AMBER Archive (2004)

Subject: AMBER: Problem with divcon - increasing MXATTS?

From: steven.j.enoch_at_gsk.com
Date: Wed Jul 14 2004 - 10:33:38 CDT

Amber users,

I am trying use antechamber to create prep (and frcmod) files for a large
ligand (126 atoms). It is essentially two protien chains linked together
by an organic backbone. When I run antechamber I get the following error
message:

CARTESIAN AM1 STANDARD DIRECT OPT=BFGS XTEST=0.0001 CHARGE=-3

created by wmopcrt()

 CARTESIAN - CARTESIAN COORDINATE FORMAT SELECTED
 AM1 - AM1 HAMILTONIAN TO BE USED
 STANDARD - STANDARD CLOSED-SHELL CALCULATION (NO D&C)
 DIRECT - DO NOT STORE 2-ELECTRON INTEGRALS IN SCF CALCULATIONS
 OPT=BFGS - CARRY OUT A BFGS GEOMETRY OPTIMIZATION
 XTEST=0.0001 - GEOMETRY OPTIMIZATION COORDINATE CHANGE CRITERION =
0.00010
 CHARGE=-3 - A NET CHARGE OF -3.00 TO BE PLACED ON SYSTEM

 NUMBER OF ATOMS = 124
 TOTAL NUMBER OF BASIS FUNCTIONS = 313
 TOTAL NUMBER OF ELECTRONS = 352

 DEFAULT SETTINGS FOR CALCULATION
 --------------------------------

 - D&C TEMPERATURE OF 1000.00 K WILL BE USED
 - NO TIME LIMIT ON CALCULATION
 

 ETEST = 0.002000
 XTEST = 0.000100
 GNTEST = 2.410913
 GTEST = 0.500000

 NUMBER OF VARIABLES= 372

 UPDATING PAIRLIST
 NPAIRS = 7626

  CYCLE = 0 TIME = 47.565 ENERGY = -405.624288
  GNORM = 473.513 GRDMAX = 114.8450 GRDAVR = 17.6764
 DELTAE = 0.000000 DELTAX = 0.000000

  MAXIMUM NUMBER OF ATOMS IN BFGS SEARCH EXCEDEED

  Max Degree = 180 NPAR= 372

  INCREASE MXATTS IN divcon.dim AND RECOMPILE

 FINAL QUANTITIES:
 -----------------

 NUMBER OF FULLY DIAGONALIZED SUBSYSTEMS: 1
 NUMBER OF FROZEN SUBSYSTEMS: 0
 NUMBER OF SCF CALCULATIONS = 1

 ELECTRONIC ENERGY = -143107.91168290 EV
 CORE-CORE REPULSIONS = 130880.02283529 EV
 TOTAL ENERGY = -12227.88884761 EV
 HEAT OF FORMATION = -405.62428815 KCAL/MOL
 FERMI ENERGY = 0.00000000 EV

I have tried increasing MXATTS to 150 (and then 400) and then recompiled
divcon, but I still get the same error message. Does anybody have any
advice on how to get around this problem? Or can divcon not deal with such
a large ligand?

Many thanks

Dr Steve Enoch
GlaxoSmithKline
UK

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu