[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ccp4bb]: Re:[ccp4bb] SUMMARY loop lmsd's, was tricking TOPP

***  For details on how to be removed from this list visit the  ***
***    CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html    ***

Thank you to everyone who replied.  Sorry to take so long to
send the summary...  

Summary of the question:

I want to calculate rmsd's of specific loops in the active site.

I'd tried:

 1) CNS rmsd.inp:  rmsd for each residue, but both structures have
 to be described by the same .mtf file, so I can't evaluate same
 protein from different species or mutants easily.

 2) DALI gives a single overal rmsd for the whole protein, but I
 am interested in certain loops. can't look at segments of protein

 3) TOPP breaks the rmsd's down into 2nd structure, calculating
 rmsd's for each alpha-helix and beta-strand, but not for loops.
    NOTE: You can trick TOPP by making fake 'HELIX' cards
           where there are really loops (not helices), but
           if the rmsd's of those fake helices <2 ... rejected!

Summary of the suggestions with my comments from trying them...

1) Edit .pdb's to only have the loops of interest, send to DALI:
    Argh!  I have 9 complexes of dimers with a TIM barrel
           and 8 loops of interest in each active site!
           That's 144 edited .pdb files pairwise to DALI.  ;-)
    Also: if I send edited loops to DALI, the overall superposition
           of the protein is not possible and rmsd's are isolated
           numbers with little meaning.

2) LSQMAN:  http://xray.bmc.uu.se/usf/welcome2usf.html:
    This is a nice program, but with 144 loops to compare, an 
    interactive program can't be left to run overnight.
3) CCP4's LSQKAB or COMPAR: These seem convenient, but I didn't 
    try them, because I found another one I liked before I got 
    to this suggestion.... sorry!

4) ALIGN (Cohen, NIH; J. Appl. Cryst, 1997, 30:1160-1161) or 
   HOMOLOGY (Rossmann's):
   I tried to find these to download from the web, but the names 
   are not unique enough ...too many hits... so I gave up after 
   half hour or so.

4) ProFit:  http://www.biochem.ucl.ac.uk/~martin/swreg.html
                  This one I liked the best!!!
   -It's very flexible!
   -It's script-run... so I wrote the script once and just
    changed the filenames to compare different protein pairs.  
   -You can specify exactly which residues to use for the
    LSq-superposition, and exactly which ones to calculate
    the rmsd, and they can be different.
   -You can specify 'C-alpha' or 'all atom' rmsd, so in the CA
    mode, it overlooks mutations (not like CNS).

    NOTE:  I can't find a reference for ProFit now that I used
    it and want to put the results in a paper...???

thanks for all the responses!
have fun


donate free food	www.thehungersite.com

diane h peapus, phd		dp72@cornell.edu
Chemistry & Chemical Biology	+1-607-255-2174 (office)
Cornell University		+1-607-255-2600 (graph room)
Ithaca, NY 14853  USA		+1-607-255-1227 (fax)