[ccp4bb] Correction: Refmac and MSE?
In the previous message I forgot to mention: When MET were modelled, there were no difference peaks anywhere. When I changed them all to MSE, the large difference density peaks showed up. Large *negative* difference density.
[ccp4bb] electrostatics of a small molecule & calculation of desolvation energies
Hi CCP4ers, I am looking for some advice on electrostatic maps of small molecules. What is the best program for this? Is there anyway I can calculate the desolvation energy at the same time? I have tried to create a PQR file for APBS (through PDB2PQR) with the pdb of the small molecule (and the mol2 file) - but it seems to be hanging after 4 days. Any help would be greatly appreciated Rose Harrison
[ccp4bb] Refmac and MSE?
Hello, I am at a loss on what's going on: I am refining SeMET containing structure and using REFMAC 5.2.0005 on Linux and, the same thing happening, using REFMAC 5.5.0070 on Windows. When MET were modelled, there were no difference peaks anywhere. When I changed them all to MSE, the large difference density peaks showed up. So either the protein does not contain SeMet or Refmac somehow uses sulfur scattering factors during refinement. I have hard time believing the former because 1) the protein was checked by mass spec to be correct size for SeMet derivative, 2) the structure was solved by SAD with all five sites correctly found (196 residues total), 3) first 50 aa of the protein are identical to a known structure. This is 2.4A resolution and at this point R/Rfree = 25/29. Refmac log shows correct scattering factors read out: SE17.0006 2.4098 5.8196 0.2726 3.9731 15.2372 4.3543 43.8163 2.8409 The PDB has this for MSE: ATOM 1140 N MSE A 143 35.708 161.163 13.715 1.00 78.82 N ATOM 1141 CA MSE A 143 35.995 162.467 13.106 1.00 77.86 C ATOM 1142 CB MSE A 143 36.307 162.307 11.617 1.00 78.79 C ATOM 1143 CG MSE A 143 37.755 162.127 11.119 1.00 80.89 C ATOM 1144 SE MSE A 143 37.503 161.104 9.363 1.00 91.22 SE ATOM 1145 CE MSE A 143 39.169 160.021 9.696 1.00 86.47 C ATOM 1146 C MSE A 143 34.709 163.226 13.030 1.00 77.70 C ATOM 1147 O MSE A 143 34.682 164.436 12.737 1.00 77.33 O Any clues greatly appreciated! Dima
Re: [ccp4bb] ccp4-6.1.0: which ccp4.setup in Ubuntu Intrepid?
When I "upgraded" to ibex it hosed my display, so this might not be of
much help, but in general,
1. You don't HAVE to use bash. You can use tcsh or zsh or any other
shell that suits your personal tastes. It just comes with a minimal
install to fit on a CD. Just issue
sudo apt-get install tcsh zsh
and you will have three shells to choose from.
2. To use tcsh, you can issue a command such as (after installing the
above),
chsh -s /bin/tcsh
(or even better, chsh -s /bin/zsh ).
3. You are right: setenv is indicates tcsh or csh
4. You can add directories to your path in any relevant startup
script or on the command line for that matter.
For all but tcsh and csh, use this syntax:
PATH=${PATH}:/your/new/directory ; export PATH
For tcsh and csh, use
setenv PATH ${PATH}:/your/new/directory
On Dec 19, 2008, at 6:01 PM, Victor Alves wrote:
*
*
*
Hi
Doing my first installation of CCP4 and also a Linux newbie, so try
to ?bear with me? ;-)
The ?install.sh? CCP4 script created a ?setup-scripts? directory,
with 2 shell subdirectories; ?csh? and ?sh?
In ?/home/victor/.bashrc? I found 2 lines:
source /usr/local/xtal/ccp4/setup-scripts/sh/ccp4.setup
source /usr/local/xtal/ccp4/setup-scripts/sh/ccp4-others.setup
So, I assume it is using the ccp4.setup-sh file.
But Ubuntu Intrepid 8.10 uses ?bash?. If I type ?echo $SHELL?, it
gives me /bin/bash, and Bash doesn?t use ?setenv? but ?export? (so I
read).
