A bug documented on the reflector in 2003 does not seem to have been fixed in
Amber 8
When you source leaprc.ff99 from leap, these commands are processed:
#
# Load the main parameter set.
#
parm99 = loadamberparams parm99.dat
....
#
# Load main chain and terminating
# amino acid libraries.
#
loadOff all_amino94.lib
loadOff all_aminoct94.lib
loadOff all_aminont94.lib
While parm99.dat has correct torsion terms:
X -CA-OH-X 2 1.80 180.0 2. Junmei et al, 99
X -C -OH-X 2 4.60 180.0 2. Junmei et al, 1999
the all_amino*94.lib files have Tyrosine still parameterized as type C instead
of CA. Thus, tyrosine is too rigid under simulation.
This was better explained by others back in 2003:
http://amber.ch.ic.ac.uk/archive/200305/0180.html
I'm rementioning it because it might impact simulation results, and it really
should be fixed.
I can understand reluectance to change the parm*.dat and the all_amino*.lib
files, as "fixing" them in one place might botch their user in other
combinations. But, the leaprc.ff99 appears to definitely give a "bad" result.
Perhaps TYR should just be redefined in that file, going forward.
Thanks
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Thu Sep 01 2005 - 23:53:00 PDT
