
|
ROTAMER (CCP4: Supported Program)
NAME
rotamer - list amino acids whose side chain torsion angles
deviate from the "Penultimate Rotamer Library"
SYNOPSIS
rotamer xyzin foo_in.pdb
[Keyworded input]
This program reads a protein coordinate file in PDB format and lists
all amino acids whose side chain torsion angles deviate more than a
user-defined threshold from the rotamers of the "Penultimate
Rotamer Library" (ref. 1, Table I,
"common-atom values").
An amino acid is listed if at least one side chain torsion angle deviates
more than a user-defined threshold DELT from the
equivalent one of the nearest rotamer.
Empirically, the following two-step procedure to find the nearest
rotamer seems to work quite reasonably, producing only a few false
negatives or false positives: The first step is to choose a
subset of potential nearest rotamers in a hierarchical order going from
the first side chain torsion angle Chi(1) to the last side chain
torsion angle Chi(n). For a side chain with four Chi-angles, first,
all rotamers deviating less than the user-defined threshold in Chi(1) are
chosen, then of those all rotamers that deviate less in
Chi(2) are chosen, then of those all rotamers that deviate less in
Chi(3) are chosen, and finally of those all rotamers that deviate less
in Chi(4) are chosen. If a subset for Chi(i) is empty, the subset for
the previous Chi(i-1) is chosen and the hierarchical search is
terminated. If the subset for Chi(1) is empty, simply all rotamers are
chosen as a "subset". The second step is to find the nearest
rotamer out of the previously chosen subset of rotamers. Here, the nearest
rotamer is the one with the minimum sum over the absolute Delta-Chi(i)
values.
Some remarks and hints
A reasonable threshold to start with is 30°, which is also the
default. In my experience, the deviation in Chi(1) is most important.
So, watch out for amino acids that deviate in Chi(1). For amino acids
with longer side chains, it is worth to look at residues with at least
two torsion angle deviations in the printer
output. This could be a hint that an alternative (rotameric)
conformation might fit the electron density as well. Remember, that
even with current maximum-likelihood refinement programs the electron
density maps are still somewhat biased towards the current model. So, even if
an alternative conformation doesn't fit the electron density map
perfectly, it might be worth to try a refinement of that conformation and
inspect the electron density map again.
For proline residues, it would be probably much better to check for
possible ring pucker modes as done in the program WHAT_CHECK (ref. 2). Here, only the Chi(1) value of proline residues
is checked.
Known limitations:
- All necessary coordinates for calculating side chain torsion angles must
be provided, otherwise the program issues a warning and skips the residue.
- The output side chain torsion angles for residues with multiple
side chain conformations depend on the order of the atoms: the
last atoms in the coordinate file of all alternate conformations determine the
values. For example, the following order of atom
records in a valine residue with two alternate side chain conformations
will result in the output of the side chain torsion angle of conformation B:
ATOM 650 CG1AVAL C
444 4.997 -1.309
3.806 0.60 19.67
ATOM 651 CG1BVAL C
444 2.959 -1.564
5.384 0.40 29.48
ATOM 652 CG2AVAL C
444 3.111 -1.499
5.347 0.60 12.83
ATOM 653 CG2BVAL C
444 5.395 -2.257
5.300 0.40 12.68
whereas the following order of atom records will produce nonsense:
ATOM 651 CG1BVAL C
444 2.959 -1.564
5.384 0.40 29.48
ATOM 650 CG1AVAL C
444 4.997 -1.309
3.806 0.60 19.67
ATOM 652 CG2AVAL C
444 3.111 -1.499
5.347 0.60 12.83
ATOM 653 CG2BVAL C
444 5.395 -2.257
5.300 0.40 12.68
-
XYZIN
-
Input coordinates in PDB format.
List all amino acids with at least one side chain torsion angle
deviating more than |threshold| from its nearest rotamer. If this
keyword is not supplied or if its value is outside the allowed range of
[0,180]°, the default value of 30° is taken.
(Optional) Specifies the end of keyworded input and starts ROTAMER
running.
List all amino acids with at least one side chain torsion angle
deviating more than 40° from its nearest rotamer:
rotamer XYZIN foo_in.pdb << eof
delt 40
end
eof
Unix examples script found in $CEXAM/unix/runnable/
rotamer.exam
Each amino acid is listed with its chain name, residue number, residue
name, a flag, and the observed Chi(i) values followed by the Chi(i)
values of the nearest rotamer in brackets. The output flag has a "*"
in column "i" if the absolute value of Delta-Chi(i) is
greater than the user-defined threshold, and a "o" otherwise.
In the example below of a partially refined protein, the two residues A444 Val and
C450 Arg were
intentionally modelled with wrong side chain torsion angles into the
electron density map. The residues A461 Lys and B461 Lys are very
poorly defined in the electron density map.
Rotamer list with |Delta-Chi| = 30
Residue Flag
Chi1
Chi2
Chi3 Chi4
===================================================================
A 434 GLU oo* -71 ( -67) 174 (
180) -74 ( -10)
A 444 VAL * 104 ( 63)
A 445 GLN oo* -79 ( -67) 178 (
180) 122 ( -25)
A 446 LYS ooo* -71 ( -67) -179 ( 180) 178 (
180) -145 ( 180)
A 448 GLN oo* -80 ( -67) 174 (
180) 96 ( -25)
A 453 ARG ooo* -73 ( -67) 171 ( 180) -169 (
180) -143 ( 180)
A 461 LYS *ooo -146 (-177) 83 ( 68)
-177 ( 180) 179 ( 180)
B 445 GLN oo* 178 (-177) 180 ( 180)
-121 ( 0)
B 448 GLN oo* -75 ( -67) -177 (
180) 83 ( -25)
B 461 LYS oo** -60 ( -67) -176 ( 180) 148 (
180) 145 ( 180)
C 432 ARG ooo* 175 (-177) 71 (
65) -161 ( 180) -88 ( 180)
C 445 GLN oo* -73 ( -67) -177 (
180) 62 ( -25)
C 448 GLN oo* -175 (-177) 161 ( 180) -103 (
0)
C 450 ARG o*** -72 ( -67) -135 ( 180) -101 (
-65) 138 ( 105)
-
Lovell, S.C., Word, J.M., Richardson, J.S.
& Richardson, D.C. "The penultimate rotamer library",
Proteins: Structure, Function and Genetics, Vol.40, 389-408 (2000).
-
Hooft, R.W.W., Vriend, G., Sander, C.
& Abola, E.E. "Errors in protein structures", Nature,
Vol. 381, 272 (1996).
AUTHORS
Dirk Kostrewa, Paul Scherrer Institute, Villigen, March 2002
SEE ALSO
Rotamer tables
|