[AMBER] Paramfit issues

From: David Poole <thepoole.ucdavis.edu>
Date: Mon, 14 Sep 2015 12:31:57 +0200

Hello Everybody,

I'd been continuing my effort to produce some specialized heme parameters
and decided that it might be the simplest to use paramfit and a large
number of QM derived structures+energies. But after the QM is done, I've
questions.


1) During the optimization process Gaussian produces a number of structures
with energies that are sub-optimal configurations, is there any reason why
these cannot or should not be used for parameter fitting? or other concerns
in that regard?

2) Is there a way to run paramfit on multiple processors?

3) I derived the correction constant K using the method described in the
tutorial, and then ran it to find parameters by list as well as the default
(since I want to optimize all parameters), in both circumstances it set the
initial value of K to zero, this obviously deviates from the output files
in the tutorial.. so I probably need help with this.

4) I am getting some warning related to dihedral spans lacking data
coverage, given that Is fitting dehedral parameters to the heme system
unnecessary or is this not a significant message?

5) Any other pointers would be greatly appreciated.

Thank you,
-David Poole

Here are the relevant input and output files, the latter is truncated at
points just for space on the list messages.


fit_parameters.in
RUNTYPE=FIT
COORDINATE_FORMAT=TRAJECTORY
NSTRUCTURES=1001
K=-1922732.27618371
PARAMETERS_TO_FIT=DEFAULT
FUNC_TO_FIT=SUM_SQUARES_AMBER_STANDARD
QM_ENERGY_UNITS=HARTREE

ALGORITHM=GENETIC
OPTIMIZATIONS=500
MAX_GENERATIONS=50000
GENERATIONS_TO_CONV=20
GENERATIONS_TO_SIMPLEX=10
GENERATIONS_WITHOUT_SIMPLEX=10
MUTATION_RATE=0.075
PARENT_PERCENT=0.25
SEARCH_SPACE=1.0

WRITE_ENERGY=fit_parameters_04_energy.dat
WRITE_FRCMOD=fit_parameters_04.frcmod

fit_parameter.out
                * AMBER Bond Angle and Dihedral Parameter Optimiser *
                * v3.0.0 *
                                Execution started at:
| Mon Sep 7 11:38:13 2015
|
| Random seed = 1441651093
 Reading job control file: fit_parameters.in
 Job Control: Read a total of 23 lines from job_control file. 18 options
set.

  Reading prmtop file : fixed_2.prmtop
   Prmtop (unique): Found 13 unique bonds.
   Prmtop (unique): Found 22 unique angles.
   Prmtop (unique): Found 42 unique dihedrals.
   Prmtop (unique): Found 44 unique dihedral terms.
 Reading mdcrd file: all_log_structures.mdcrd
  Reading mdcrd file : all_log_structures.mdcrd
    Coordinate file passed format check
  Reading energy file or directory : all_log_energies.dat

---------------------------------------------------------------------
        | OPTIONS SUMMARY |
        | ---------------
 |
        | Summary of Run Type Options:
 |
        | Run Mode = FIT, Minimiser = GENETIC
|
        | Function to be Fit = SUM_SQUARES_AMBER_STANDARD
|
        |
|
        | Terms to be fit:
 |
        | UNIQUE_BONDS = 13, UNIQUE ANGLES = 22, UNIQUE DIHEDRALS = 42
 |
        | NBONDS = 91, NANGLES = 168, NDIHEDRALS = 320
 |
        | Total dimensions of fit = 114
|
        |
|
        | Sample structures for least squares fit = 1001
 |
        |
|
        |
|
        | Energy Correction Term (K) = 0.00000000
|
        |
|
        | OPTIMIZATIONS = 500
 |
        | MAX GENERATIONS = 50000
 |
        | MUTATION RATE = 0.07
 |
        |
|
        | Estimate Memory Usage (per cpu):
 |
        | Coordinate info will be read from disk as required.
 |
        | OPTION STORAGE = 131 bytes
|
        | PRMTOP STORAGE = 874562 bytes
 |
        | COORDINATE STORAGE = 2074112 bytes
|
        | GA ARRAY STORAGE = 456000 bytes
 |
        |
|
        | TOTAL ESTIMATED MEMORY USAGE = 3404805 bytes
|
        |
|

---------------------------------------------------------------------

   WARNING: h4-c2-na-hn dihedral is missing 9 data points in the range
0.0000 to 3.1416 radians.
 <Truncated>
   WARNING: c3-c2-c2-c3 dihedral is missing 9 data points in the range
0.0000 to 3.1416 radians.
 * Input structures passed dihedral span check.

