[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [ccp4bb]: Refmac vs. cns
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Dear All,
From my personal experience and from working with other CNX users on their
problems comparing performance of the different programs is a valuable and
productive way to learn when and how these programs can be applied. In
fact, X-PLOR/CNS/CNX and REFMAC-ARP/wARP can be quite complimentary to each
other depending on model quality, the amount and accuracy of X-ray data
available and type of refinement. As has already been stated, ARP/wARP
works the best with high resolution data, I have seen CNS/CNX do an
excellent job with data in the (3.5-2.2) A range, particularly when a
starting protein model is relatively crude. In many cases loops can move
more than 2 A to their correct positions and side-chains are correctly
placed into the density.
Another advantage that X-PLOR/CNS/CNX provides is the ability to
write/merge various task files in order to make your own refinement protocols.
To summarize, a direct comparison of REFMAC-ARP/wARP and X-PLOR/CNS/CNX
should not focus on which program is better or worse but rather how one can
benefit from using many programs with different data.
Cheers,
Igor
At 05:11 PM 1/21/2002 +0000, James Naismith wrote:
>*** For details on how to be removed from this list visit the ***
>*** CCP4 home page http://www.ccp4.ac.uk ***
>
>Dear All,
>
>I pretty much agree with everything Dirk has said. I pretty much shudder at
>my late 80's low R-factors which I annealed. I am sure CD will do me one day
>(but hey I deposited the data).
>
>I find torsional dynamics excellent for getting errors out of mol repl slns
>where bias is a severe problem at 2.5 + A. Like Dirk I chuck the model (I
>use them as bkgrnd objects) but keep the omit maps.
>
>CNS has some very nice features, the ability to define NCS is very strong
>also dictionary design is easier.
>
>I have hacked a script because I am old and stupid which runs refmac and
>arp_warp in a cyclical manner. It is absolutely great for putting in waters
>etc. Use the gui where possible.
>
>A side note is if you ask any program to put waters in, it will put waters
>in. It does not put in glycerols, sulfates etc etc. I only run my script
>once I have by hand checked my maps to put these things in first. If you
>don't do this serious errors can result.
>
>When using unusual ligands, I struggle (see above I'm dumb) to avoid refmac
>and arp_warp choking on each others atom types.
>Here is my work around
>
>grep -v 'XYL' {$name}{$count}.pdb > test.pdb
>grep 'YXL' {$name}{$count}.pdb > pre_arp_xyl.pdb
>
>cat test.pdb test_xyl.pdb > pre_arp_{$name}{$count}.pdb
>
>/bin/rm {$name}{$count}.pdb
>
>$CBIN/arp_warp \
>xyzin pre_arp_{$name}{$count}.pdb \
>mapin1 arp_2fofc.map \
>mapin2 arp_fofc.map \
>xyzout post_arp_{$name}{$count}.pdb \
><< eof-arp
>mode update waters
>symmetry p212121
>resolution 100 1.8
># refine waters
>find atoms 50 chain W cutsigma auto
>remove atoms 8 cutsigma 1 merge 2.2
>refine waters
>end
>eof-arp
>#
>#
>
>grep -v 'xyl' post_arp_{$name}{$count}.pdb > post_arp_test.pdb
>cat post_arp_test.pdb pre_arp_xyl.pdb > {$name}{$count}.pdb
>
>Replace the arp_warp below in the iterative script with the above. If my
>funky ligand is xyl, the script strips out the latest refined coords puts
>them in a file. It then adds some coordinates you specify to this file
>before going into arp.
>Your specified file has the starting coordinates of your ligand but with
>atom names warp_arp likes (C1- C100 is fine!). This stops arp_warp choking
>but it also stops it putting water in where your ligand should be. After
>arp_warp the script strips out the place holder ligand putting back in your
>refmac refined coordinates for xyl.