1-Academic question: If I were to do a manual installation and
chose, as a template for ?ccp4.setup?, the file?ccp4.setup-bash?
(which uses ?setenv?) would I get into trouble?
2-If Ubuntu uses a ?bash? shell, why it is not using ?ccp4.setup-
bash?
3-Which ?ccp4.setup? file is CCP4 using?:
a) ?usr/local/xtal/ccp4/ccp4-6.1.0/include? or
b) ?/usr/local/xtal/ccp4/setup-scripts/sh/ccp4.setup?
4-I manually installed CCP4MG and the instructions state to ?Add
the CCP4mg "bin" directory to your PATH in ccp4.setup?, but the only
place I found to do this was in ?/usr/local/xtal/ccp4/setup-scripts/
sh/CCP4-OTHERS.SETUP? not ?ccp4.setup?.
So I did:
sudo gedit /usr/local/xtal/ccp4/setup-scripts/sh/ccp4-others.setup
Added the following line, (right after a similar one for COOT):
/usr/local/xtal/ccp4/ccp4mg-1.1.1/bin:
After Login, I now have 2 lines for CCP4MG-1.1.1/BIN. What did I do
wrong?
echo $PATH (I reformatted one statement per line, for easier reading)
/usr/local/xtal/ccp4/graphviz-2.20.2/bin:
/usr/local/xtal/ccp4/fasta-35.4.1/bin:
/usr/local/xtal/ccp4/clustalw-2.0.9/bin:
/usr/local/xtal/ccp4/tcltk++/bin:
/usr/local/xtal/ccp4/Python-2.4.5/bin:
/usr/local/xtal/ccp4/ccp4-6.1.0//bin:
/usr/local/xtal/ccp4/ccp4-6.1.0/bin:
/usr/local/xtal/ccp4/Coot-0.5/bin:
/USR/LOCAL/XTAL/CCP4/CCP4MG-1.1.1/BIN:
/USR/LOCAL/XTAL/CCP4/CCP4MG-1.1.1/BIN:
/usr/local/xtal/ccp4/ccp4-6.1.0/share/dbccp4i/bin:
/usr/local/xtal/ccp4/ccp4-6.1.0/ccp4i/bin:
/usr/local/xtal/ccp4/ccp4-6.1.0/bin:
/usr/local/xtal/ccp4/ccp4-6.1.0/etc:
/usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2//binaries/
linux_386:
/usr/local/sbin:
/usr/local/bin:
/usr/sbin:
/usr/bin:
/sbin:
/bin:
/usr/games:
/usr/local/xtal/chimera/bin:
/usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2core//Test:
/usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2core//Test/
Compiled/linux_386:
/usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2//Applications
Thank you for your patience
Victor Alves
__
VICTOR DIOGO ALVES, DVM PhD
Professor Auxiliar
Faculdade de Medicina Veterinaria
UTL - Lisboa
__
This message was sent using IMP, the Internet Messaging Program.
Re: [ccp4bb] ccp4-6.1.0: which ccp4.setup in Ubuntu Intrepid?
Sorry for typos. The " were replaced with ? Victor > Hi Doing my first installation of CCP4 and also a Linux newbie, so try to ?bear with me? ;-) This message was sent using IMP, the Internet Messaging Program.
[ccp4bb] ccp4-6.1.0: which ccp4.setup in Ubuntu Intrepid?
