|
| Chazin Home | Ca-binding Protein DB | Vanderbilt Home |
|
| Research Description | Publications | Wisdom | Search | ||
| How to contribute | About this page |
Here are detailed instructions for using GENXPK to assign peaks.
We focus on 2D peaks with special emphasis on NOE peaks for which
distances from preliminary structures may be used.
This tutorial assumes that you have GENXPK version 1.0.9 or later, and
are running it in the felix/genxpk directory with a suitable refparm or
history file for the spectrum from which the unassigned peaks originate.
Commands that you type into GENXPK follow the gx> prompt.
G Gippert May 12 1997
==========================================================================
Prepare xpk and (optional) volume files and settings.
__________________________________________________________________________
Pick peaks in FELIX and write crosspeak entity to xpk.txt.
Optional: measure volumes and write to vol.txt.
Set up to read these files in GENXPK.
gx> file xpk ../text/xpk.txt
gx> file vol ../text/vol.txt
Optional: Choose an intensity->volume conversion in GENXPK
or 0 for none. (here are the choices, number)
volume 0 off
volume 1 I
volume 2 I/M
volume 3 sign(I)(abs(I)/M)^(-1/6)
volume 4 scale*sign(I)(abs(I)/M)^(-1/6)
volume 5 scale*sign(I)(abs(I{*,/}factor)/M)^(-1/6)
Note, a non-zero choice assumes that the volume file name has
been defined and written from FELIX.
gx> vol 4
If you chose vol 4 or 5, you need to set the distance scale to
produce a reasonable intensity->distance conversion.
gx> scale 50 (will be different for each NOE spectrum)
Start with vol 0 until the rest of this procedure is clearly working.
--------------------------------------------------------------------------
Prepare assignment parameter settings.
__________________________________________________________________________
Determine current assignment parameters in GENXPK.
gx> sap
You should see something like this list
# dim 1 tol 0.05 atm hx
# dim 2 tol 0.05 atm hx
# asg_typ any (ignore)
# asg_cor none
# asg_ali 0
# asg_dfi is "cat"
# tasfile temp.asg
which describes a chemical shift tolerance of +-0.05 ppm and
an atom name filter of "hx" for both dimensions 1 and 2. "hx"
means "all protons in the assignments list". To change these
settings use GENXPK commands like these (examples given for
dimension 1 settings):
gx> sap tol 1 0.05
gx> sap atm 1 hx
For 2D NOE, set the asg_cor (assign correlation) setting to
none; this flag is used to correlate proton-heteronuclei for
3D NOE experiments involving one pair of (1H,15N) or (1H,13C)
HSQC-like dimensions.
Ignore the asg_typ setting which doesn't do anything.
A dfi (distance filter) setting of "cat" means that a distance
filter is NOT applied to candidate assignments based on chemical
shift. Other possibilities are described below.
All the sap parameter settings are stored in the history file
when you exit GENXPK, so you don't need to reset them next time
you work with the same spectrum. Write them down anyway.
You need to write the temporary assignments file once at the
beginning of the assignment session, and if you change the
atom name filter or change the assignments.
gx> sap write
--------------------------------------------------------------------------
Select a distance filter.
__________________________________________________________________________
The "distance filter" is a Unix command used to filter candidate
crosspeaks. A filter of "cat" does nothing, i.e., provides for no
additional pruning of candidate assignments other than by chemical
shift.
There are two ways to set the "distance filter". The first uses
the "sap dfi" command.
gx> sap dfi cat
Another way to set the distance filter is to examine the list of
current definitions and to choose one. New distance filters can be
defined by the user. Properties of the distance filters are described
at the end of this document.
gx> dfi
In addition to command syntax for dfi, you should see a list that
looks something like this (genxpk-1.0.9 and later):
dfi 0 cat
dfi 1 $GENXPK/gx_dfilter file.pdb 5.5
dfi 2 $GENXPK/gx_bfilter 1 2 distance.out 4.5 5.0
Select one of these.
gx> dfi 0
--------------------------------------------------------------------------
Case 0: assignment by chemical shift alone.
__________________________________________________________________________
Select the distance filter of "cat" by one of the above methods.
gx> dfi 0
Read the xpk file.
gx> rxp
For each newly-picked peak you should see comments like the following:
# Unassigned peak in dimension 1 at 8.29 ppm
# Unassigned peak in dimension 2 at 5.023 ppm
Invoke the assignment command.
gx> asg
The gx_asg program attempts to assign each peak one by one on the basis
of chemical shift matching (and distance filter), and writes the
candidates to a temporary file called ASG_TEMP. The interactive
version of gx_asg then opens the "vi" editor on this file. Your job is
to further eliminate poor assignment candidates by deleting those lines
in the file. Then, save your work by exiting the vi session with :wq (or
:q if no changes were made/necessary).
Ideally there will be only one good choice remaining, although
you should keep all real possibilities.
You will get a "vi" session for each unassigned peak in the input.
