Dear All,
I met a problem when using Xplor PARArestraints xpcs module.
The script I use is based on the sampe input file downloaded from
www.postgenomicnmr.net, and revised as the following (see the bottom).
Briefly speaking, I generate the psf file for the protein plus ANI, with or
without the file for cobalt.
The problem is that the the cobalt tensor parameters $chiax and $chirh should
converge after several iterations (50 structures x 5 interations), but my
results show that they don't. Something like this:
saCONV_1_10.pdb:REMARK xpcs coefficient.: 9480.79 -6012.09
saCONV_1_11.pdb:REMARK xpcs coefficient.: 13515 -7734.1
saCONV_1_12.pdb:REMARK xpcs coefficient.: 9331.34 -6128.62
saCONV_1_13.pdb:REMARK xpcs coefficient.: 26393.5 -17358.1
.....
saCONV_2_11.pdb:REMARK xpcs coefficient.: 2.389801E+06 -1.571383E+06
saCONV_2_12.pdb:REMARK xpcs coefficient.: 173188 -111816
saCONV_2_13.pdb:REMARK xpcs coefficient.: 2.043607E+06 -1.352478E+06
saCONV_2_14.pdb:REMARK xpcs coefficient.: 6.396403E+06 -4.250398E+06
....
saCONV_3_10.pdb:REMARK xpcs coefficient.: 2210.77 -1206.25
saCONV_3_11.pdb:REMARK xpcs coefficient.: 1200.21 -324.95
saCONV_3_12.pdb:REMARK xpcs coefficient.: 1338.83 -793.126
saCONV_3_14.pdb:REMARK xpcs coefficient.: 1928.36 -508.823
....
saCONV_4_10.pdb:REMARK xpcs coefficient.: 916869 -602598
saCONV_4_11.pdb:REMARK xpcs coefficient.: 56444.6 -36374.3
saCONV_4_12.pdb:REMARK xpcs coefficient.: 36311.4 -20406.6
saCONV_4_13.pdb:REMARK xpcs coefficient.: 80822.9 -50343.3
....
saCONV_5_10.pdb:REMARK xpcs coefficient.: 1914.45 -802.257
saCONV_5_11.pdb:REMARK xpcs coefficient.: 1782.43 -935.74
saCONV_5_12.pdb:REMARK xpcs coefficient.: 949.112 -348.363
saCONV_5_13.pdb:REMARK xpcs coefficient.: 2037.34 -923.342
....
In some interations (saCONV_2_*.pdb, saCONV_4_*.pdb, etc), the absolute value
of $chiax and $chirh goes extremely large, but the output structures seem ok,
with normal energy terms and coordinates. In other interations
(saCONV_3_*.pdb, saCONV_5_*.pdb, etc), their range look normal, but, in such
iterations, there is no noe or xpcs violations, the enegy terms are very low
(overall energy ~10^-2, noe energy ~10^-2), and the structure calculation
obviously fails, since in the output pdb file, the atom coordinates are like:
ATOM 1 CA ALA 1 -5.594 3.278 2.303 1.00 0.00 PDZ2
ATOM 2 HA ALA 1 604.428 354.780 285.185 1.00 0.00 PDZ2
ATOM 3 CB ALA 1 -7.131 2.805 1.318 1.00 0.00 PDZ2
ATOM 4 HB1 ALA 1 603.126 354.814 283.656 1.00 0.00 PDZ2
ATOM 5 HB2 ALA 1 601.765 353.323 285.126 1.00 0.00 PDZ2
ATOM 6 HB3 ALA 1 604.055 353.268 284.413 1.00 0.00 PDZ2
ATOM 7 C ALA 1 -4.757 1.565 3.007 1.00 0.00 PDZ2
ATOM 8 O ALA 1 -4.389 1.031 3.855 1.00 0.00 PDZ2
ATOM 9 N ALA 1 -4.755 4.462 2.234 1.00 0.00 PDZ2
ATOM 10 HT1 ALA 1 -4.673 4.437 3.255 1.00 0.00 PDZ2
ATOM 11 HT2 ALA 1 -5.589 4.552 1.511 1.00 0.00 PDZ2
ATOM 12 HT3 ALA 1 -4.115 4.075 1.669 1.00 0.00 PDZ2
ATOM 13 N GLU 2 -4.614 0.980 1.619 1.00 0.00 PDZ2
ATOM 14 HN GLU 2 606.603 356.257 286.373 1.00 0.00 PDZ2
ATOM 15 CA GLU 2 -3.806 0.003 1.743 1.00 0.00 PDZ2
.............
I tried several ways of calculation, with or without the cobalt psf and pdb
file, with or without the initial setting of xpcs coefficient, and with
different force constant from 2 to 5. However, it makes no difference. I
don't know whether the problem lies in the xplor scripts I used, or the
paramagnetic restraints table, or the molecular topology file, or
something..... Could anyone give me some advice and help? Thanks a lot.
The xplor script I used:
===================================
evaluate ($init_t = 1000)
evaluate ($high_steps = 12000)
evaluate ($cool_steps = 6000)
parameter
@TOPPAR:protein-1.0.par
@TOPPAR:axis-1.0.par
end
topology
@TOPPAR:protein-1.0.top
@TOPPAR:top_axis.pro
MASS CO 58.9332 ! Cobalt
residue COB
atom CO type=CO charge=2.0 end
end
end
structure @./reference/pdz2_ani_template.psf end
structure @./reference/cobalt.psf end
coordinates @./reference/pdz2_ani_template.pdb
coordinates @./reference/cobalt.pdb
noe
nres=5000
class tensor
@./restraints/tensors.tbl
class 1
@./restraints/pdz2_nnoe.tbl
end
restraints dihedral
reset
@./restraints/pdz2_dihed.tbl
end
flags exclude * include bonds angle impr vdw noe xpcs end
vector do (fbeta=10) (all)
vector do (mass=100) (all)
noe
ceiling=1000
averaging * cent
potential * soft
scale * 50.
sqoffset * 0.0
sqconstant * 1.0
sqexponent * 2
soexponent * 1
asymptote * 0.1
rswitch * 0.5
end
xpcs
nres=2000
class cobalt
force 2
coeff 1600 -500
@./restraints/pcstable_pdz2_3.tbl
end
parameter
nbonds
repel=1.
rexp=2 irexp=2 rcon=1.
nbxmod=3
wmin=0.01
cutnb=4.5 ctonnb=2.99 ctofnb=3.
tolerance=0.5
end
end
set abort off end
evaluate ($end_count=50) { 50 structures per step }
coor copy end
evaluate ($count1 = 0)
while ($count1 < 5 ) loop media { 10 steps }
evaluate ($count1=$count1+1)
evaluate ($count = 0)
while ($count < $end_count ) loop main
evaluate ($count=$count+1)
coor swap end
coor copy end
noe
potential * soft
end
{* ============================================ Initial minimization.*}
.............
{* ======================================= High-temperature dynamics.*}
.............
{* ============= Tilt the asymptote and increase weights on geometry.*}
.............
{* ================================================ Cool the system.*}
.............
{===============================================>} {*Abort condition.*}
............
{* ============================================== Final minimization.*}
............
xpcs
foff
son
evaluate
($filename="./pdz2/test16/1pse"+encode($count1)+"_"+encode($count))
save $filename
frun 1
soff
end
{* ================================ Write out the final structure(s).*}
...........................
end loop main
xpcs
fmed 10 1
class cobalt
coeff $chiax $chirh
fmed 1 0
end
end loop media
xpcs
erron 10 35
frun 1
end
stop
====================================
Thanks,
Victor
