Since the last Newsletter, versions 2.15, 3.0 and 3.1 of the CCP4 suite
have been released. Version 3.0 included a significant addition to the MTZ
format, namely the use of Missing Number Flags (MNF) to indicate missing data.
Most programs now read and write MNFs, and it is recommended that
MNFs are incorporated into existing data files using the jiffy MTZMNF.
With a policy for handling missing data, it is now straightforward
(and strongly recommended) to work with complete datasets, i.e. to
include all reflections up to the high resolution limit whether
or not the corresponding datum is present (see the UNIQUE documentation
for more details). A Users' Guide to MNFs can be found at
http://www.dl.ac.uk/CCP/CCP4/user.html
In addition to improvements to existing programs, a number of new
programs were introduced. In particular, have a look at:
RESTRAIN from the Birkbeck group: refinement program including,
amongst many options, TLS refinement.
REFMAC from Garib Murshudov: Maximum Likelihood refinement program.
SOLOMON from Jan-Pieter Abrahams: density modification
program using solvent flipping.
XDLDATAMAN, XDLMAPMAN from Gerard Kleywegt: manipulation and reformatting
of data and map files.
Version 3.1 is primarily to fix a number of bugs contained in
version 3.0, and to do some general tidying. The most
significant/visible changes are as follows: Missing data in MTZ
files are now listed by MTZDUMP as the string 'MNF'. ROTAPREP can now
convert user-defined formats to a multi-record MTZ file (INPU USER).
The functionality of the script $CETC/uniqueify for completing a
dataset has been extended, see unique.doc for details. A version
for VMS, $CETC/UNIQUIFY.COM, has been added. There is a patch file
from v3.0 to v3.1.
A full list of changes since version 2.14 (announced in the
last newsletter) is given below. Details of program fixes since
version 3.1 have been placed on a Problems Page.
Perennial problems appear on our FAQ.
If you still have problems, then email us at
ccp4@dl.ac.uk.
Finally, CCP4 will have a booth at the Commercial Exhibit Show
of the XVII IUCr Congress in Seattle. Come and visit us!
Changes to the Suite
Building etc.
configure: fftbig, tracer and amore have individual compile
flags for AIX. OSF/1 optimisation reduced to -O1; problem
with FFT SCALE card suggests compiler problem.
configure tries to avoid messing with the `unsupported' directory if it's
not present.
HPUX: don't set (invalid) amore (non-)optimisation flags. The correct one
reported to be needed, though.
include/ccp4.setup-dist: Incorrectly (for any system) set _RLD_LIST,
causing chaos with dynamic loading on Irix and similar systems.
MAKEFILE.COM: typo in linking of refmac corrected. Can now compile
just one designated program with no checks.
Fix problems building procheck on VMS.
ccp4/etc/PROCHECK*.COM: hopefully the problem with PRODIR$ is now
sorted.
XCCPJIFFY: SUNOS_5 added to Imakefile, SUNOS_4 hints in comments.
Irix6 configuration is provided, your mileage may vary -- it did for me.
Feedback still wanted.
Program changes
General: A number of programs were found to close files on program
termination. This may cause problems on some machines. Here
is a list of progs. that were changed (not all programs have been
checked): abscale, act, agrovata, almn, contact, coordconv, crossec,
extend, f2mtz, fft, fftbig, gensym, geomcalc, hgen, hklplot, icoefl,
mlphare, mtz2various, mtztona4, npo, pdbset, peakmax, polarrfn, prolsq,
protin, rfcorr, rotaprep, rstats, sortwater, undupl, vectors, volume.
AMORE: The HKLM keyword has been reinstated. Default NM=25. Fix: all
0kl reflections of table were 0. Keyword END now accepted. Fix possible
problems with inverse FFT due to dimensioning error. Fix failure to get
FFT prime factors right. Less likely to overflow limits on HKL. SORTFUN
can generate a table -- enables use of Es for TRAFUN. Maximum value for
TRAFUN NMOL increased to 9. The ROTATE NSR keyword reinstated.
CAD: Allow longer LABIN/LABOUT lines. Accept LABIN FILE <i> ALL.
The scale factor is now applied to the correct columns and not to
inappropriate ones like phase.
CONTACT: The output has changed to: Res type, Res ID, Atom name.
DM: correct multi-domain averaging bug. DM 1.6 now current version.
ECALC: Use LABIN and LABOUT rather than FOBS/ISODIFF -- incompatible
change. Introduced SCALE card to scale F and F2OR.
F2MTZ: Deal with SHELX free R.
FFT/FFTBIG: A new keyword has been added FILLIN which substitutes F2 for
F1 if F1 is missing. This only works if the MTZ file has been properly
converted to Missing Number Flags.
FFTBIG: Compiler/runtime problems caused by undefined LNAN fixed.
FHSCAL: Resolution card added so that data can be truncated in output
file. Problem with eof on input corrected. LIST can take a no. > no.
of refls. without crashing. Statistics for old style MTZ the same as
new.
FREERFLAG: Default semantics of free R column changed, along with its
type: it's picked from a set of integers, allowing full cross-validation.
