AMBER Archive (2001)

Subject: ANAL (Amber6.0)

From: Aldo Jongejan (jongejan_at_chem.vu.nl)
Date: Wed May 02 2001 - 08:01:31 CDT

Dear users,

I have been playing around now with ANAL for some time
in order to get lists of interaction energies of specific
amino acids with my ligand/substrate. I tried to use
the GROUP format to put every residue in a separate
group (RES -resNr1 resNr2), but this allready failed.
After inspection of the source code this seemed to be
due to a programming 'error' in the ./lib/rgroup.f code: 

----
----
C         ----- 1ST GROUP OF 1ST CARD MUST BE - IF ANY - NUMBERS ARE
C               FOUND -----
C
c original code:      IF(ITIME.EQ.1.AND.IGRP(1).LT.0) LSIGN = 1
         IF(ITIME.EQ.2.AND.IGRP(1).LT.0) LSIGN = 1
          DO 12 I = 1,7

----
----

I am not a (real) programmer, so if somebody wants to take 
a closer look at this, I would be most happy. I just had the
program write its variables and saw the ITIME variable
had to be increased in order to enter the loop. Anyway, it
seems to work....

A closely related subject is the discrepancy between the
energies outputted by ANAL and SANDER. It still seems to be
an 'un-fixed bug'. I am sorry to say, I didn't solve it either.
I have just noticed, that the values you enter for DIELC and
SCEE are very important (and as was already reported, the value
for the cut-off distance!). Still, energy differences are to
be found.
However, what I find particularly frustrating is the fact that
the energies printed for the INITIAL ENERGY COMPONENTS and the
ENERGY CONTRIBUTION BY GROUPS differ (notably the energies
involved in the non-bonded interactions). Obviously, (as I noticed
reading the source code) these are calculated using different
protocols. Documentation is rather poor, and I don't understand
why different protocols are used. Now, it looks just like the program
can't add the numbers right!!! :-)

Anyway, is the ANAL program still used? and if so, whatfor?
Can interaction energies (residue based!) be calculated using
other methods in a fast way? And finally, what numbers should
be trusted, those printed by SANDER or by ANAL....??

with kind regards,

aldo

-- 
###########################################

Aldo Jongejan
Molecular Modeling Group
Dept. of Pharmacochemistry
Free University of Amsterdam
De Boelelaan 1083
1081 HV Amsterdam
The Netherlands

e-mail: jongejan_at_chem.vu.nl
tlf:    +31 (0)20 4447612
fax:    +31 (0)20 4447610

###########################################