>This has some advantages
>(1)It works easily and avoids a headache. You've spent ages getting that
>refmac dictionary right, have a coronary trying to get it compatable with
>warp_arp is just a pain. In water mode you just don't want waters where your
>ligand should be.
>The assumption is that atoms in your ligand have not moved more than 2.0A.
>If you put in right at the start and refined it already this is reasonable.
>By inserting more greps you can strip more funky ligands etc.
>
>
>Automated insertions of waters
>
>#!/bin/csh -f
>#
># user input
>#
>set name = 'yfpjn_'
>set count = 0
>set cycles = 30
>#
>#
>#
>set last = 0
>@ last = $count + $cycles
>#
>set current = 0
>set next = 0
>#
>while ( $count != $last )
>#
>@ count ++
>@ current = $count - 1
>@ next = $count + 1
>#
>echo '###################################'
>echo ' CYCLE-NO. '$count
>echo '###################################'
>#
>#
>#
>
>refmac5 \
> XYZIN {$name}{$current}.pdb \
> XYZOUT {$name}{$count}.pdb \
> HKLIN data.mtz \
> hklout {$name}{$count}.mtz \
> tlsin {$name}{$current}.tls \
> tlsout {$name}{$count}.tls \
> <<eof-refmac
>LABIN FP=F SIGFP=SIGF FREE=FreeR_flag
>LABO FC=FC PHIC=PHIC FWT=2FOFCWT PHWT=PH2FOFCWT -
> DELFWT=FOFCWT PHDELWT=PHFOFCWT FOM=FOM
>REFI TYPE RESTrained
>refi tlsc 10
>REFI RESI MLKF resol 50 1.5
>REFI BREF isotropic
>REFI METH CGMAT
>WEIG EXPE MATR 1.35
>NCSR NCHAIN 2 CHAINS A B NSPAN 3 1 44 1 46 177 1 182 202 1
>SCAL TYPE BULK LSSC ANIS FIXBulk BBulk 200
>#SCAL LSSC ANIS
>make_restraints hydrogens no
>SOLVENT YES
>NCYC 15
>MONI MEDI
>!BINS 15
>PNAME YFP
>DNAME EPIM NATIVE
>USECWD
>
>make newligand noexit
>make check 0
>
>END
>eof-refmac
>#
>
>#
>grep "REMARK 3" {$name}{$count}.pdb
>
>#
>#
>#
>#
>#
>#
>#
>#
>#
>fft \
>hklin {$name}{$count}.mtz \
>mapout map2fo.temp
><< EOF-fft > /dev/null
>labin F1=2FOFCWT PHI=PH2FOFCWT
>fftsymmetry p21212
>END
>EOF-fft
>#
>mapmask mapin map2fo.temp mapout arp_2fofc.map << eof > /dev/null
>xyzlim 0 1 0 0.25 0 1
>symmetry p21212
>eof
>#
>fft \
>hklin {$name}{$count}.mtz \
>mapout mapfo.temp \
><< EOF-fft > /dev/null
>labin F1=FOFCWT PHI=PHFOFCWT
>fftsymmetry p21212
>END
>EOF-fft
>#
>mapmask mapin mapfo.temp mapout arp_fofc.map << eof > /dev/null
>xyzlim 0 1 0 0.25 0 1
>symmetry p21212
>eof
>#
>#
>
>cp {$name}{$count}.pdb pre_arp_{$name}{$count}.pdb
>
>
>$CBIN/arp_warp \
>xyzin pre_arp_{$name}{$count}.pdb \
>mapin1 arp_2fofc.map \
>mapin2 arp_fofc.map \
>xyzout post_arp_{$name}{$count}.pdb \
><< eof-arp
>mode update waters
>symmetry p21212
>resolution 100 1.5
># refine waters
>find atoms 50 chain W cutsigma auto
>remove atoms 8 cutsigma 1 merge 2.2