   ------------------------------- INITIAL PARAMETERS
--------------------------
------
   Parameters for force field equation: AMBER_STANDARD:
   (* means parameter is NOT constant during fit)
                         K = 0.000000 kcal/mol
             (c2-h4)*Kr = 348.6000 kcal/(mol A)^2,*r_eq = 1.0840 A
<Truncated>
   IMP (c2-c2-c2-c3)*Kp = 1.1000 kcal/mol, Np = 2.0000, Phase =
180.0001 Deg


 -----------------------------------------------------------------------------
------
   Sum of squares for initial parameters = 3700596316869135.0000000000
kcal^2/mo
l^2
   R^2 value for initial parameters = 0.588313
   Calculated energy with initial parameters for structure 1 = 372.804549
KCal/m
ol
   Actual energy for structure 1 should be =
-1922382.389732 K
Cal/mol

   --------------------------GENETIC ALGORITHM MINIMISATION
--------------------
-------
   Minimising function SUM_SQUARES_AMBER_STANDARD, using the GENETIC
ALGORITHM
   Running SIMPLEX REFINEMENT every 10 converged gen, then break for 10

 -----------------------------------------------------------------------------
--------
   GENERATIONS_TO_CONVERGE = 20 MAX_GENERATIONS = 50000 CONV_LIMIT =
1.00E
+01
   OPTIMIZATIONS = 500 PARENT_PERCENT = 0.25
   SEARCH_SPACE = 1.00 MAX_GENERATIONS = 50000
   ------------------------------------ CONVERGENCE
----------------------------
--------
   Gen 0: Best= 3737064547674845.50000 Mean= 4.3387e+16 Elapsed=

  0/3
   Gen 1: Best= 3734895501632929.00000 Mean= 3.8512e+15 Elapsed=

  1/3
   Gen 2: Best= 3722321266414401.00000 Mean= 3.7601e+15 Elapsed=

:
   R^2 value for initial parameters = 0.588313
   Calculated energy with initial parameters for structure 1 = 372.804549
KCal/mol
   Actual energy for structure 1 should be =
-1922382.389732 KCal/mol

   --------------------------GENETIC ALGORITHM MINIMISATION
---------------------------
   Minimising function SUM_SQUARES_AMBER_STANDARD, using the GENETIC
ALGORITHM
   Running SIMPLEX REFINEMENT every 10 converged gen, then break for 10

 -------------------------------------------------------------------------------------
   GENERATIONS_TO_CONVERGE = 20 MAX_GENERATIONS = 50000 CONV_LIMIT =
1.00E+01
   OPTIMIZATIONS = 500 PARENT_PERCENT = 0.25
   SEARCH_SPACE = 1.00 MAX_GENERATIONS = 50000
   ------------------------------------ CONVERGENCE
------------------------------------
   Gen 0: Best= 3737064547674845.50000 Mean= 4.3387e+16 Elapsed=
      0/3
   Gen 1: Best= 3734895501632929.00000 Mean= 3.8512e+15 Elapsed=
      1/3
   Gen 2: Best= 3722321266414401.00000 Mean= 3.7601e+15 Elapsed=
      2/3
   Gen 3: Best= 3711452850436978.50000 Mean= 3.7361e+15 Conv=
       0/20
   Gen 4: Best= 3707060632432145.00000 Mean= 3.7221e+15 Conv=
       0/20
   Gen 5: Best= 3700834433631478.50000 Mean= 3.714e+15 Conv=
       0/20
   Gen 6: Best= 3698366613824640.50000 Mean= 3.7083e+15 Conv=
       0/20
   Gen 7: Best= 3694454498694381.00000 Mean= 3.7041e+15 Conv=
       0/20
   Gen 8: Best= 3692951446031217.50000 Mean= 3.7009e+15 Conv=
       0/20
   Gen 9: Best= 3692354203638458.50000 Mean= 3.6988e+15 Conv=
       0/20
   Gen 10: Best= 3691484864346694.00000 Mean= 3.6972e+15 Conv=
       0/20
   Gen 11: Best= 3689650891060643.00000 Mean= 3.6961e+15 Conv=
       0/20
   Gen 12: Best= 3689650891060643.00000 Mean= 3.695e+15 Conv=
       1/20
   Gen 13: Best= 3689650891060643.00000 Mean= 3.6943e+15 Conv=
       2/20
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Sep 14 2015 - 04:00:03 PDT
Custom Search