Dear Francois,
On Fri, Dec 13, 2013 at 4:24 AM, FyD <fyd.q4md-forcefieldtools.org> wrote:
> Dear Kamali,
>
> Parameterizing a TS is quite interesting; that being said this is far
> more difficult than a regular true minimum. Moreover you use the
> OPT_Calc = "OFF1" option in the Configuration.py (quite dangerous):
> # OFF1 Geom. opt. is skipped & the PDB input file is directly used in
> MEP comput.
>
I also agree that I shouldn't use these results for my final calculations;
I just used the OPT_Calc = "OFF1" option as part of a "practice run" so
that I could see how to successfully submit a job to R.E.D Python and how
the outputs would look.
In a previous message you also asked why I used the INTRA-MCC constraints
that I did: this was also part of the "practice". I chose atoms to
constrain that made sense to me, and I just wanted to see what would
happen. I actually don't think I need to constrain any atoms beyond the
default setting; is it possible to also use defaults for INTRA-MCC?
Maybe I should start a different thread for this question, but I have
actually tried to submit job P8902 with OPT_Calc = OFF and providing my own
QM output. However, I've gotten errors in the R.E.D.-
Server-14651.master0.q4md-forcefieldtools.org file:
*...*
*Reading JOB1*.log for molecule : 1*
*Find "Normal termination of Gaussian"*
*Find "Normal termination of Gaussian"*
*Find "Stationary point found"*
*Find "Standard orientation"*
*Element in JOB1 is not the same with PDB*
*Find big error in log : /home/u0263B7BZ/P8902/JOB1-gau_m1-1.log*
*JOB1*.log is not correct , can not continue......*
*Information of end*
*Time : Wed Dec 11 22:26:46 CET 2013*
But to me the JOB1-gau_m1-1.log looks fine. Did I have to modify my QM
output (similar to what one must do for R.E.D Server Perl)?
> In these conditions _the key_ is the definition of the atom
> connectivities from this PDB file (atom typing & FF parameters are
> based on these atom connectivities). You have C-H bonds with 1.5 Angs
> length in your PDB input file. Are you sure an atom connectivity is
> created in this case (VMD does not...)?
>
You're correct; those bond lengths should be fixed, and I will do that
before the next time I submit my structure.
> Your best bet here (because RED Python computes atom connectivities
> after the geometry optimization step and you skip this step...) is to
> provide the atom connectivities *YOU* want in the PDB input file and
> *prevent* their computation.
>
> See
> http://q4md-forcefieldtools.org/REDS-Development/Demo2-Files/Project.config
> MOLECULE1-CALCONECT = OFF
>
> then you need to carefully think which atom types you want: obviously
> two CT-H1 or CT-H2 bonds (CT, H1, H2 are atom types) cannot have two
> different equilibrium values; i.e. ~1.0 and ~1.5 Angs -> a new atom
> type has to be created...
>
> Moreover your phosphor atom has 5 bonds I would set up a new atom type
> for it...
>
Thank you for the suggestion, Francois. But I am wondering if the new
phosphorus atom type is necessary? I ask because it is possible in nature
for phosphorus to form 5 bonds.
>
> Finally I really hope the Cartesian coordinates you have were checked
> by QM: once again better providing a QM output (with one imaginary
> frequency) as input for RED Python...
>
For job P8902, I used Gaussian09 with these options:
#P hf/6-31G* Opt=(Tight,CalcFC) Freq SCF(Conver=8) Test
I tried to answer this question on my own, but I got confused: why do I
need imaginary frequencies for my calculations, and why only just one?
Thank you very much again, Francois, and have a great night,
Kamali
>
> I hope this helps...
>
> regards, Francois
>
>
> Quoting FyD <fyd.q4md-forcefieldtools.org>:
>
> > Kamali,
> >
> > please check the PDB file format you use as input...
> > see
> http://q4md-forcefieldtools.org/REDS-Development/Demo2-Files/readme.txt
> >
> > regards, Francois
> >
> >
> > Quoting FyD <fyd.q4md-forcefieldtools.org>:
> >
> >> Dear Kamali,
> >>
> >>> I have been trying a few things out with my transition state structure
> of
> >>> RNA cleavage, and I had a few follow-up questions. My questions are
> >>> regarding my job P8824:
> >>
> >> ok
> >>
> >> in the PDB input file:
> >> ATOM 15 O2P RT0 1 31.511 16.369 54.925
> >> O
> >> ATOM 16 O5' RT0 6 32.667 14.580 56.630
> >> O
> >> ATOM 17 P RT0 1 30.934 15.202 55.888
> >> P
> >> ->
> >> ATOM 15 O2P RT0 1 31.511 16.369 54.925
> >> O
> >> ATOM 16 O5' RT0 1 32.667 14.580 56.630
> >> O
> >> ATOM 17 P RT0 1 30.934 15.202 55.888
> >> P
> >>
> >> & look at the residue name generated by R.E.D. Python
> >>
> >> You deactivated many keywords in the Project.config file letting the
> >> defaults to be executed; this is correct. I am glad to see you can do
> >> the job without a tutorial. We need to provide a tutorial. The
> >> description of the keywords in the Project.config and Configuration.py
> >> files is too limited...
> >>
> >>> Following your suggestion in the private communication (when we first
> >>> talked about this structure), for this job I just ran the PDB through
> the
> >>> RED Server Dev. with the option OPT_Calc = Off1 in the Project.config
> file,
> >>> as a practice run.
> >>
> >> OPT_Calc = Off1 means you skip the QM job and use the Cartesian
> >> coordinates from the PDB file to be used in MEP computation. (I am
> >> really not a big fan of this approach; I would prefer you provide a QM
> >> job for a Cartesian coordinate set with one imaginary frequency). ok
> >> for a 'dirt cheap' job ;-) (just my personal opinion)
> >>
> >>> I was wondering if there is any documentation for the
> >>> files generated by RED Python?
> >>
> >> Not yet - so far we work on a new version of R.E.D. Python with new
> >> features...
> >>
> >> Just ask for help; sorry for that...
> >>
> >>> I found two types of files in the
> >>> P8824/Data-R.E.D.Server/Mol_m1 folder: *sm* files (which I know are
> files
> >>> containing intra-molecular charge constraints) and the *ia*: what does
> the
> >>> *ia* mean? I also have two .mol2 files: Mol-ia0_m1-c1.mol2
> >>
> >> sm means files related to single molecule charge fit (without intra-mcc)
> >> ia means files related to charge fitting carried out with intra-mcc
> >> (because you use 'MOLECULE1-INTRA-MCC1 = 0.0 | 1 4 11 | Keep')
> >> mm means files related to multiple molecule fit (i.e. number of
> >> molecules > 1)
> >>
> >>> and Mol-sm_m1-c1.mol2, but these files appear to have two columns of
> >>> charges rather than one next to the rightmost **** column. Could you
> tell
> >>> me which column is the correct one for charges, or alternatively point
> me
> >>> to documentation that answers my questions?
> >>
> >> See
> >>
> http://cluster.q4md-forcefieldtools.org/~x/P8824/Data-R.E.D.Server/Mol_m1/Mol-sm_m1-c1.mol2
> >>
> >> 1 C4' 30.706000 13.015000 52.944000 CT 1 U01 0.0520 0.8780
> ****
> >>
> >> 0.0520 is the column of the atomic charge value
> >> 0.8780 is the column of the atomic polarizability
> >> (useful for the non-additive FF model)
> >>
> >> ---
> >>
> >> I do not understand why you use: MOLECULE1-INTRA-MCC1 = 0.0 | 1 4 11 |
> Keep
> >> in
> >>
> http://cluster.q4md-forcefieldtools.org/~x/P8824/Data-R.E.D.Server/Mol_m1/Mol-ia0_m1-c1.mol2
> >>
> >> i.e. 0.6957-0.5092-0.1865=0.0000
> >>
> >> then look at the impact of the INTRA-MCC1:
> >>
> http://cluster.q4md-forcefieldtools.org/~x/Project/P8824/Data-R.E.D.Server/Mol_m1/Statistics_m1.txt
> >>
> >> See the table:
> >> Charge values for single molecule fit; without (SM) vs with (IA)
> INTRA-MCC:
> >> -> the impact seems quite important...
> >>
> >> You could also look at :
> >> - the RRMS values of the corresponding fits.
> >> - the es*.pdb files to display where the errors are.
> >> See http://q4md-forcefieldtools.org/RED/resp/#intro
> >> espdb
> >> -j PDB-like file with relative residual in the TempFactor field
> >> esqpotpdb
> >> -y PDB-like file with input MEP values in the TempFactor field
> >> (atomic units)
> >> esmpotpdb
> >> -z PDB-like file with MEP values for new charges in TempFactor field
> >> (atomic units)
> >>
> >> let us know if you need more information...
> >>
> >> regards, Francois
> >>
> >>
> >>> On Tue, Dec 10, 2013 at 8:12 AM, Kamali Sripathi
> >>> <ksripath.umich.edu> wrote:
> >>>
> >>>> Dear Francois,
> >>>>
> >>>> Thank you very much for the detailed information, and I'm sorry for
> the
> >>>> confusion. I would have mentioned that I had questions about a topic
> >>>> related to our private discussion, but I didn't want you to have
> >>>> to look up
> >>>> that communication, and so I tried to start a new thread :)
> >>>> Looking back, I
> >>>> see that my initial email in this thread was unfortunately less clear
> than
> >>>> I intended.
