Elaine Meng wrote:
On Jul 27, 2008, at 10:38 AM, Francesco Pietra wrote:
Hi Elaine: After a brief playing with Multiscale Models, I was considering your last suggestion below. I modeled each chain separately, also because the X-ray diffr data show some conformational differences. Modeling gave very good DOPE score. Now the problem is appending Chain B to chain A, etc (each chain opens correctly in Chimera). I have a Python script (not tried yet) to insert the Chain identifier (which Modeller did not print), update the Atom serial number, and then append the files. May I ask, in view of Chimera, if the "Atom number + 1" printed in Modeller pdb files at column 80 can be left at it is, i.e. not updated on appending the files? I forgot about that and I have already problems enough with the script at my minimal level with Python. I assumed that Python identifies the columns one unity less than PDB numbering. Finally, do you know what means "1SG" at columns 74-76 of Modeller pdb files?
Modeller questions are best asked on the Modeller mailing list, of course. I will follow up there.
According to the description of ATOM lines: http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/tutorials/framepdbintro.html 67-76 is supposed to be blank, 77-78 element symbol, 79-80 charge. I believe Chimera tolerates lines without the last few columns
Modeller still writes old-style PDB files, in which these columns were used for the sequence number.
My suggestion of modeling the multimer from a template multimer would still incorporate the chain differences - you would still use A as the template for A, B for B, etc. However, I don't know how to make modeller do that.
I agree. It is very easy to build such models in Modeller.
Following the "appending" plan below, is it correct to append pdb files lacking all hydrogens (as provided by Modeller) and then rely on Amber's LEAP to complete the multimer?
Modeller can also produce output files containing hydrogens. We don't do it by default because that doubles the number of atoms (thus quadrupling the computational time) for no real benefit - most templates don't contain hydrogens anyway. Ben -- ben@salilab.org http://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle