[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[ccp4bb]: [summary]secondary structure assignments
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Dear All ;
Here are suggestions from several people.
Thank you all.
The original question was
>I am using PyMOL and it requires secondary-structure-assigned pdb file for
>the correctly assigned secondary structure.
>I have successfully ran DSSP for the assignment of the secondary structure
>but don't know how to add this information to pdb file.
>Does anyone know how to do?
>Is there any program to do this?
#####################################################################
Dr. Hyun Kyu Song provided the information about Dr. Kristian Rother's
Python script.
OK, heres what we've got:
The DSSP program is available for Linux and Windows from
http://www.cmbi.kun.nl/swift/dssp/ .
I have written a Python script that parses the output of DSSP and alters
the annotation in PyMOL. You just start the script, then You have a
dssp <object>
command available.
The script can be found on my homepage,
http://www.rubor.de/bioinf/pymol_extensions.html
along with some useful stuff
http://www.rubor.de/bioinf/index.html
However, the script works for sure only on Unix. For Windows, i still
need to figure out how to
- redirect the output of Dssp on Windows
- assign platform-independent Paths correctly.
If someone could test the script, that would be great.
Kristian
###############################################################
Dr. Michel Fodje suggeted the following web site.
If you want to accurately identify Pi-helices also, use this site:
http://www.mbfys.lu.se/Services/SecStr/
Both DSSP has a few problems with Pi-helices which as we and others have
shown could be important for the function of your protein. STRIDE is
worse than DSSP in this area.
Weaver, T.M (2000) Protein Sci, 9 201-206
Fodje, M.N & Al-Karadaghi, S (2002) Protein Eng, 15:5 353-358
The program is basically an improved implementation of DSSP to correctly
identify Pi-helices. The output can be pasted at the top of your PDB
file to be recognized by PyMOL, rasmol, SPDBV etc. It is not 100% PDB
format but all the essentials are there for graphical programs plus
extra information such as helical geometric parameters.
########################################################
Dr. Laurence Pearl suggested KSDSSP.
We use a program called KSDSSP from Conrad Huang at UCSF - this generates
PDB secondary structure records directly from the coordinates, using the
Kabsch and Sander dssp algorithm. You just have to insert the output into
the pdb above the first atom record, and off you go ....
###################################################################
Dr. James Stroud suggested the program STRIDE and Stride2pdb.
I do not use DSSP, but I use the very similar (and some say better) stride.
Stride, I believe, is mostly an implementation of DSSP with some
"improvements". It mostly gives the same assignments as DSSP with rare
exception. Secondary structure assignment specialists may correct me on this
one. The stride homepage is/was
http://www-db.embl-heidelberg.de/jss/servlet/de.embl.bk.wwwTools.GroupLeftEM
BL/argos/stride/
(the link was split by my email program, so you will have to concatenate the
two preceding lines of text)
I have compiled an OSX version of stride (not m$ windoze--HA!) at the link
below. I believe that I am making this software available in accordance with
the license in the documentation.
http://keres.colorado.edu/stride2pdb/stride-osx.tar
Since the stride output is not readable by PDB parsers in general I made a
utility called stride2pdb that solves this problem. It is available at
http://keres.colorado.edu/stride2pdb/
stride2pdb is a perl script so you should have perl running. Perl is
standard on unix boxes (Linux, OSX, IRIX, etc-x, but not m$ windoze--again,
HA!).
Please read the README file for usage.
#############################################################
Dr. Kim Henrick suggested the follwing method.
for dssp/promotif to pdb helix_strand
you can try www.ebi.ac.uk/~henrick/doss.tar.gz
this requires a f90 compiler
and is run as
doss 1abc
expecting a file pdb1abc.ent
to change the file opening to
doss my.brk
i.e. use arg1 as the complete file name
shouldnt be a problem
The main o/p is mmcif or a direct oracle read/write
there is a dummy set of oracle calls as a .f file
to write complex sheets into the limited pdb format
is difficult as a strand can only have 2 neighbours
in the pdb while in some cases one strand can have
more than 2 partners by linking 2 sheets - therefore
the 'helix_sheet' o/p is not entirely 'correct'
as long strands that link 2 sheets have h-bonds broken
deliberately to make a pdb format (and in about 1 in
a 1000 pdb entries it may not give the correct
SHEET format)
The code is dssp+promotif written for the mmcif o/p
(or oracle) to give always unique id's and for
very large structures not to either core dump
or reuse strand/sheet id's as in promotif and dssp
respectively
nass.tar.gz does sort of similar things for nucleic acids
but does not yet do the complexities of rna secondary
structure
Many thanks to all.
Cheom-Gil Cheong
_________________________________________________________________
Add photos to your e-mail with MSN 8. Get 2 months FREE*.
http://join.msn.com/?page=features/featuredemail