Re: [AMBER] MMPBSA.py problem

From: marawan hussain <marawanhussain.yahoo.com>
Date: Tue, 15 May 2012 20:22:06 -0700 (PDT)

HI Jason,
Thanks for help..The QM/MM option i figured out that it's not compatible with the energy decomposition option...

King regards
Marawan


________________________________
 From: Jason Swails <jason.swails.gmail.com>
To: marawan hussain <marawanhussain.yahoo.com>; AMBER Mailing List <amber.ambermd.org>
Sent: Wednesday, 16 May 2012 12:02 PM
Subject: Re: [AMBER] MMPBSA.py problem
 
Hello,

I'll just add a couple comments here in addition to what Bill said.

On Tue, May 15, 2012 at 5:13 PM, marawan hussain
<marawanhussain.yahoo.com>wrote:

> Dear AMBER user,
>
> I'm doing a MMPBSA.py calculation on a ligand protein complex..The final
> file (FINAL_DECOMP_MMPBSA.dat) is empty. I have read similar problem but if
> my understanding was correct, it said that i need to mention the ligand as
> well in the print residue section. I have done so, i got the output this
> time, but i don't know why i should mention the ligand because the best
> guess mask has already mentioned the ligand correctly, here is my input:
>

Yes, you currently do still have to mention a ligand residue (the ligand
residue if the ligand is just 1 residue).  I've been meaning to fix this,
but I haven't gotten around to it yet.

Note that this issue is independent of the mask prediction (which is done
correctly in your case).

Input file for running PB and GB
> &general
>    startframe=1, endframe=2, verbose=1, netcdf=1
> #  entropy=1,
> /
> &gb
>  igb=2, saltcon=0.100, qm_residues="18-19"
> /
> &decomp
>  idecomp=1, print_res="18-19"  (this gives empty output, but
>  print_res="320, 18-19" gives the results below )
>  dec_verbose=1,
> /
>
> I have to other questions:
> 1-Can i use the ifqnt option eventhough my original trajectory is purely
> classical, i have read the manual and the AMBERtool but this was not
> clear...???.And if this is possible, do i need to include the ligand number
> as well or the script well correctly identify it to be included, and can
> this be paralleled or not....???
>

Yes, you can.  However, if the shape of the potential energy surface of the
original force field and the semi-empirical quantum method are
significantly different, I expect you would get horrible results.

As an aside, I don't know how well decomp and QM/MM work together.  I'm
pretty sure the program will not quit gracefully.  It may even produce
results, but I'm also pretty sure the decomposed values will be
meaningless.  The non-local nature of QM/MM energies does not lend itself
to decomposition analysis, and there is no code that I know of in place to
handle this combination.

2-The output i got is messy, do i need a special program to open:
>

As Bill mentioned, this is a CSV-format file.  Most spreadsheets can open
CSVs, and most scripting languages have an API to read/write them, so it's
a very convenient file type to dump large amounts of data to (and
decomposition has large amounts of data).

You can add the variable "csv_format=0" to the &decomp section to get an
ASCII printout that's the same as the AmberTools 1.5 printout.  (You can
modify the _MMPBSA_info file and change INPUT['csv_format'] to 0 and just
run the command

MMPBSA.py -rewrite-output

to get the new output file without redoing the calculation).

HTH,
Jason

-- 
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue May 15 2012 - 20:30:03 PDT
Custom Search