Re: [AMBER] Antechamber and parmchk2 problem in case of processing mol2 input files

From: Scott Brozell <sbrozell.iqb.rutgers.edu>
Date: Thu, 20 Aug 2020 21:02:00 -0400

Hi,

Marek, thanks for the detailed bug report.
I have reproduced the problem and will start debugging.

scott

ps
why did the chicken cross the road ?
it was amberlating !

On Thu, Aug 20, 2020 at 12:21:08PM +0200, Marek Maly wrote:
>
> I would like to draw attention to a problem regarding the specific use of
> the antechamber and parmchk / parmchk2 programs, starting with the
> AmberTools17 version. It is the use of these routines to create mol2 and
> frcmod outputs from the mol2 input, where the only operation required is to
> add gaff atom types to the mol2 output file and eventually to create also
> the appropriate frcmod file. The input mol2 file is in my case obtained as
> the output from R.E.D. IV program, used for calculation of partial charges
> using RESP method. At first I thought that the problem with this relatively
> simple action, which the antechamber and parmchk programs handled easily
> until version 16, is due to the presence of the foreign chemical element
> "Si" (atom no. 80) in my residues, but as it turned out (after the Si -
> substitution Si - > C in the input mol2 file) this problem is more general
> in nature. Unfortunately, the problem persists even after this substitution
> and even after disabling the acdoctor routine (running antechamber with "-dr
> no") and possibly "freezing" non-standard bonds in the input mol2 file with
> "f". In the best (AT19) case, all key Amber files are created:
>
> ANTECHAMBER_AC.AC
> ANTECHAMBER_AC.AC0
> ANTECHAMBER_BOND_TYPE.AC
> ANTECHAMBER_BOND_TYPE.AC0
> ATOMTYPE.INF
>
> which are moreover identical (except for the letter "f" for frozen bonds) to
> the same files produced by AmberTools16.
>
> However, my calculation script in AmberTools19 case still ends with an
> absurd error:
>
> "Unable to find charges in file (Mol_m1-o1-sm.mol2)"
>
> where Mol_m1-o1-sm.mol2 is the input mol2 file. This is of course nonsense,
> considering the existence of all the Amber files mentioned above. So in
> spite the perfect creation of all *.AC, *.AC0 and INF files neither the
> output mol2 file nor the frcmod file is created.
>
> I would be very grateful for advice on how to get such great output with
> AmberTools19 or later (mol2 file with added atom types and possibly frcmod
> file - it is clear to me that in the case of a foreign chemical element, the
> frcmod file will not contain concrete bond / nonbond values parameters) as
> could be obtained using antechamber and parmchk in AmberTools16 and older
> versions. This functionality is key for my work and I hope there exists a
> more elegant solution than maintaining an alternative old version of
> AmberTools16 just for this specific need.
>
> The inputs and outputs for one specific residue are available in the
> attached file, namely:
>
> #1 AmberTools16
> a) using parmchk
> b) using parmchk2
> There are identical mol2 outputs but different frcmod outputs (in case b) is
> this file empty)
>
> #2 AmberTools19
> a) using parmchk2 (without modifying the input mol2 script and without
> adding the "-dr no" parameter to the antechamer command)
>
> b) using parmchk2 (unusual "Si-x" bonds are frozen with "f" and the
> antechamber is started with the parameter "-dr no")
>
> You may easily try/check also the case, were the "Si" atom (number 80) is
> substituted with "C" atom, and "-dr no" + "f" are eventually used.
>
> FILES
>
> input mol2 file: Mol_m1-o1-sm.mol2
> calculation script: MakePREPandCHECK_MOL2_TO_MOL2.sh
> output mol2 file: Mol_m1-o1-sm_withAT.mol2 (created just in AT16 case)
> output frcmod file: Mol_m1-o1-sm.frcmod (created just in AT16 case, please
> notice the difference between parmchk ans parmchk2 case).
>
> Then there are the usual Amber outputs and StdOut and ErrOut files
> (SCREEN_OUTPUT_ATxy.txt, ERR_ATxy.txt) with informations/errs reported by
> antechamber/parmchk/2 programs on the screen.

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Aug 20 2020 - 18:30:02 PDT
Custom Search