> >>>>
> >>>> I will try the protocol you suggested, but I think it would be better
> for
> >>>> me to use the RESP-A1 charge type because I'm working with RNA rather
> than
> >>>> carbohydrates, and using AMBER10 for my simulations.
> >>>>
> >>>> Thank you very much again, and have a great day,
> >>>>
> >>>> Kamali
> >>>>
> >>>>
> >>>> On Tue, Dec 10, 2013 at 2:52 AM, FyD <fyd.q4md-forcefieldtools.org>
> wrote:
> >>>>
> >>>>> Dear Kamali,
> >>>>>
> >>>>> > No, thank you very much for the offer. I have the PDF and refer to
> it
> >>>>> often
> >>>>> > because there are concepts that I'm still learning :)
> >>>>>
> >>>>> Yes a looot to learn in this work...
> >>>>>
> >>>>> > I have two final question: from examples on the RED Server (perhaps
> >>>>> this is
> >>>>> > not applicable to RED Perl or RED Python), it appears that one can
> also
> >>>>> set
> >>>>> > the INTRA-MCC constraints to defined values, e.g., sugar carbons to
> >>>>> > standard values in the AMBER force field instead of setting the
> sum of
> >>>>> > certain atoms to 0 or to another value. Do you suggest one way
> over the
> >>>>> > other?
> >>>>>
> >>>>> I would constrain hydrogen atoms (only these in the methylene &
> methyl
> >>>>> groups, using the 'F' flag in the intra-mcc to keep them in the FF
> >>>>> lib) _only_ if you use the GLYCAM force field. So the 'RESP-C2'
> charge
> >>>>> model (to be provided in the Configuration.py file to overwrite the
> >>>>> default RESP-A1):
> >>>>> see
> http://q4md-forcefieldtools.org/REDS-Development/popup/popkeyword.php
> >>>>>
> >>>>> Here once again I would use INTRA-MCC1; in general one always uses
> >>>>> INTRA-MCC1 except when one realizes that some atoms are not
> >>>>> equivalenced as they should be ;-)
> >>>>>
> >>>>> > I understand that it very much depends on one's case, but I was
> >>>>> > wondering if you had any thoughts. I was thinking that I might
> >>>>> constrain
> >>>>> > the charges of some of the sugar atoms that are 5' of the
> >>>>> transition-state
> >>>>> > pentacovalent phosphate to their values in the AMBER force fields
> >>>>> instead
> >>>>> > of summing their charges to 0.
> >>>>>
> >>>>> oh oh I got it: You continue the discussion from the private
> >>>>> assistance...
> >>>>> Not easy to follow you. In your case I would:
> >>>>> -1 characterize the TS using the QM program of your choice
> >>>>> -2 run RED Python using CHR_TYP = "RESP-C2", OPT_Calc = "Off" &
> >>>>> MEPCHR_Calc = "On" using the QM log file & PDB file as inputs
> >>>>> -3 study what you get: here you will be able to:
> >>>>> - check the atom connectivities generated in the mol3 file
> >>>>> - re-run RED Python using Re_Fit = "On" (i.e. upload/provide the
> >>>>> entire/previous RED Python job in the archive)
> >>>>> - provide new FF parameters in the frcmod.user (see the
> >>>>> frcmod.unknown file generated in the previous job) in the archive
> file
> >>>>> See
> >>>>>
> http://q4md-forcefieldtools.org/REDS-Development/RED-Server-demo1.php
> >>>>>
> >>>>>
> http://q4md-forcefieldtools.org/REDS-Development/popup/poptestcase.php
> >>>>>
> >>>>>
> http://q4md-forcefieldtools.org/REDS-Development/Demo2-Files/frcmod.user
> >>>>> - provide new FF atom types in the Project.config file for key
> atom
> >>>>> center(s)
> >>>>> MOLECULE1-ATMTYPE = ...
> >>>>> (here you can load your own FF as input to R.E.D. Python)
> >>>>>
> >>>>> When ready just request a private assistance and provide the PXXXX
> >>>>> R.E.D. Server job name in the body of your email.
> >>>>>
> >>>>> > I lastly wanted to double-check that the INTER-MCC and INTER-MEQ
> >>>>> keywords
> >>>>> > are only necessary if one is deriving charges for more than one
> >>>>> molecule,
> >>>>> > and so would not be applicable in my one-omolecule one-conformation
> >>>>> case?
> >>>>>
> >>>>> Yes INTRA means within a molecule; INTER means between molecules.