An existing free R set may be completed when reflections are added (only
if using missing data conventions).
Systematic absences are now not assigned a freeR flag but
are flagged as missing. When using the COMPLETE keyword the freeR flag
information is taken from the MTZ file and not the keyword input.
HKLPLOT: Redimensioning, improved output.
HKLVIEW: minv replaced by minv3 because of routine renaming in
modlib.f
IPDISP: Support MacScience scanner. spdfil.mar_n now image type
"bigovf", to indicate overflow
table. ipdisp -m option points to spdfil.mar_n (bigend rather than
littleend image)
MAMA2CCP4: problem fixed wrt quotes in call to ccperr.
MAPMASK: Symmetry expansion speeded up. New SIGMA option for SCALE
(useful when contouring).
MLPHARE: Phasing power was printed as 0.0 with PRINT AVE. SCALE FP option
removed. Resolution in Angstroms is added to tables but not plotted on
xlog-graphs. SCRIPT keyword will produce either a VMS COM file or unix
script with refined heavy atom parmeters. Ensure Angstrom coordinates are
in PDB format - Orthogonal code 1. Will now produce NaNs for unphased
reflections. Compiler problems caused by undefined CCPNUN fixed.
MTZ2VARIOUS: Remove RESTRAIN input type (never actually necessary
since restrain can specify the input format) now restrain reads MTZ. OUTPUT
USER will now output DP and SIGDP. For USER the ordering of the columns is
taken from the ordering on LABIN not LSPRGI table. If reflection record
has NaN, by default reflection not output. MISS keyword will output them.
MTZDUMP: Missing data identified through MNFs. File statistics table
now gives single "Mean" and "Mean abs." which is calculated from all
non-missing data in a column. LRES option added to list resolution.
More flexible statistics available via STAT keyword. Missing reflections
are printed as "MNF". The VALM keyword has been removed.
MTZUTILS: Fix zeroed columns from hklin2 for reflexions not in hklin1 with
UNIQUE option.
NA4TOMTZ: Fix incorrect call of `qnan'. (Unlikely to have caused problems
before missing data changes.)
NCSMASK: A crude `sphere mask' can be constructed by making a PDB file
with only a few atoms and specifying a large radius. Output map grid was
always 3A round coordinates -- now 2A+.
NPO: colours are now consistant with those defined in XCCPJiffy. CHAIN
sub-keyword removed. Input handling improved a little. JOIN RADII HBOND
now works as it should. Msecst has been reinstated. Can now have maps
stacks like 3,5,7 or 5,10,15 etc.. RADII BOND ALL fixed.
OVERLAPMAP: changed code for EXCLUDE option. Will have non-zero point in
mapout if |grid_point| < 1.0e-06 in map1.
PEAKMAX: symmetry related peaks at the edge of the map are not output.
PDBSET: new keywords added EXCLUDE HEADERS, EXCLUDE HYDROGENS and SEQUENCE
[PDB|SINGLE].
PHISTATS: output improved + documentation.
PLTDEV: `-pen' option now works (e.g. for colour PostScript). Colours
are now consistant with those defined in XCCPJiffy.
PROLSQ: the keyword RTEST has been reconstructed. Now you can only have
RTEST -1|0. This means that you cannot recalculate structure factors i.e.
RTEST +1.
PROCHECK: Updated to v. 3.3.2 plus fixes. Now the first atom in *.new is
printed only once. CONVAX has now been added to distribution.
PROTIN: Didn't find contacts between the same residue on symmetry related
molecules; error in packed symmtery number caused some contacts not to be
picked up by prolsq; increase some dimensions; error in checking special
inter-chain distances; no longer monitors short hydrogen bonds. Can now
use atom names which represent ions e.g. Fe+2, U+6 etc. Non-protein residue
atoms all assigned as side-chain not as main-chain and side-chain.
Error in handling of multiple disordered planar groups fixed?
Atoms now must have VDW radii defined. Hence, changes in dictionaries.
RFCORR: arrays for PARSER redimensioned to avoid problem with TITLE. Polar
angles are output correctly when have exact 2-fold axes.
ROTAPREP: Format for SCALEPACK options changed from F8.1 to F8.0. Also,
updated man page. Can now convert user-defined format to multi-record MTZ
file (INPU USER).
ROTMAP: the table printed after the OUTPUT ... keyword was incorrect, now
fixed.
RSPS: Failed to handle SGs with Patterson symmetry groups 143-145.
RSTATS: OUTPUT ASIN changed to LABOUT ALLIN. PROCESS FOBS should work.
Ranges in F applied to Fobs only. Xloggraph altered and new column added
ln(/) v resolution.
SCALEIT: failed for more than 9 derivatives
SFALL: Reads formfactors automatically if not specified, but still no way
to distinguish Fe Fe+2, Fe+3 except to enter manually. Obsolete keywords
removed.
SIGMAA: `PARTIAL DAMP' didn't work. HLAC ... HLDC now output correctly.
Map cofficients can now be constructed for missing reflections.
SOLOMON: (solomon.c) parameters for rbfro1 were undefined.
SORTMTZ: now not only does there need to be an equal number of columns in
each file but the column types must be equal.