>refine waters
>end
>eof-arp
>#
>#
>
>cp post_arp_{$name}{$count}.pdb {$name}{$count}.pdb
>
>
>
>echo 'Finished internal cycle'
>end
>echo 'Finished big cycles'
>
>
>
>
>
>refmac5 \
> XYZIN {$name}{$count}.pdb \
> XYZOUT final_{$name}{$count}.pdb \
> HKLIN data.mtz \
> hklout final_{$name}{$count}.mtz \
> tlsin {$name}{$count}.tls \
> tlsout final_{$name}{$count}.tls \
> <<eof-refmac
>LABIN FP=F SIGFP=SIGF FREE=FreeR_flag
>LABO FC=FC PHIC=PHIC FWT=2FOFCWT PHWT=PH2FOFCWT -
> DELFWT=FOFCWT PHDELWT=PHFOFCWT FOM=FOM
>REFI TYPE RESTrained
>refi tlsc 10
>REFI RESI MLKF resol 50 1.5
>REFI BREF isotropic
>REFI METH CGMAT
>WEIG EXPE MATR 1.35
>NCSR NCHAIN 2 CHAINS A B NSPAN 3 1 44 1 46 177 1 182 202 1
>SCAL TYPE BULK LSSC ANIS FIXBulk BBulk 200
>#SCAL LSSC ANIS
>make_restraints hydrogens no
>SOLVENT YES
>NCYC 15
>MONI MEDI
>!BINS 15
>PNAME YFP
>DNAME EPIM NATIVE
>USECWD
>
>make newligand noexit
>make check 0
>
>END
>eof-refmac
>
>
>
>
>grep "REMARK 3" final_{$name}{$count}.pdb
>
>fft \
>hklin final_{$name}{$count}.mtz \
>mapout map2fo.temp \
><< EOF-fft > /dev/null
>labin F1=2FOFCWT PHI=PH2FOFCWT
>fftsymmetry p21212
>END
>EOF-fft
>
>fft \
>hklin final_{$name}{$count}.mtz \
>mapout mapfo.temp \
><< EOF-fft > /dev/null
>labin F1=FOFCWT PHI=PHFOFCWT
>fftsymmetry p21212
>END
>EOF-fft
>
>
>
>mapmask:
>mapmask mapin map2fo.temp mapout 2fofc_xyl.map \
>xyzin final_{$name}{$count}.pdb << eof > /dev/null
>border 15
>symmetry p21212
>eof
>
>mapmask:
>mapmask mapin mapfo.temp mapout fofc_xyl.map \
>xyzin final_{$name}{$count}.pdb << eof > /dev/null
>border 15
>symmetry p21212
>eof
>
>maps:
>
>echo read m1 fofc_xyl.map ccp4 > script
>echo brix m1 fofc_yfp.omap >> script
>echo read m2 2fofc_xyl.map ccp4 >> script
>echo brix m2 2fofc_yfp.omap >> script
>echo quit >> script
>
>run mapman -b < script > mapman.log
>
>
>
>/bin/rm -f *.temp
>
>best
>Jim
>
>James H. Naismith | Research mailto:naismith@st-and.ac.uk
>Professor of Chemical Biology | Teaching mailto:jhn@st-and.ac.uk
>BBSRC Career Development Fellow |
>Centre for Biomolecular Sciences | Office: 1334-463792 (24 hr)
>The North Haugh | Fax : 1334-467229
>The University | Lab : 1334-467245
>St. Andrews | In UK add 0 to start of number
>Fife Scotland, U.K., KY16 9ST | www http://speedy.st-and.ac.uk/
==========================================
Igor Mochalkin, PhD.
X-ray Product Specialist, Life Science
Accelrys Inc.
9685 Scranton Road
San Diego, CA 92121-3752
Phone: (858) 799-5224
Fax: (858) 799-5100
Mobile: (858) 945-5224
e-mail: igor@accelrys.com
URL: http://www.accelrys.com
+++++++++++++++++++++++++++++++++++++++++++