Dear Anagha,
just a guess...
I'm not sure I understood you properly, but you are only trying to
make a model of a solved protein which will lack 3-4 amino acids in a
loop... therefore, why not using a modeling soft, such as Coot,
remove the amino acids, and do one run of "Regularize Zone"?? You
might need to renumber your chain once you have removed the amino
acids, but it might work...
just a guess...
Best regards.
Leo
On Mar 1, 2007, at 4:11 AM, anagha gupta wrote:
Hi CCP4 community!
I have constructed a homology model of a deletion variant of a
protein whose structure has already been solved. These deletions
are 3-4 amino acid in length and are in a loop that connects two
helices. The model structures look good with respect to bond
lengths in the aforementioned loop region but the bond angles
(especially Phi and Psi) are very distorted. Ramachandran plot
suggests the same and some of the amino acids flanking the loop are
now in the disallowed region.
I thought one way to avoid this is to delete one amino acid at a
time and construct the model and use the previous model as the
template to construct the next model. This is not working very
well in fixing the wrong angles.
I was wondering if
1) Anybody out there knows good homology modeling software. I used
SWISS model and CPH model to create my models. I have heard about
prime but right now am waiting to get the software .
2) Is there a way I can fix the angles if I perform an energy
minimization of the model.
Suggestions are appreciated.
Thanks,
Anagha.
==========================
Chavas Leonard M.G., Ph.D.
Structural Research Center
KEK,PF. 1-1 Oho
Tuskuba, Ibaraki - Japan
---------------------------------------------
PHS: +81(0)29-864-5200 (ext: 2682)
e-mail: [EMAIL PROTECTED]
URL: http://pfweis.kek.jp/~leo
==========================