* * * Hi Doing my first installation of CCP4 and also a Linux newbie, so try to ?bear with me? ;-) The ?install.sh? CCP4 script created a ?setup-scripts? directory, with 2 shell subdirectories; ?csh? and ?sh? In ?/home/victor/.bashrc? I found 2 lines: source /usr/local/xtal/ccp4/setup-scripts/sh/ccp4.setup source /usr/local/xtal/ccp4/setup-scripts/sh/ccp4-others.setup So, I assume it is using the ccp4.setup-sh file. But Ubuntu Intrepid 8.10 uses ?bash?. If I type ?echo $SHELL?, it gives me /bin/bash, and Bash doesn?t use ?setenv? but ?export? (so I read). 1-Academic question: If I were to do a manual installation and chose, as a template for ?ccp4.setup?, the file?ccp4.setup-bash? (which uses ?setenv?) would I get into trouble? 2-If Ubuntu uses a ?bash? shell, why it is not using ?ccp4.setup-bash? 3-Which ?ccp4.setup? file is CCP4 using?: a) ?usr/local/xtal/ccp4/ccp4-6.1.0/include? or b) ?/usr/local/xtal/ccp4/setup-scripts/sh/ccp4.setup? 4-I manually installed CCP4MG and the instructions state to ?Add the CCP4mg "bin" directory to your PATH in ccp4.setup?, but the only place I found to do this was in ?/usr/local/xtal/ccp4/setup-scripts/sh/CCP4-OTHERS.SETUP? not ?ccp4.setup?. So I did: sudo gedit /usr/local/xtal/ccp4/setup-scripts/sh/ccp4-others.setup Added the following line, (right after a similar one for COOT): /usr/local/xtal/ccp4/ccp4mg-1.1.1/bin: After Login, I now have 2 lines for CCP4MG-1.1.1/BIN. What did I do wrong? echo $PATH (I reformatted one statement per line, for easier reading) /usr/local/xtal/ccp4/graphviz-2.20.2/bin: /usr/local/xtal/ccp4/fasta-35.4.1/bin: /usr/local/xtal/ccp4/clustalw-2.0.9/bin: /usr/local/xtal/ccp4/tcltk++/bin: /usr/local/xtal/ccp4/Python-2.4.5/bin: /usr/local/xtal/ccp4/ccp4-6.1.0//bin: /usr/local/xtal/ccp4/ccp4-6.1.0/bin: /usr/local/xtal/ccp4/Coot-0.5/bin: /USR/LOCAL/XTAL/CCP4/CCP4MG-1.1.1/BIN: /USR/LOCAL/XTAL/CCP4/CCP4MG-1.1.1/BIN: /usr/local/xtal/ccp4/ccp4-6.1.0/share/dbccp4i/bin: /usr/local/xtal/ccp4/ccp4-6.1.0/ccp4i/bin: /usr/local/xtal/ccp4/ccp4-6.1.0/bin: /usr/local/xtal/ccp4/ccp4-6.1.0/etc: /usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2//binaries/linux_386: /usr/local/sbin: /usr/local/bin: /usr/sbin: /usr/bin: /sbin: /bin: /usr/games: /usr/local/xtal/chimera/bin: /usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2core//Test: /usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2core//Test/Compiled/linux_386: /usr/local/xtal/ccp4/ccp4-6.1.0/share/XIAROOT/xia2//Applications Thank you for your patience Victor Alves __ VICTOR DIOGO ALVES, DVM PhD Professor Auxiliar Faculdade de Medicina Veterinaria UTL - Lisboa __ This message was sent using IMP, the Internet Messaging Program.
