AMBER Archive (2002)

Subject: AM1-BCC sensitivity to AM1 charges

From: Matthew Lee (mrlee_at_cgl.ucsf.edu)
Date: Tue Jul 30 2002 - 12:31:43 CDT

Hi--

        I am wondering if anyone has analyzed the sensitivity of
correlation between AM1-BCC charges and RESP charges to single point
AM1 vs. AM1 after geometry optimization. It seems the two approaches
for AM1 charges differ, on average, by about 10% and 0.01 charge units
on a small set of small molecules I've looked at.

        For comparison, on these same small molecules, the BCC alters
the charges, on average, by about 60% and 0.07 charge units.

        Incidentally, the way to use sybyl's mopac is rather unusual
in that it will not run unless there are 3 arguments after the STDOUT
redirect. There was some questions on this reflector awhile back
about this and I don't think a response was ever posted. The
following mopac.sh works with the Sybyl distribution of mopac:
--------------
/home/tripos/sybyl6.7/bin/mopac < $1 > $2 junk.arc junk.syb
/usr/bin/rm -f junk.arc junk.syb
--------------

        To elaborate more on my concern, here is an example of the AM1
charge differences I see between single point and geometry
optimazation approaches on Raloxifene:

        Single Geom. Absol %
ATOM Point Optimiz Change Change
C1 -0.1591 -0.1365 0.0226 14.2
C2 -0.1688 -0.2471 0.0783 46.4
C3 -0.0931 -0.0824 0.0107 11.5
C4 0.3323 0.3401 0.0078 2.3
C5 -0.0624 -0.0709 0.0085 13.6
S6 0.3472 0.5077 0.1605 46.2
C7 -0.1854 -0.2752 0.0898 48.4
C8 -0.0439 -0.0498 0.0059 13.5
C9 -0.1967 -0.1924 0.0043 2.2
O10 -0.2887 -0.2909 0.0022 0.8
C11 -0.0626 -0.0698 0.0072 11.6
C12 -0.0643 -0.0757 0.0114 17.7
C13 -0.1297 -0.1329 0.0032 2.5
C14 -0.2205 -0.2055 0.0150 6.8
C15 -0.0331 -0.0415 0.0084 25.4
C16 -0.0372 -0.0393 0.0021 5.5
C17 -0.1591 -0.1566 0.0025 1.6
C18 -0.2289 -0.2171 0.0118 5.2
C19 0.1071 0.0870 0.0201 18.8
C20 -0.1769 -0.1751 0.0018 1.0
C21 -0.2499 -0.2303 0.0196 7.8
C22 0.1063 0.0901 0.0162 15.3
O23 -0.2464 -0.2501 0.0037 1.5
C24 0.1066 0.1133 0.0067 6.3
O25 -0.2478 -0.2477 0.0001 0.1
O26 -0.1545 -0.1917 0.0372 24.1
C27 -0.0394 -0.0271 0.0123 31.2
C28 -0.1061 -0.0644 0.0417 39.3
N29 -0.1812 -0.2465 0.0653 36.0
C30 -0.1002 -0.0656 0.0346 34.6
C31 -0.1081 -0.0691 0.0390 36.0
C32 -0.1557 -0.1557 0.0000 0.0
C33 -0.1551 -0.1555 0.0004 0.3
C34 -0.1621 -0.1578 0.0043 2.7
H35 0.1416 0.1475 0.0059 4.2
H36 0.1585 0.1590 0.0005 0.3
H37 0.1381 0.1373 0.0008 0.6
H38 0.1534 0.1572 0.0038 2.5
H39 0.1335 0.1373 0.0038 2.9
H40 0.1544 0.1561 0.0017 1.1
H41 0.1421 0.1454 0.0033 2.3
H42 0.1517 0.1561 0.0044 2.9
H43 0.1325 0.1355 0.0030 2.3
H44 0.1503 0.1570 0.0067 4.4
H45 0.1422 0.1391 0.0031 2.2
H46 0.2270 0.2202 0.0068 3.0
H47 0.2248 0.2190 0.0058 2.6
H48 0.0834 0.0813 0.0021 2.6
H49 0.0881 0.0848 0.0033 3.7
H50 0.1099 0.1086 0.0013 1.2
H51 0.0686 0.0585 0.0101 14.7
H52 0.0940 0.0887 0.0053 5.7
H53 0.0550 0.0470 0.0080 14.6
H54 0.0540 0.0441 0.0099 18.3
H55 0.0955 0.0927 0.0028 2.9
H56 0.0993 0.0942 0.0051 5.1
H57 0.0805 0.0818 0.0013 1.6
H58 0.0798 0.0807 0.0009 1.1
H59 0.1000 0.0943 0.0057 5.7
H60 0.0835 0.0823 0.0012 1.5
H61 0.0759 0.0764 0.0005 0.7

Sincerely,
--Matt

*******************************************************************
Matthew Randolph Lee, Ph.D.

Computational Scientist phone: (858)410-6534
Lion Bioscience Inc. fax: (858)410-6665
9880 Campus Point Drive e-mail: matthew.lee_at_lionbioscience.com
San Diego, CA 92121
*******************************************************************