hi -
my two cents:
1. I am puzzled to see only positive and no negative density in the
picture
2. I would contour difference maps at 3.5 in general for significant
features
3. why is the 2fofc 'heavily biased' ? Its a 2mFo-DFc map presumably,
and lots have been said about these maps not being so biased these days
of maximum likelihood. With an Rfree of 27% and 80% similarity (!not
homology, please!)
model to start with, even at 3.0 A I cant agree I expect maps to be
'heavily biased'.
Sorry James, but I find these maps essential, and not essentially
useless.
4. if you want to really look for wrong direction, registry trouble,
pep flips,
you are much better of looking at validation reports for that regions
and less at the density.
If these residues look nice in e.g. Ramachandran plots, I would guess
the green density
might at the end be a 'technicality'.
What I would do is apply first some small random shift (0.5 A) to all
atoms,
go back to Refmac, start with a new TLD tensors from 0 and reset B
factors,
refine, make sure my RMSZ scores for distances and angles is 0.5-1.0,
look at maps again and make sure I see both red and green peaks at 3.5,
run a validation test with molprobity or whatcheck, and then look for
and correct mistakes.
best,
Tassos
PS Since there were a few posting like that lately, with pictures
attached, could we please
follow the suggested ccp4bb "etiquette" and not post attachments of
pictures,
but make them available online for whoever is really interested only?
On Aug 21, 2009, at 7:28, James Stroud wrote:
I agree with Artem. I also don't think you have a register problem.
All of the residues show typical 2fofc density for their types,
especially considering the 3A data. The disappearing act of the asp
carboxyl is typical of solvent exposed aspartates. Assuming the
register is correct, you have a carbonyl flip somewhere throwing off
the whole segment. The phe is telltale. The 2 dimensionality and
limited region of the picture makes a full diagnosis difficult.
Also, the 2fofc is heavily biased. Don't trust it when it tells you
where things are. I find 2fofcs essentially useless. Your milage may
vary.
James
On Aug 20, 2009, at 5:24 PM, Mike England wrote:
Hi all,
I will appreciate the comments on the 2Fo-Fc (1.5 sigma) and Fo-Fc
(3.0 sigma) (at coot) maps as shown in attached picture .
I am working at 3.0 A resolution (MR phasing with more than 80%
homology) and current Rfree is 0.27 and FOM of 0.8.
How to interpret the positive density (green) in Fo-Fc map which
overlaps with the C-alpha tube? The side-chains of the polypeptide
around this regions seems to be properly fitting and registry of the
sequence seems to be OK.
Is this due to some kind of model bias?
Thanks in advance.
Mike
<Picture 1.png>
P please don't print this e-mail unless you really need to
Anastassis (Tassos) Perrakis, Principal Investigator / Staff Member
Department of Biochemistry (B8)
Netherlands Cancer Institute,
Dept. B8, 1066 CX Amsterdam, The Netherlands
Tel: +31 20 512 1951 Fax: +31 20 512 1954 Mobile / SMS: +31 6 28 597791
