[BiO BB] Generate Quaternary structure of protein from monomer

Boris Steipe boris.steipe at utoronto.ca
Fri Mar 4 00:12:08 EST 2005

Depending on the crystal symmetry and the orientation of the monomer in  
the asymmetric unit, this all can be nontrivial.

Consider a dimer "()" in a crystal lattice  ...()()()()()()...
The asymmetric unit contains the monomer "(" or its rotated copy ")"  
but, depending on where the monomer is located in the unit cell,  
applying the symmetry matrix once may give you "()" or it may give you  
")(". So in order to cover all cases, you may have to apply the matrix  
to two copies and _translate_ one by one unit cell to give ")()", then  
decide which of the pairs is the one you want.

Pragmatically, and depending how familiar you are with "O" or other  
crystallography tools, it  may be faster downloading the biological  
unit from PDB or EBI-PQS and then using a superposition tool or Web  
service to superimpose your model on each of the monomers. It really  
depends on the spacegroup.

Also, depending on the graphics program you use, you may either have to  
give each of the monomers a separate chain identifier "A", "B", "C" ...  
(check the biological unit file to see how this is done) or put a  
"MODEL  1", "MODEL  2" etc. line before each copy and  an "ENDMDL" line  
after it (check any NMR structure ensemble to see how that is used).

Hope this helps,


On Thursday, Mar 3, 2005, at 23:16 Canada/Eastern, James Stroud wrote:

> Hi Subhash,
> You will need to get the transformation matrices from the template.  
> This can
> be done in the program O < http://www.bioxray.dk/~mok/o-files.html >.
> The commands you will need are:
> pdb_read        (to read your pdb-file)
> lsq_explicit    (to determine the transformation matrix and store it)
> The o manual is at < http://www.bioxray.dk/~mok/o/o_man/manual.html >
> Once you have the transformation stored in the o database  
> (".lsq_rt_?????",
> where "?????" is the name you provide to the lsq-expl command), you  
> can load
> your protomer with
> pdb_read
> Then you can read in the second copy, again with "pdb_read".
> Apply the transformation matrix from the template to the second copy  
> with
> lsq_mol
> You can read in further monomers and apply the transformation matrix N  
> times
> until you fill your polymer.
> You should then write out every monomer with
> pdb-write
> and join the files with a text editor, taking care to remove headers  
> and "END"
> statements between files.
> Things to think about:
> (1) Molecule names in O are the first 5 characters only (this can cause
> serious confusion to first-timers.)
> (2) If you have a cyclic N-mer where (N > 2) and (N mod 2 = 0) [i.e.  
> even and
> greater than 2], then you should take care to choose the right two  
> monomers
> from the template polymer to generate the transformation matrix or you  
> will
> skip monomers in your model (think about starting at one vertex of a  
> hexagon
> and transforming 120 degrees versus 60 degrees, etc.--both relate  
> equivalent
> vertices but the former skips three under all transformations).
> James
> On Thursday 03 March 2005 07:25 pm, Subhash Agarwal wrote:
>> Hi all
>> Reply to Dan Bosler answer
>> I knew of these sites and visited them again after u
>> referred. These sites are useful if u have a pdb code
>> avaliable. In my case i have built a homology model
>> for which i need to generate quaternary structure.
>> Reply to James Strond answer
>> Yes, for template structure has a quaternary structure
>> and therefore are known.
>> My Query:
>> How can i use these transformations for generating
>> Quaternary structure? Moreover from the pdb of the
>> template i also need to understand that the
>> transformation are in regard to which monomer. How can
>> i know that.
>> ______________________________________________________________________ 
>> __
>> Yahoo! India Matrimony: Find your life partner online
>> Go to: http://yahoo.shaadi.com/india-matrimony
>> _______________________________________________
>> Bioinformatics.Org general forum  -   
>> BiO_Bulletin_Board at bioinformatics.org
>> https://bioinformatics.org/mailman/listinfo/bio_bulletin_board
> -- 
> James Stroud, Ph.D.
> UCLA-DOE Institute for Genomics and Proteomics
> Box 951570
> Los Angeles, CA 90095
> _______________________________________________
> Bioinformatics.Org general forum  -   
> BiO_Bulletin_Board at bioinformatics.org
> https://bioinformatics.org/mailman/listinfo/bio_bulletin_board

More information about the BBB mailing list