> >>>>>
> >>>>> regards, Francois
> >>>>>
> >>>>>
> >>>>> >> regards, Francois
> >>>>> >>
> >>>>> >>
> >>>>> >> > On Sat, Dec 7, 2013 at 4:30 PM, FyD <
> fyd.q4md-forcefieldtools.org>
> >>>>> >> wrote:
> >>>>> >> >
> >>>>> >> >> Dear Kamali,
> >>>>> >> >>
> >>>>> >> >> > I am using the RED Server Development to parameterize a model
> >>>>> compound
> >>>>> >> >> for
> >>>>> >> >> > the transition state and different protonation states of one
> or
> >>>>> both
> >>>>> >> of
> >>>>> >> >> the
> >>>>> >> >> > nonbridging oxygens in ribozyme cleavage.
> >>>>> >> >>
> >>>>> >> >> Classical geometry optimization as implemented in R.E.D.
> >>>>> will likely
> >>>>> >> >> generate a true minimum; not a TS.
> >>>>> >> >>
> >>>>> >> >> So better first characterizing the TS & loading the
> >>>>> corresponding QM
> >>>>> >> >> log file as input by setting OPT_Calc = "Off" & MEPCHR_Calc =
> "On"
> >>>>> in
> >>>>> >> >> the Configuration.py file
> >>>>> >> >>
> >>>>> >> >> > My questions are regarding a
> >>>>> >> >> > portion of the Project.config file that one may optionally
> >>>>> include to
> >>>>> >> >> > override default options (
> >>>>> >> >> >
> >>>>> >> >>
> >>>>> >>
> >>>>>
> http://q4md-forcefieldtools.org/REDS-Development/Demo2-Files/Project.config
> >>>>> >> >> ).
> >>>>> >> >> > There is a portion of the Project.config file that deals with
> >>>>> >> >> > intramolecular constraints (INTRA-MCC). My questions are:
> >>>>> >> >>
> >>>>> >> >> INTRA-MCC in R.E.D. Perl = INTRA-MCC1 in R.E.D. Python,
> >>>>> which is the
> >>>>> >> >> classical way to set up intra-molecular charge constraint...
> >>>>> >> >> See Cieplak et al. in
> >>>>> >> >>
> >>>>>
> http://www3.interscience.wiley.com/cgi-bin/abstract/109583237/ABSTRACT
> >>>>> >> >>
> >>>>> >> >> > - How does one decide what value to assign for
> intramolecular
> >>>>> >> >> > constraints?
> >>>>> >> >>
> >>>>> >> >> In general when one wants to remove a part of a whole molecule
> to
> >>>>> >> >> generate a molecular fragment the intra-mcc total charge
> >>>>> value takes
> >>>>> >> >> an integer value (& this integer value is very often zero).
> >>>>> >> >>
> >>>>> >> >> See also a discussion about the relation between the chemical
> group
> >>>>> >> >> involved in the constraint & the total charge of this
> constraint:
> >>>>> >> >> http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2918240/
> >>>>> >> >>
> >>>>> >> >> > - How does one decide whether or not to keep the
> >>>>> intra-molecular
> >>>>> >> >> charge
> >>>>> >> >> > constraints through the second stage of fitting (i.e.,
> >>>>> INTRA-MCC1
> >>>>> >> vs.
> >>>>> >> >> > INTRA-MCC2)?
> >>>>> >> >>
> >>>>> >> >> In general you want to use the way defined by Cieplak et al.
> i.e.
> >>>>> >> >> INTRA-MCC1
> >>>>> >> >> See
> >>>>> >>
> http://www3.interscience.wiley.com/cgi-bin/abstract/109583237/ABSTRACT
> >>>>> >> >>
> >>>>> >> >> run both types of intra-mcc on the CH3CO group in a dipeptide
> model
> >>>>> to
> >>>>> >> >> study the differences; in short an intra-mcc1 is only
> >>>>> applied in the
> >>>>> >> >> 1st resp input, while intra-mcc2 is applied in both resp
> >>>>> inputs; the
> >>>>> >> >> later allows charge equivalencing for CH2 and/or CH3 groups
> when
> >>>>> >> >> involved in the constraint...
> >>>>> >> >>
> >>>>> >> >> > If my questions are trivial ones, please direct me to the
> >>>>> appropriate
> >>>>> >> >> > references.
> >>>>> >> >>
> >>>>> >> >> no trivial question - do not worry ;-)
> >>>>> >> >>
> >>>>> >> >> R.E.D. Python is not yet published - still coding new
> features...
> >>>>> >> >> R.E.D. Perl is published:
> >>>>> >> >> http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2918240/
> >>>>> >> >>
> >>>>> >> >> regards, Francois
>
>
>
> _______________________________________________
> 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 Fri Dec 13 2013 - 07:00:03 PST
