Improvements introduced in rhofit 1.2.4 (05 May 2014)
- Fix bug that caused slightly incorrect correlation coefficient to be
reported when a ligand had any atoms named "C " or "CA "
or "O " or "N ".
- rhofit included in BUSTER academic release for the first time.
Improvements introduced in rhofit 1.2.3 (27 November 2012)
- rhofit has been extended so as to be able to handle cif dictionaries from CORINA
and PRODRG which lack atom coordinates in the dictionary itself. Previously it
was necessary to specify -lp ligand.pdb to handle such a dictionary.
Improvements introduced in rhofit 1.2.1 (June 2011)
- This is a minor release to address user requests to control rhofit output, in
particular the residue number used for the ligand and the number of hits
output (see Output options).
Improvements introduced in rhofit 1.2.0 (October 2011)
- Fitting is more robust in the presence of incorrect planarity
assignments in the input dictionary; planes which are not compatible
with a good fit to the density are deleted in the output dictionary
- The output directory contains CIF-format rather than TNT-format
dictionaries
- Fitting starting with a grade-generated dictionary is quicker,
since we trust grade's generated coordinates for the ligand and
don't have to regenerate them
- Torsion restraints are used when fitting with a grade-generated dictionary
- When running chiral scanning in a case with multiple clusters,
there is more of a distinction in the output between different
chirality choices and different cluster choices
Improvements introduced in rhofit 1.1.1 (April 2010)
- This is a minor release dealing with user requests and fixing bugs.
- rhofit no longer needs ccp4 utilities to be set up before it is invoked.
Instead it will use the ccp4 utilities already set up in $BDG/setup.csh or
$BDG/setup.sh.
- rhofit will no longer leave CifParser.log or #####.setvar.lis files
cluttering up the directory you run it in.
- Bugfix: ignored protein side chains (with large B range) are no longer left out of merged.pdb.
- User-requested options to manually control the clustering process.
Improvements introduced in rhofit 1.1.0 (March 2010)
- The fitting algorithm itself
- We now support starting with a refmac-compatible CIF-format dictionary for the input ligand; you are recommended to use this option if you have a reliable CIF dictionary for your ligand.
- If you specify an dictionary for a ligand then you need not specify any coordinates for it. Instead rhofit will generate initial optimized coordinates from the restraints.
- Clustering is more robust: if no clusters of the expected size are found at first, we relax the standards and try again. This is useful in cases where the ligand is partially disordered.
- Better sampling in the rhofit fitting algorithm means that it is much less likely to get stuck in local minima.
For instance, ligands with cyclohexyl-type rings are reliably output in unstrained chair conformations.
- Better ordering of ligand hits found.
- There is an option to handle ligands with hydrogens, provided you have a dictionary with them.
- The B-factors in the output ligand are now estimated; atoms not fitting the cluster are indicated with a B-factor penalty
- The -nochiral option works more reliably. This option allows chiral centres to invert at will during the run and can be much faster than -scanchirals for large ligands.
- Partial-occupancy heavy atoms in an input PDB file are fitted as such, which helps with fitting of radiation-damaged ligands
- Side chains with large B-factor range are ignored in clustering and in clash calculations - so we will fit correctly a cluster with a lysine sticking through it as a result of a poorly-refined input protein.
- We are less pedantic about the format of input PDB files for ligands.
- MTZ files produced by refmac are supported in the fitter and the visualiser.
- Temporary files generated during the fitting are now by default deleted afterwards.
- More useful output from rhofit
- A results.txt file listing correlation coefficients, ligand strain score and protein-clash metrics for each hit is generated.
- The command line used is recorded in results.txt so that you can see what job produced a given rhofit output directory.
- A merged.pdb file in which the best hit for each cluster are merged with the input PDB, and bad waters removed, to provide an input ready for refinement is generated.
- The cluster files themselves are recorded in the output directory.
- In the visualise-rhofit-coot visualiser
- Information about the strain and clash scores of the currently-visible conformation is displayed.
- You can (optionally) see the clusters which the fitter is using to define the fit region.
- Issue with selecting atoms which aren't currently displayed has been fixed.
First release of rhofit 1.0.0 (November 2009)