STEREO: keyword CONSTRAIN could not be used, now working.
SURFACE: trapped uninitialised variable, important for output.
TRACER: prog. now works correctly.
TRUNCATE: MTZ columns were potentially scrambled for data not from
agrovata containing anomalous data.
ZEROED: initialised some variables to stop IOT trap failure.
CROSSEC: Interpolate cross sections and compute anomalous scattering
factors (thanks to Don Cromer).
GEOMCALC: Geomtery calculations on molecule (some overlap with `angles')
HGEN: Generates hydrogen atoms for a protein coordinate file with standard
geometry.
MAPROT: Map skewing, interpolating, rotating and averaging program.
MTZMNF: Attempts to identify missing data in an MTZ file, and replaces
the entry with a Missing Number Flag (e.g. NaN).
REFMAC: Garib's Maximum Likelihood refinement program. Brand new from York.
RESTRAIN, TLSANL: `restrain' does constrained anisotropic thermal
parameter refinement using the translation/libration/screw-rotation model.
Its feature list says: refinement with restrained geometry, rigid body,
use of amplitude and phase observations, group anisotropic displacement
parameters, disordered solvent corrections (thanks to Birkbeck group).
`tlsanl' provides TLS analysis. Note restrain's keyword input isn't
in the standard CCP4 form.
SOLOMON: density modification (phase improvement) by solvent flipping.
SORTWATER: Replaces `watersort'.
xdlmapman, xdldataman: have been added thanks to Gerard K.
Library
QNAN from the diskio routines (almost unused) changed from a function to a
subroutine to avoid lossage with f2c's calling conventions.
protin.dic: Last block wasn't read. WAT residue changed to contain more
extensive list of waters (oxygen) only. All other atoms (ZN1,CU) have been
removed. All atoms have VDW radii defined. Changed OXT and OE
for NFORMYL and NACETYL so they are consistant with OT.
protin_jp.dic: extra terminal atoms (OE, OXT) added to terminal
residues similar to protin.dic.
protin_vl.dic: extra terminal atoms added to terminal
residues similar to protin.dic.
symop.lib: Errors corrected in SGs 43, 112, 115, 160, 210, 228, 230 (a
subset of errors reported with the X-PLOR version...)
library.nw removed after too many complaints -- source is now library.c
lib/data/atomsf.lib has comment section added; some programs which read it
altered as a consequence.
lib/data/crossec.lib: Data file for `crossec'.
CCPOPN uses /dev/null and NL: interchangeably
Fatal errors on diskio files now report the filename in question.
Packed image routines (for mosflm, ipdisp): New version.
MODLIB: MINV has been renamed to MINV3. A new routine has been added to
calculate the inverses of nth order matrices MINVN.
MTZLIB: Facilities for dealing with missing data items. [Null attributes
for relational tuples? -- surely not!] Writing files with this facility
should probably not be done since the programs aren't guaranteed to
swallow the results.
PARSER: Backslash is now a continuation character like `&' and `-'.
SYMLIB: new routine HKLRANGE. Maximum no. of symmetry operators
increased to 192. Now consistent with MTZLIB. Error in handling P2 C
unique corrected.
symop.lib: correct groups 120, 166, 210, 219, 220, 227, 228. File now
subjected to the overdue check for fulfilling group conditions on the
operators.
x-windows/XCCPJIFFY/XCCPJiffy: colours changed to be consistant with those
defined in NPO. Also, Xloggraph and xplot84driver pen colours are the same.
Altered write statements so they do not exceed vms write limit of 132
characters.
Examples
*.log-dist: all files except completedata and na4tomtz have been removed
from the $CEXAM/toxd area. Some files have been added to the
$CEXAM/unix/runnable area.
All examples updated, inline with current progs.
$CETC/uniqueify: original cell now retained rather than rounded up version.
Problem with some refls. not being assigned a FreeR flag overcome.
-f switch modified to accept
Documentation
Several .doc files corresponding to man .1 files were missing. diskio.doc
was out-of-date.
Various fixes/updates to documents for amore, ccplib, contact,
libsymmetry, lsqkab, ncsmask
FAQ: the plain text version with searching scripts (ccp4faq) are avialable.
Just type 'ccp4faq'.
HGEN: Included this time.
MLPHARE: more information about using MAD data.
MTZMNF: Clarified that columns not specified in LABIN are output
unchanged.
MTZUTILS: RZONE described correctly now.
SOLOMON: MULSOLV keyword incorrectly referenced in the doc as
SLVMUL.
UNIQUE: Improved description of uniqueify script.
Out-of-date manual-usletter.ps removed -- the current one should print OK
on letter paper anyhow.
Updates to mtzlib, npo, sfall, sigmaa documentation and the ccp4.1 man
page.
Withdrawal
COMPLETE removed, see unique.doc for replacement procedure.
FAQ in $CCP4 removed. It will be added to FAQ on the Web.
ENVELOPE, FLATMAP, RESLICE, PRMAP moved unsupported status.
LCF2MTZ, EXCHANGE and LCFLIB removed in the hope that everyone's using MTZ
files -- still available from Daresbury if anyone is in need.