CHANGE LOG for PHENIX distribution
                ==================================

Version 1.6.4 (July 2010)
=========================

- Bug fix release
- Multi-criterion plot in validation

Version 1.6.3 (July 2010)
=========================

- General:
  - SOLVE converted to C++; FORTRAN compiler no longer required
  - CCP4 map output supported for all applications that write XPLOR maps
  - New commands: phenix.superpose_maps, phenix.apply_ncs,
    phenix.phase_and_build, phenix.build_one_model, phenix.assign_sequence

- Graphical interface:
  - Supports execution of jobs on Sun Grid Engine queueing system
  - Maps for PyMOL now output in (smaller, faster) CCP4 format
  - new GUIs for phenix.find_ncs, phenix.superpose_maps
  - support for execution on Sun Grid Engine queueing system
  - display of sequence and secondary structure in wizards and validation

- phenix.refine:
  - improved map parameters (identical to phenix.maps)
  - new custom_nonbonded_symmetry_exclusions atom selection parameter
  - reference model refinement now supports selections for both the
    reference and working models (refinement.reference_model.reference_group)
    - this method is most useful in cases such as the reference model having
      one chain and the working model having multiple copies of the same
      chain, all of which should be restrained to the single chain in the
      reference model.

- phenix.autosol:
  - much faster and better model-building after density modification

- phenix.find_ncs:
  - now supports identification of NCS operators from density map

- phenix.fit_loops:
  - two main algorithms now used: fit short gaps using a loop library derived
    from high‐resolution structures in the PDB, or build loops directly by
    iterative extension with tripeptides
  - the loop‐library approach (specified with loop_lib=True) is very fast,
    and is currently applicable for short gaps of up to 3 residues
  - the iterative extension approach is slower, but can be used for longer
    gaps (typically up to 10‐15 residues)

- phenix.superpose_maps:
  - transforms maps to follow a molecular superposition
  - output in CCP4 or XPLOR format

- phenix.phase_and_build:
  - carries out an iterative process of building a model as rapidly as
    possible and uses this model in density modification to improve the map
  - up to 10x faster as AutoBuild, but model quality is nearly as good

- phenix.build_one_model:
  - quickly build a single model from a map and sequence file, or extend an
    existing model. 
  - example of use (for building a new model):
      phenix.build_one_model my_map.mtz my_sequence.dat

- phenix.assign_sequence:
  - carry out an improved sequence assignment of a model that you have already
    built
  - once the sequence has been assigned, this method will use the sequence
    and proximity to identify chains that should be connected, and it will
    connect those that have the appropriate relationships using the new loop
    libraries available in phenix.fit_loops

- phenix.apply_ncs:
  - apply NCS operators from a my_ncs.ncs_spec file (from phenix.find_ncs) to
    a single copy of your protein to create all the NCS‐related copies
  - example of use:
      phenix.apply_ncs my_ncs.ncs_spec my_model_one_ncs_copy.pdb

Version 1.6.2 (June 2010)
=========================

- Disk image-based installation for Macs

- Graphical interface:
  - new GUIs: eLBOW, phenix.cif_as_mtz
  - structure comparison tool: evaluate multiple structures of a single
    protein, and identify differences in geometry; superposes models and maps
    for viewing in Coot and PyMOL
  - rearranged phenix.refine, validation, and wizard GUIs for wider screens
  - many bug fixes

- phenix.refine:
  - automatic flipping of incorrect N/Q/H rotamers
  - custom planarity restraints
  - hydrogen-bond restraints for Watson-Crick base pairs
  - reflections with Fobs=0 used in refinement
  - automatic optimization of mask parameters
  - reference model restraints: used to steer refinement of working model
    - advantageous in cases where the working data set is low resolution, but
      there is a know related structure solved at higher resolution.
    - the higher resolution reference model is used to generate a set of
      dihedral restraints that are applied to each matching dihedral in the
      working model.
    - to use (also available in GUI):
        main.reference_model_restraints=True
        reference_model.file=my_reference.pdb
  - Alternate ideal angles for torsion definitions: CIF definitions in geostd
    now have an 'alt_value_angle' tag for torsion angle definitions.
    'alt_value_angle' allows for non-periodic specification of target angle
    values, and can take a single value or a set of comma-separated values.
  - phenix.refine can now write a geometry restraints file *after* refinement
    for comparison with the standard .geo file that is written *before*
    refinement.  To turn on this option, use output.write_final_geo_file=True.
    After refinement, a <run name>_final.geo file will be written.

- New program - KiNG:
  - The Java kinemage viewer KiNG (Protein Science 2009;18:2403-2409) can
    now be launched via 'phenix.king'.  KiNG requires a Java version of
    1.5.1 or later to be installed on your system.

- phenix.hyss:
  - Phaser rescoring in HySS: Trial solutions after real/reciprocal space
    recycling are evaluated using the SAD likelihood target in phaser (only
    applicable to SAD). LLG scores are then used for identifying top solutions
    and terminating the search if a clear solution has been found.