Re: [ccp4bb] refmac 5.5.063 vs. 5.5.066
The latest version with all the bugfixes will of course appear in CCP4-6.1.1. Until then, the latest binaries and source codes are available on the mentioned webpage. Pavol
Re: [ccp4bb] refmac 5.5.063 vs. 5.5.066
Since version 5.5.0066 is the version in ccp4-6.1.0, is there going to be a patch? On Dec 19, 2008, at 12:32 PM, Jan Abendroth wrote: Thanks a lot, Pavol, that indeed did the trick. refmacRwork Rfree RMSDb RMSDan 5.5.0063 0.1682 0.2118 0.0204 1.717 5.5.0066 0.1781 0.2247 0.0205 1.764 5.5.0070 0.1663 0.2102 0.0204 1.719 Cheers Jan On Dec 19, 2008, at 2:21 AM, Pavol Skubak wrote: Hi Jan, the difference between 5.5.0063 and 5.5.0066 is just a few bugfixes. However, unfortunately one of them introduced a new bug to 5.5.0066 which has been fixed in the later versions. Could you please upgrade to the latest version (currently 5.5.0070) from http://www.ysbl.york.ac.uk/~garib/refmac/data/refmac_stable/ That should fix the problem (please let us know if it does not) Pavol. -- Jan Abendroth deCODE biostructures Seattle / Bainbridge Island WA, USA work: JAbendroth_at_decode.is home: Jan.Abendroth_at_gmail.com
Re: [ccp4bb] refmac 5.5.063 vs. 5.5.066
Thanks a lot, Pavol, that indeed did the trick. refmacRwork Rfree RMSDb RMSDan 5.5.0063 0.1682 0.2118 0.0204 1.717 5.5.0066 0.1781 0.2247 0.0205 1.764 5.5.0070 0.1663 0.2102 0.0204 1.719 Cheers Jan On Dec 19, 2008, at 2:21 AM, Pavol Skubak wrote: Hi Jan, the difference between 5.5.0063 and 5.5.0066 is just a few bugfixes. However, unfortunately one of them introduced a new bug to 5.5.0066 which has been fixed in the later versions. Could you please upgrade to the latest version (currently 5.5.0070) from http://www.ysbl.york.ac.uk/~garib/refmac/data/refmac_stable/ That should fix the problem (please let us know if it does not) Pavol. -- Jan Abendroth deCODE biostructures Seattle / Bainbridge Island WA, USA work: JAbendroth_at_decode.is home: Jan.Abendroth_at_gmail.com
[ccp4bb] Need help for solving a tough problem of phasing
Dear all, I have determined a structure with the resolution 2.0A and the spacegroup P3121 from crystals grown at 4 degree. This apo-form structure is composed of an N-terminal (a/b)8 TIM barrel (~300 AA) and two C-terminal domains (~100 AA) which are named as Sub1 and Sub2. Meanwhile, I have also got crystals at 20 degree. Although the diffraction of these crystals is not good, I could confirm that the spacegroup has changed to C2. Recently, I have collected several datasets from crystals with different substrates at 20 degree. All of them have relative low resolution from 3.5A to 2.8A and completeness from 80% to 90% and have the same spacegroup and almost the same cell parameters. Considering that the biological unit is dimer and the C-terminal domains have high flexibility, I use both monomer and dimer forms of the TIM barrel as the search models when I do MR in the best dataset (2.8A resolution and 85% completeness). I use three programs CNS, Molrep, and Phaser and they all could find the same optimal solution which contains a dimer and about 68% solvent in one asymmetric unit. After several rounds of refinement, the R and freeR have dropped to 35.6 and 41.6, respectively and the density fit well at the TIM barrel region. Between the symmetric molecules exists a lot of discontinuous positive peaks which have no obvious secondary structure patterns. All the information above show that something unknown is not determined in this structure. I superpose the full-length dimer to the determined TIM barrel dimer but find that the C-terminal domains have severe conflict with the symmetric molecules at this packing form. I continue the MR using the TIM barrel, Sub1, Sub2, or Sub1+Sub2 as search models with fixed TIM barrel dimer, however no right solutions are obtained. I also do the MR with the full-length monomer and dimer forms but there are still no right solutions. Then I have tried the other datasets, however I get the same results. Now I am trying to build the model manually. Because of the discontinuous densities, it is hard to finish the building work at a relatively short time and I think it is not a right way to cut the knot. I don't know how and what to do next. Could anyone give me a reasonable method to solve this tough problem? Any suggestion is appreciated. Jian Wu -- Jian Wu Ph.D. Student Institute of Biochemistry and Cell Biology Shanghai Institutes for Biological Sciences Chinese Academy of Sciences (CAS) Tel: 0086-21-54921117 Email: prote...@gmail.com
Re: [ccp4bb] install ccp4 in OS X
On Dec 19, 2008, at 7:43 AM, Eric Liu wrote:
Hi All
Hi Eric:
I am a new user for OS X system and this question may sound naive. I
followed the instruction on website Crystallography on OS X,
installed X11,
Xcode and fink on my OS X 10.5.5, but still could not install ccp4
with
fink. Here is the err message" no package found for specification
"ccp4""
after I typed the command " fink -b install ccp4"
As Graeme mentioned, you can just install ccp4's own pre-compiled
binaries.
However, if you want to do it as above, you need to activate the
"unstable" branch of
fink, then issue the command "fink selfupdate-rsync" (or selfupdate-
cvs) and then you
will see it. If you want to take advantage of the fink -b option,
there are a couple more
things you need to do, also described on that website (cf http://tinyurl.com/fink-precompiled
).
Finally, to get all the imosflm components to work, you can install a
placeholder package
called ("imosflm") that in reality installs the various itck,itk add-
ons. Coot and cctbx are also
separate packages, and I am working on a few others.
Also, Is there any place offering service for installing the
crystallographic softwares in OS X?
The hope of above procedure is to enable you (and others) to do this
without having to rely on the
services of anyone else, and the corresponding wiki is there to help
make this a community
effort. It grew out of my own struggles as a naive user and
struggling assistant professor
at a "do it yourself university" where support for infrastructure is
quite minimal. I've kept it
going, even though there is less of a need for it now than in the
early days of OS X, because
many people have relied on it for speed and efficiency. Charging for
it would seem somehow
morally wrong, and probably would be a license violation as well.
Any piece of help is appreciated.Thanks!
Email me if I can be of any help.
Best wishes.
Bill
Re: [ccp4bb] install ccp4 in OS X
Dear Eric, A native installer for CCP4 6.1.0 for this platform is available from the download web page: http://www.ccp4.ac.uk/download/downloadman.php this is a disk image and installs in the usual manner for macintosh applications. This includes all of the components needed. Best wishes, Graeme From: CCP4 bulletin board [mailto:ccp...@jiscmail.ac.uk] On Behalf Of Eric Liu Sent: 19 December 2008 15:44 To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] install ccp4 in OS X Hi All I am a new user for OS X system and this question may sound naive. I followed the instruction on website Crystallography on OS X, installed X11, Xcode and fink on my OS X 10.5.5, but still could not install ccp4 with fink. Here is the err message" no package found for specification "ccp4"" after I typed the command " fink -b install ccp4" Also, Is there any place offering service for installing the crystallographic softwares in OS X? Any piece of help is appreciated.Thanks!
Re: [ccp4bb] Ordering a cDNA library
Rajan, have a look at http://www.imagenes-bio.de. I guess there are more suppliers. Please send a summary to the board. Good luck, Guenter Rajan Pillai wrote: Hi All, Apologies for a non-CCP4 question. I want to clone a couple of proteins of human and mouse origin. Can anyone tell me where from I can order Human and mouse (brain/liver) cDNA library? Any suggestions are welcome. Thanks, Rajan -- *** Priv.Doz.Dr. Guenter Fritz Fachbereich Biologie Sektion Naturwissenschaften Universitaet Konstanz http://www.biologie.uni-konstanz.de/fritz Universitaetsstrasse 10 Postfach M665 D-78457 Konstanz e-mail: guenter.fr...@uni-konstanz.de Tel. Office: +49-(0)7531 88 3205 Tel. Lab : +49-(0)7531 88 3687 Fax: +49-(0)7531 88 2966
[ccp4bb] install ccp4 in OS X
Hi All I am a new user for OS X system and this question may sound naive. I followed the instruction on website Crystallography on OS X, installed X11, Xcode and fink on my OS X 10.5.5, but still could not install ccp4 with fink. Here is the err message" no package found for specification "ccp4"" after I typed the command " fink -b install ccp4" Also, Is there any place offering service for installing the crystallographic softwares in OS X? Any piece of help is appreciated.Thanks!