The results of all vi sessions accumulate to a file called
ASG_RESULTS.
--------------------------------------------------------------------------
Case 1: Single-structure distance filter.
__________________________________________________________________________
Select or define a distance filter involving gx_dfilter.
gx> dfi 1
The pre-defined distance filter 1 reads a file called "file.pdb" in the
GENXPK startup directory, and reports matching H-H distances less than
a specified cutoff (units of Angstrom). Either copy your coordinate
file or create a link to your file.
You can define a customized distance filter "myfilter" using gx_dfilter.
(Name your distance filter any string other than "add", "def" or "del".)
gx> dfi add myfilter $GENXPK/gx_dfilter /my/path/my/file 4.5
Select it by name.
gx> dfi myfilter
Read the xpk file and invoke the assignment command as in Case 0.
--------------------------------------------------------------------------
Case 2: Distance bounds filter (for a family of preliminary or homologous
structures).
__________________________________________________________________________
Select or define a distance filter involving gx_bfilter.
gx> dfi 2
The pre-defined distance filter 2 reads a file called "distance.out" in
the GENXPK startup directory and prints lines that match atom names for
the candidate assignment, and distance extrema (minimum and maximum) that
satisfy specified limits (units of Angstrom).
$GENXPK/gx_bfilter 1 2 distance.out 4.5 5.0
You can add a new distance filter definition (as described above in
Case 1) using your own file name and/or threshold values. See $GENXPK/
gx_bfilter for more details if necessary.
The distance.out file is produced using the GAP distance program (version
0.9 or later). Here is how:
ls my*.pdb > family
distance -def > distance.out
The GAP suite comes with documentation for the distance program. You may
need to add entries to the pseudomap file in case you use a different
atom nomenclature than I do. The pseudomap file is used to identify
equivalent atoms for r**-6 averaging.
You can also create your own "distance.out" file in a format similar to
that below. Fields must be separated by one or more blanks or tabs.
Comment lines beginning with "#" and blank lines are ignored.
# ATOM I ATOM J MINIMUM MAXIMUM AVERAGE RMS
2 TYR HN 2 TYR HA 2.87 2.95 2.91 0.03
2 TYR HN 2 TYR HB2 2.25 2.76 2.44 0.14
2 TYR HN 2 TYR HB3 3.50 3.82 3.61 0.08
2 TYR HN 2 TYR HD 2.43 3.64 2.85 0.46
2 TYR HN 2 TYR HE 4.74 5.95 5.11 0.33
2 TYR HN 3 LYS HN 4.12 4.68 4.53 0.16
2 TYR HN 3 LYS HA 4.77 5.67 5.35 0.23
2 TYR HN 3 LYS HG2 4.49 8.26 6.40 1.24
2 TYR HN 3 LYS HG3 3.66 8.04 5.65 1.21
2 TYR HN 3 LYS HD2 4.58 10.09 7.18 1.39
2 TYR HN 4 CYS HN 4.87 7.36 6.83 0.73
2 TYR HN 10 SER HA 4.01 5.20 4.78 0.46
2 TYR HN 10 SER HB2 4.61 8.01 6.47 0.89
2 TYR HN 10 SER HB3 4.44 7.31 5.69 0.86
2 TYR HN 11 PHE HN 3.15 3.77 3.52 0.20
2 TYR HN 11 PHE HA 4.37 6.08 5.18 0.61
2 TYR HN 11 PHE HB2 3.07 4.78 3.73 0.53
2 TYR HN 11 PHE HB3 4.11 6.11 4.88 0.61
2 TYR HN 11 PHE HD 4.55 6.38 5.67 0.51
2 TYR HN 12 VAL HN 2.44 4.99 4.12 1.03
2 TYR HN 12 VAL HA 1.94 4.42 3.44 0.82
2 TYR HN 12 VAL HB 3.56 6.39 5.63 0.84
2 TYR HN 12 VAL MG1 3.94 6.93 6.06 0.89
# etc.
Read the crosspeak file (rxp) and invoke the assignment procedure (asg)
as described above in Case 0.
--------------------------------------------------------------------------
More about distance filters.
__________________________________________________________________________
A distance filter for GENXPK can be any program (shell, awk, C, etc.)
that reads N-column crosspeak format on standard input and produces
N-column crosspeak format on standard output. Comment lines (input and
output) begin with "#".
The distance filter can contain a Unix pipeline connecting several
programs, provided that the final input and output are Ncolumn
crosspeak format as described.
--------------------------------------------------------------------------
Additional topics (heteronuclear NOE, correlated peak assignments, etc.)
__________________________________________________________________________
Garry Paul Gippert
Physical Chemistry 2
Lund University
Chemical Center Box 124
S-22100 Lund, Sweden
(+46) 46 222 4470 phone
(+46) 46 222 4543 FAX
garry@bor.fkem2.lth.se
http://www.fkem2.lth.se/~garry