Hello there,
I have obtained ~12 kcal/mol difference in energies from different ways of
minimizing a small organic compound (20 atoms) with cut=999.0. The two
minimization protocols differ only in their specified NSNB values (25
versus 20000). The resulting, minimized structures are in different
conformations, as you might expect. I don't understand why updating the
nonbonded pairlist periodically vs. no updating would have any effect on
the conformational sampling for a small system like mine. Have any of you
encountered such cases before?
My two minimization protocols along with their resulting energies are
provided below:
=============
Protocol #1:
=============
&cntrl
imin=1, maxcyc=1000, ntpr=50, scee=1.2, nsnb=25, cut=999.0,
&end
-------------
Results (min.info file):
NSTEP ENERGY RMS GMAX NAME NUMBER
1000 -5.2065E+01 9.3692E-03 2.8413E-02 C4 4
BOND = 0.6229 ANGLE = 10.1613 DIHED =
35.3476
VDWAALS = -1.2898 EEL = 77.0412 HBOND =
0.0000
1-4 VDW = 3.7663 1-4 EEL = -177.7250 CONSTRAINT =
0.0105
EAMBER = -52.0754
=============
Protocol #2:
=============
&cntrl
imin=1, maxcyc=1000, scee=1.2, ntpr=50, nsnb=20000, cut=999.0,
&end
-------------
Results (min.info file):
NSTEP ENERGY RMS GMAX NAME NUMBER
718 -6.4619E+01 8.0970E-05 2.2583E-04 C2 2
BOND = 0.3891 ANGLE = 6.6194 DIHED =
18.1204
VDWAALS = -0.2256 EEL = 72.9805 HBOND =
0.0000
1-4 VDW = 3.7723 1-4 EEL = -166.4282 CONSTRAINT =
0.1534
EAMBER = -64.7722
Lillian
----------------------------------------------------------------------------------------------------
Lillian Chong lab: (415) 476-7985
Kollman Group fax: (415) 502-1411
University of California
513 Parnassus Avenue
San Francisco, CA 94143-0446
Received on Wed Mar 28 2001 - 10:59:15 PST
