As per my understanding goes in QM/MM molecular dynamics you must treat
some active site residues and ligand with QM mask with some keywords. Rest
part of your simulation goes via MM. In QM there are different algorithms
to parameterize but based on one report I presume DFTB-SCC is the best one.
Then you should perform and generate .mdcrd file based on QM/MM molecular
dynamics protocols with keywords specific. And then add QM keywords in the
GBSA or PBSA script to perform binding free energy based on QM/MM method.
Some inputs I already shared in the mailing list but I asked for
confirmation of these scripts. I am resharing them again please amber
developers or some experts verify these scripts.
Full_Mini.in
Initial minimization of QMMM: solvent molecules and added ions
&cntrl
imin = 1,
maxcyc = 2500,
ncyc = 750,
ntb = 1,
ntr = 1,
cut = 12.0,
ifqnt = 1,
/
&qmmm
qmmask= '250,238.75,86,89,75',
qmcharge=0,
qmtheory=1, *(please suggest number based on DFTB-SCC)*
qmshake=1,
qm_ewald=1,
qm_pme=1,
/
Hold the Protein fixed
10.0
RES 1 557
Heating.in
*Heating Step of QMMM: stage-2 &cntrl imin= 0, irest=0, NTX=1, ntb=
1, NTPR=500, NTWX=500, NTWR=500, ntr=1, Tempi=0.0, Temp0=300.0,
ifqnt=1, NTT=3, gamma_ln=1.0, NTC=2, NTF=2, cut= 12.0,
nstlim=2500, dt=0.002,/&qmmmqmmask=
'250,238.75,86,89,75',qmcharge=0,qmtheory=1,qmshake=1,qm_ewald=1,qm_pme=1,
/Keep Protein and inhibitor fixed with weak restraints10.0RES 1 557ENDEND*
equil.in
Equilibration Step of MMP3 (MMMM): stage-1
&cntrl
imin= 0,
irest=1,
NTX=7,
ntb=2,
ntp=1,
PRES0=1.0,
TAUP=2.0,
NTPR=500,
NTWX=500,
ntr=0,
Tempi=300.0,
Temp0=300.0,
NTT=3,
ifqnt=1,
gamma_ln=1.0,
NTC=2,
NTF=2,
cut=12.0,
nstlim=250000,
dt=0.002
/
&qmmm
qmmask= '250,238.75,86,89,75',
qmcharge=0,
qmtheory=1,
qmshake=1,
qm_ewald=1,
qm_pme=1,
/
*md.in <
http://md.in>*
Equilibration Step of MMP3 (MMMM): stage-1
&cntrl
imin= 0,
irest=1,
NTX=7,
ntb=2,
ntp=1,
PRES0=1.0,
TAUP=2.0,
NTPR=500,
NTWX=500,
ntr=0,
Tempi=300.0,
Temp0=300.0,
NTT=3,
ifqnt=1,
gamma_ln=1.0,
NTC=2,
NTF=2,
cut=12.0,
nstlim=2500000,
dt=0.002
/
&qmmm
qmmask= '250,238.75,86,89,75',
qmcharge=0,
qmtheory=1,
qmshake=1,
qm_ewald=1,
qm_pme=1,
/
*MMGBSA script*
Input file for running PB and GB in serial
&general
startframe=1, endframe=5000, interval=5verbose=2,entropy=1,keep_file=0,
/
&gb
igb=5,saltcon=0.15,ifqnt=1,qmmask= '250,238.75,86,89,75',
qmcharge=0,qmtheory=1,
/
On Mon, May 19, 2014 at 1:30 AM, Nitin Sharma <sharmanitin.nus.edu.sg>wrote:
> Hello Soumendranath,
>
> I went through few papers and followed the approach mentioned in J Comput
> Chem 32: 866-877, 2011 . I did minimization of the water and then of the
> selected residues + ligand in consecutive steps. Then I created .mdcrd file
> from minimized structure and calculated MMGBSA energy. However, as you
> results were scary.
>
> Let me know your inputs on my approach and the paper I mentioned
>
> Best,
> Nitin
>
> -----Original Message-----
> From: Soumendranath Bhakat [mailto:bhakatsoumendranath.gmail.com]
> Sent: Monday, May 19, 2014 3:48 AM
> To: AMBER Mailing List
> Subject: Re: [AMBER] QM/MM based binding free energy analysis AMBER
>
> Hii Nitin;
>
> Thats a bit of scary data 1000 kcal/mol. Though I am looking at some
> papers saying for QM/MM GBSA or PBSA such as
> http://www.ncbi.nlm.nih.gov/pubmed/24490903 and
> http://www.ncbi.nlm.nih.gov/pubmed/24631364
>
> Did u perform a QM/MM based MD taking catalytic and crucial residues
> treated with QM parameter and then perform QM/MM GBSA or PBSA? Because 1000
> kcal/mol is the most false approximation ever.
>
>
> On Mon, May 19, 2014 at 1:08 AM, Nitin Sharma <sharmanitin.nus.edu.sg
> >wrote:
>
> > Hello Soumendranath,
> >
> > I just worked on same thing and didn't get what I was expecting after
> > reading all the papers recommending MMGBSA method over other scoring
> > methods. I wished to rescore docking poses after minimization BUT to
> > my surprise I got energy values more than 1000 Kcal/mol which I found
> amusing.
> > Moreover, there was no trend in the values. Hence, I moved to rescore
> > with docking software
> >
> >
> > Do let me know if you have some tips to improve the results
> >
> > Best,
> > Nitin
> >
> > -----Original Message-----
> > From: Soumendranath Bhakat [mailto:bhakatsoumendranath.gmail.com]
> > Sent: Monday, May 19, 2014 3:30 AM
> > To: AMBER Mailing List
> > Subject: [AMBER] QM/MM based binding free energy analysis AMBER
> >
> > Dear Amberists;
> >
> > It has been noticed that recent developments in QM/MM based
> > parametirization raised some better approximation of binding free
> > energy close to experimental counterpart. For literature:
> > http://www.ncbi.nlm.nih.gov/pubmed/22210962
> >
> > Literature and some other trails suggested that DFTB-SCC level QM/MM
> > approximation might leads to a better approximation than MM one.
> >
> > I am just qurious if AMBER developers tried some test runs to check
> > the effect of QM/MM based free energy calculations compared to
> > experimental one for a better approximation of binding free energy?
> >
> >
> > --
> > Thanks & Regards;
> > Soumendranath Bhakat
> > _______________________________________________
> > 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
> >
>
>
>
> --
> Thanks & Regards;
> Soumendranath Bhakat
> _______________________________________________
> 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
>
--
Thanks & Regards;
Soumendranath Bhakat
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun May 18 2014 - 14:00:02 PDT
