On 20/05/14 15:10, George Devaniranjan wrote:
Hi,
Hi George,
Sorry for the delay.
I am new to COOT
Hello. Welcome.
and electron density analysis and I have been reading the manual and
the tutorials so far. I have a question to which I have yet been
unable to find an answer.
It's hard to answer - there isn't a tool that gives you a number from
which you can make a yes/no decision.
We like to look at the model and the electron density to see if it
"looks good" :-)
I took a fragment (a part of the whole protein) modified it slightly
(made all the omega angles close to 180) and I want to see "how well
it still fits" the e-density. A quantitative measure of how well this
"ideal" protein fits into the e-denisty.
Can it still be identified ?
Have you tried density-score-by-residue?
$ density-score-by-residue
Usage:
/home/paule/Projects/build-enhanced-ligand/src/.libs/lt-density-score-by-residue-real
--pdbin pdb-in-filename --hklin mtz-filename --f f_col_label --phi
phi_col_label --weight weight_col_label --score-atoms --cut --cut-level
(absolute) level --pdbout pdb-out-filename
where pdbin is the protein (typically)
I want to "measure" how well my "new" fragment's C-ALPHA carbon's fit
the e-density (not really worried about the side chains at this point).
To use density-score-by-residue you'd need to create a pdb file without
side-chains.
Basically I want to know from more experienced crystallographers
whether if I gave this fragment you would accept it as a good
"starting" structure for the given e-density?
We would look at the correlation coefficients. Or I would at least.
You can get something like that - and more sophisticated scoring even -
using CCP4's EDSTATS. That may help.
Paul.
