R.E.D. to
>>>>>> generate the P2N file(s) using PDB file(s) as input and then using
>>>>>> the
>>>>>> P2N file(s) you execute RED in a second step
>>>>>>> Just to double check: is this approach suitable for a
>>>>>>> methyl-histidine
>>>>>>> residue that is a part of a protein (actin)? I will extract this
>>>>>>> residur
>>>>>>> from the original pdb file and then upload it to the server.
>>>>>> You extract this residue from the protein (or you construct it by
>>>>>> controlling the conformation i.e. the phi, psi and chi dihedrals),
>>>>>> transform it into a dipeptide (PDB file to be transformed into P2N
>>>>>> file) and then from this dipeptide you generate the central (and
>>>>>> N-term & C-term) fragments (to be re-inserted in your protein) using
>>>>>> R.E.D. Server/R.E.D.
>>>>>> See:
>>>>>> then all together:
>>>>>> and finally all together automatically from a single dipeptide:
>>>>>> regards, Francois
>>>>>>> Dear George,
>>>>>>>> Does R.E.D. III.x need a GAMESS installation which the perl script
>>>>>>>> will
>>>>>>> somehow locate?
>>>>>>> See the installation procedure described in the RED version II pdf
>>>>>>> file.
>>>>>>> See the part "-III- HOW TO USE R.E.D. & X R.E.D.?" page 9
>>>>>>> i.e. you need to:
>>>>>>> - Install GAMESS (or Firefly or Gaussian) _and_ RESP.
>>>>>>> - Check that GAMESS (Firefly or Gaussian) works from your
>>>>>>> X-terminal.
>>>>>>> (i.e. the binaries and scratch path are defined and found)
>>>>>>> - Same remark for RESP: install & test it before interfacing it with
>>>>>>> R.E.D.
>>>>>>> you can use the standalone version of the RESP program from our web
>>>>>>> site:
>>>>>>>>> From R.E.D. III.x, we obtain a Tripos mol2 file that we can
>>>>>>>>> directly
>>>>>>> then
>>>>>>>> load into leap and get the .lib and .frcmod files we want?
>>>>>>>> After loading the mol2 file into leap, do we need to run some kind
>>>>>>>> of
>>>>>>> script to change atomnames etc?
>>>>>>> If you use R.E.D. Server/Ante_R.E.D. 2.0 atom names are checked
>>>>>>> (i.e.
>>>>>>> in
>>>>>>> the philosophy of a FF library two atoms can NOT share the same
>>>>>>> name
>>>>>>> in
>>>>>>> a given residue). See
>>>>>>> Once you got the mol2 file(s) from R.E.D. perl you need to add the
>>>>>>> FF
>>>>>>> atom types; here we do use a LEaP script and define eaxh FF atom
>>>>>>> types
>>>>>>> using the 'set' command.
>>>>>>> See for instance:
>>>>>>> regards, Francois
>>>>>>> PS With R.E.D. Python all is done automatically from a PDB file.
>>>>>>>>> Dear George,
>>>>>>>>>> I was wondering if any user has constructed a lib/prep and a
>>>>>>>>>> frcmod
>>>>>>>> file
>>>>>>>>>> for a methylated histidine to share with me.
>>>>>>>>>> If no, what is the general procedure to make these files?
>>>>>>>>>>> From the tutorial:
>>>>>>>>>> I understand that a RESP calculation must be made with R.E.D to
>>>>>>>>>> get
>>>>>>> the
>>>>>>>> partial charges.
>>>>>>>>>> Then, I think some AmberTools must be used but I am not sure
>>>>>>>>>> which.
>>>>>>>>> See Figure 1 at

AMBER mailing list
Received on Thu Jul 11 2013 - 02:00:03 PDT
Custom Search