[ccp4bb] sftools interface problem
Yes, this is a bug. This is fixed now - you need 2 corrected ccp4i files. See http://www.ccp4.ac.uk/problems.php#6.1.0-sftools-interface which has links to the fixed files. Cheers Martyn On Thu, 2008-12-18 at 08:32 +, Winter, G (Graeme) wrote: > Dear Engin, > > Yep, I can reproduce this using a binary installation myself - that > would be a bug. We (CCP4) will look into it and get a fix to you as soon > as possible. > > Thanks for the report, > > Graeme > > CCP4 Support Team > > -Original Message- > From: CCP4 bulletin board [mailto:ccp...@jiscmail.ac.uk] On Behalf Of > Engin Ozkan > Sent: 18 December 2008 06:10 > To: CCP4BB@JISCMAIL.AC.UK > Subject: Re: [ccp4bb] iMosflm in 6.1 under OS X > > Now that everything with coot and imosflm has been settled, how about > sftools? I am getting errors when I try to open an sftools window with > ccp4i, with 6.1.0. I get the same error with a fink-installed mac > version, and a linux installation. Here is the error: > > bad window path name ".main.canvas.contents.param_1.body" > bad window path name ".main.canvas.contents.param_1.body" > while executing > "pack $wframe.body -side top -fill x" > (procedure "update_folder_display" line 23) > invoked from within > "update_folder_display 1 sftools_PROJECT view" > invoked from within > ".w_sftools_PROJECT.main.canvas.contents.param_1.top.but invoke" > (command bound to event) > > Any ideas? > > Engin -- *** * * * Dr. Martyn Winn * * * * STFC Daresbury Laboratory, Daresbury, Warrington, WA4 4AD, U.K. * * Tel: +44 1925 603455E-mail: martyn.w...@stfc.ac.uk* * Fax: +44 1925 603825Skype name: martyn.winn * * URL: http://www.ccp4.ac.uk/martyn/ * ***
Re: [ccp4bb] Transferring a Free R set.
Alun, There's only any point doing it if they're truly isomorphous (e.g. by the Crick-Magdoff test), otherwise pairs of reflections with equal indices won't sample the same points in the transforms, so you could still end up with contamination. Slow cooling or pseudo-MD such as randomly shifting co-ordinates actually moves you away from convergence, and as I said convergence is where you need to be for Rfree to be meaningful and to completely erase 'memories' in the structure. Also it's a progressive thing, the closer you get to convergence, the more meaningful Rfree becomes and the more the memories are erased. Of course convergence is very elusive, just because the parameter shifts appear to be getting smaller doesn't guarantee convergence (e.g. the series 1+1/2+1/3+1/4+1/5+... diverges to infinity!), so by all means try slow-cooling, just be prepared to carefully evaluate the results. Cheers -- Ian > -Original Message- > From: owner-ccp...@jiscmail.ac.uk > [mailto:owner-ccp...@jiscmail.ac.uk] On Behalf Of Alun R. Coker > Sent: 19 December 2008 10:37 > To: owner-ccp...@jiscmail.ac.uk > Cc: CCP4 bulletin board > Subject: Re: [ccp4bb] Transferring a Free R set. > > Hi Ian, Gerard, > > Thanks for your replies. > > Perhaps my hypothetical example wasn't a good one and may > have clouded > the issue. Could I just confirm whether or not you think its > necessary > to transfer initial free R assignment to any new data sets or to > isomorphous data sets such as substrate complexes. > > Thanks, > > Alun. > > Ian Tickle wrote: > > Hi Alun > > > > I've never seen a soaked ligand-bound structure that's > truly isomorphous > > with the native - simply soaking & freezing the crystal is virtually > > guaranteed to introduce very serious non-isomorphism, e.g. up to 5% > > change in one or more cell parameters. At least, we always > have a MR > > step to solve the structure, as rigid-body refinement alone > often fails > > to pull it in, so the R-factors always start high (> 0.4) > anyway. This > > renders the concept of a common test set totally > meaningless, since the > > native & ligand-bound reflections with the same indices > (except maybe > > the low res ones) will not be in equivalent positions in reciprocal > > space, and in any case once it has been refined down & been > through a > > few rounds of rebuilding any bias will very quickly disappear. > > > > Note that Rfree is a meaningful statistic only at convergence at the > > likelihood maximum, and at convergence all previous > 'memories' of the > > structure are erased: bias due to 'contamination' of the > test set only > > occurs as a result of failure to ensure convergence. Also I don't > > believe that any of the cases of deposited ligand structures where > > either there is clearly no density for the ligand, or the > ligand is in > > the wrong place, or even the wrong ligand has been built, > can be laid at > > the door of the test set that was used. Plain incompetence and/or > > wishful thinking is a much more likely explanation. So personally I > > think you're worrying over nothing. > > > > Just my 2p's worth. > > > > Cheers & Seasons greetings to Steve & Jon! > > > > -- Ian > > > > > >> -Original Message- > >> From: owner-ccp...