Hi,
I've found a related issue which I'd like to report. Let me first
quote a section from
http://amber.scripps.edu/doc/prep.html
'IMPROPER' Control for reading the improper torsion angles. A
proper torsion I - J - K - L has I bonded to J bonded
to K bonded to L. An IMPROPER torsion is any torsion in
which this is not the case. Improper torsions are used to
keep the asymmetric centers from racemizing in the united
atom model where all the C-H hydrogens are omitted. They
can also be used to enforce planarity. The normal case is:
J
|
K
/ \
I L
Improper I-J-K-L
where the central atom (K) is the third atom in the
improper
and the order of the other three is determined
alphabetically
by atom type and if types are the same by atom number.
According to this rule, the third atom should be the central atom and
the other three should be ordered by the atom type and then by the atom
number. But, I have found that this rule is not always followed by
leap. Here's a list of improper dihedrals for which I found
discrepancies:
residue: what leap does -- what the rule says
--------------------------------------
ARG: CD CZ NE HE -- CZ CD NE HE
HID: CG NE2 CD2 HD2 -- CG HD2 CD2 NE2
HID: ND1 NE2 CE1 HE1 -- HE1 ND1 CE1 NE2
HID: ND1 CD2 CG CB -- CB CD2 CG ND1
HIE: CG NE2 CD2 HD2 -- CG HD2 CD2 NE2
HIE: ND1 NE2 CE1 HE1 -- HE1 NE2 CE1 ND1
HIE: ND1 CD2 CG CB -- CB CD2 CG ND1
HIP: CG NE2 CD2 HD2 -- CG HD2 CD2 NE2
HIP: ND1 NE2 CE1 HE1 -- HE1 ND1 CE1 NE2
HIP: ND1 CD2 CG CB -- CB CD2 CG ND1
TRP: CD1 CE2 NE1 HE1 -- CE2 CD1 NE1 HE1
TRP: CE2 CH2 CZ2 HZ2 -- CH2 CE2 CZ2 HZ2
TRP: CG NE1 CD1 HD1 -- CG HD1 CD1 NE1
TRP: CD1 CD2 CG CB -- CD2 CB CG CD1
TYR: CD1 CZ CE1 HE1 -- CZ CD1 CE1 HE1
Of course, ordering of improper dihedrals is by no means a significant
issue practically. However, it does give trouble to someone (like me)
who is trying to port amber force field into another MD program. =)
In any case, I think this part of the leap code needs to be looked at.
Regards,
Sanghyun
PS: Trp is the only residue for which I've find inconsistency WITHIN
leap. For the other residues, leap is at least consistent although it
may not follow the rule.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Fri Mar 11 2005 - 03:53:01 PST