@jiscmail.ac.uk > >> [mailto:owner-ccp...@jiscmail.ac.uk] On Behalf Of Alun R. Coker > >> Sent: 18 December 2008 22:45 > >> To: CCP4 bulletin board > >> Subject: Re: [ccp4bb] Transferring a Free R set. > >> > >> Hi All, > >> > >> I suspect using a new free assignment for each data set is a > >> fairly wide > >> spread mistake, which would be hard to pick up by a reviewer. > >> It would > >> be interesting to survey the pdb for ligand bound structures > >> which have > >> a different free R set from their parent (isomorphous) native > >> structures. I have heard from a colleague involved in testing > >> refinement programs that there is a surprising number of deposited > >> ligand structures where re-refinement shows that the > ligand is built > >> into noise. > >> > >> Would it be fair to say that in the case of a ligand structure, > >> isomorphous with the native structure but where different > R-free sets > >> have been used, we cannot use the R-free to validate > >> refinement of the > >> ligand? My understanding is that it would be contaminated by > >> the phases > >> brought over with the model from the native data set, and is > >> no longer > >> independent, so any ligand density could just arise from over > >> fitting. > >> How could we test/retrieve such a situation? Will omit maps > >> made after > >> a round of REFMAC refinement without the ligand still have a > >> memory of > >> our over fitted model which would persist through to new > maps? If so, > >> would this be the case after a slow cool in CNS? > >> > >> Does the community still think it is valid to use a slow cool > >> to "reset" > >> the free R? > >> > >> Alun. > >
Re: [ccp4bb] Transferring a Free R set.
Hi Ian, Gerard, Thanks for your replies. Perhaps my hypothetical example wasn't a good one and may have clouded the issue. Could I just confirm whether or not you think its necessary to transfer initial free R assignment to any new data sets or to isomorphous data sets such as substrate complexes. Thanks, Alun. Ian Tickle wrote: Hi Alun I've never seen a soaked ligand-bound structure that's truly isomorphous with the native - simply soaking & freezing the crystal is virtually guaranteed to introduce very serious non-isomorphism, e.g. up to 5% change in one or more cell parameters. At least, we always have a MR step to solve the structure, as rigid-body refinement alone often fails to pull it in, so the R-factors always start high (> 0.4) anyway. This renders the concept of a common test set totally meaningless, since the native & ligand-bound reflections with the same indices (except maybe the low res ones) will not be in equivalent positions in reciprocal space, and in any case once it has been refined down & been through a few rounds of rebuilding any bias will very quickly disappear. Note that Rfree is a meaningful statistic only at convergence at the likelihood maximum, and at convergence all previous 'memories' of the structure are erased: bias due to 'contamination' of the test set only occurs as a result of failure to ensure convergence. Also I don't believe that any of the cases of deposited ligand structures where either there is clearly no density for the ligand, or the ligand is in the wrong place, or even the wrong ligand has been built, can be laid at the door of the test set that was used. Plain incompetence and/or wishful thinking is a much more likely explanation. So personally I think you're worrying over nothing. Just my 2p's worth. Cheers & Seasons greetings to Steve & Jon! -- Ian -Original Message- From: owner-ccp...@jiscmail.ac.uk [mailto:owner-ccp...@jiscmail.ac.uk] On Behalf Of Alun R. Coker Sent: 18 December 2008 22:45 To: CCP4 bulletin board Subject: Re: [ccp4bb] Transferring a Free R set. Hi All, I suspect using a new free assignment for each data set is a fairly wide spread mistake, which would be hard to pick up by a reviewer. It would be interesting to survey the pdb for ligand bound structures which have a different free R set from their parent (isomorphous) native structures. I have heard from a colleague involved in testing refinement programs that there is a surprising number of deposited ligand structures where re-refinement shows that the ligand is built into noise. Would it be fair to say that in the case of a ligand structure, isomorphous with the native structure but where different R-free sets have been used, we cannot use the R-free to validate refinement of the ligand? My understanding is that it would be contaminated by the phases brought over with the model from the native data set, and is no longer independent, so any ligand density could just arise from over fitting. How could we test/retrieve such a situation? Will omit maps made after a round of REFMAC refinement without the ligand still have a memory of our over fitted model which would persist through to new maps? If so, would this be the case after a slow cool in CNS? Does the community still think it is valid to use a slow cool to "reset" the free R? Alun. -- Alun R. Coker University College London Division of Medicine, Royal Free Campus Centre for Amyloidosis and Acute Phase Proteins Rowland Hill Street London NW32PF Tel: +44(0)20 7433 2764 Fax: +44(0)20 7433 2776 Disclaimer This communication is confidential and may contain privileged information intended solely for the named addressee(s). It may not be used or disclosed except for the purpose for which it has been sent. If you are not the intended recipient you must not review, use, disclose, copy, distribute or take any action in reliance upon it. If you have received this communication in error, please notify Astex Therapeutics Ltd by emailing i.tic...@astex-therapeutics.com and destroy all copies of the message and any attached documents. Astex Therapeutics Ltd monitors, controls and protects all its messaging traffic in compliance with its corporate email policy. The Company accepts no liability or responsibility for any onward transmission or use of emails and attachments having left the Astex Therapeutics domain. Unless expressly stated, opinions in this message are those of the individual sender and not of Astex Therapeutics Ltd. The recipient should check this email and any attachments for the presence of computer viruses. Astex Therapeutics Ltd accepts no liability for damage caused by any virus transmitted by this email. E-mail is susceptible to data corruption, interception, unauthorized amendment, and tampering, Astex Therapeutics Ltd only send and receive e-mails on the basis that the Company is not liable for any such alteration or any consequences
Re: [ccp4bb] refmac 5.5.063 vs. 5.5.066
Hi Jan, the difference between 5.5.0063 and 5.5.0066 is just a few bugfixes. However, unfortunately one of them introduced a new bug to 5.5.0066 which has been fixed in the later versions. Could you please upgrade to the latest version (currently 5.5.0070) from http://www.ysbl.york.ac.uk/~garib/refmac/data/refmac_stable/ That should fix the problem (please let us know if it does not) Pavol.
[ccp4bb] Beamline scientist position EMBL-Grenoble
Dear all, I would like to inform you all of a vacancy we have for a Staff scientist position at the EMBL-Grenoble http://www-db.embl.de/jss/servlet/de.embl.bk.emblGroups.JobsPage/08054.html We currently looking for an experienced crystallographer at the EMBL-Grenoble for the mainentance, routine operation and further development of the highly successful microfocus beamline ID23-2 at the ESRF. http://www.esrf.fr/UsersAndScience/Experiments/MX/About_our_beamlines/ID23-2 For further information please contact Dr. Andrew McCarthy email: andre...@embl.fr Please send your application (cover letter, CV (in english) and contact information of three professional references) quoting ref. no. W/08/054 in the the subject line to applicat...@embl.de. The closing date is the 08.02.2009 We are looking forward to your applications and season's greetings to all, Andrew Andrew McCarthy, EMBL-Grenoble, 6, rue Jules Horowitz, 38042 Grenoble, France. Tel: +33 (0)4.76.20.72.76 08-054.pdf Description: Adobe PDF document