- phenix.maps:
  - B-factor sharpening for improving low-resolution maps

- phenix.mtz2map:
  - new command for converting map coefficients to XPLOR or CCP4 format maps

- phenix.ligand_identification:
  - Incorporated real-space refinement in ligand fitting
    - keyword: refine_ligand=True (default)
  - Added intermolecular non-bonded interaction terms in ligand scoring and
    ranking
    - keyword: non_bonded=True (default)

- more robust build of Fortran programs (solve, resolve_pattern) with
  static ifort libraries

Version 1.6.1 (March 2010)
==========================

- New commands:
  - phenix.fmodel: structure factors from PDB file, as real or complex numbers
  - phenix.maps: generate any map coefficients or XPLOR maps

- Graphical interface:
  - numerous bug fixes
  - support for "detached" processes run outside GUI
  - new GUIs for phenix.superpose_pdbs, phenix.get_cc_mtz_mtz,
    phenix.get_cc_mtz_pdb, phenix.fmodel
  - Coot interface for running phenix.sculptor
  - general maps GUI now uses phenix.maps
  - validation GUI consolidated, works with or without reflections
  - alpha-test features available via preferences setting
  - 20x faster loading of PDB files in selection editor

- phenix.refine
  - much faster rotamer fixing
  - initial implementation of secondary structure restraints (hydrogen bonds)
    - automatic assignment of sec. str. using KSDSSP (from UCSF CGL)
  - proper treatment of charges on metal ions for scattering factors

Version 1.6 (January 2010)
==========================

- C++ version of RESOLVE

- Graphical interface:
  - new interfaces for Phaser (MR and SAD), AutoMR, ligand identification,
    ensembler and sculptor tools, NCS identification
  - reflection file editor:
    - more options for manipulating data arrays
    - assign R-free flags in thin shells
    - allow changing of symmetry by expanding or merging reflections
  - maps GUI:
    - allow custom map types
    - more options for creating XPLOR-format maps
  - improved startup time for main GUI and several apps
  - preferences setting for font sizes
  - expanded documentation
  - numerous bug fixes

- phenix.refine:
  - support for residues located on symmetry elements:
    - Atoms with occupancy == 1 are subject to automatic handling of the
      symmetry.
    - Atoms with occupancy != 1 are assumed to take the special position
      factor into account in the occupancy. Automatic handling of the
      symmetry is turned off.
    - If all atoms in a residue have occupancies != 1, all nonbonded symmetry
      interactions between the atoms in the residue are disabled.
  - Important bug fix affecting handling of branched oligosaccharides:
    - apply_cif_link can now be applied multiple times to the same residue
  - RNA sugar-pucker specific restraints, slightly revised restraints
    for all RNA residues, automatic chain links to non-standard RNA/DNA
  - Revised PRO dihedral angle restraints compatible with "endo" rotamer
  - Torsion-angle simulated-annealing bug fixes and small enhancements
  - fix_rotamers bug fixes

- Wizards:
  - optional real-space refinement of ligands in LigandFit

- New commands:
  - iotbx.reflection_file_editor: command-line version of graphical editor
  - New experimental standalone real-space refinement utility: mmtbx.lockit
    Reads a pdb file and map Fourier coefficients from an mtz file.
    (No documentation yet.)

Version 1.5 (September 2009)
============================

- General:
  - Improved installer and application launchers for Mac
    - Installation to /usr/local is no longer required
  - Nightly builds of installer available on web
  - Extensive additions to FAQs pages
  - new commands: phenix.fit_loops, phenix.fobs_minus_fobs_map,
    phenix.clashscore, phenix.grow_density, plus many graphical programs

- Graphical interface:
  - new main GUI (old GUI still used for AutoMR)
  - two-way communication with Coot, automatic updating of model/maps
  - phenix.reflection_file_editor:
    - Combine data from multiple reflection files, extend R-free sets
    - Any input format supported, output as MTZ
  - phenix.create_maps:
    - Generate standard likelihood-weighted maps, e.g 2mFo-DFc, mFo-DFc,
      anomalous, etc.

- LABELIT added to Phenix distribution (http://cci.lbl.gov/labelit)
  - autoindexing program designed to correct commonly experienced problems
  - detects pseudotranslation in diffraction patterns
  - can interface with MOSFLM for integration of data

- Phaser:
  - updated to latest version of Phaser (2.2)
  - reduced memory requirements
  - for anisotropic data, map coefficients sharpened to falloff of strongest
    direction
  - SAD phasing features
    - improvements to SAD log-likelihood-gradient substructure completion
    - automated reassignment of atom types, when more than one type of
      scatterer
    - improved sensitivity of LLG completion protocol
  - MR features
    - facility to compare MR solutions with prior ("template") solution, e.g.
      from another program
    - refine relative B-factors of different components of MR solution
    - recognize internal symmetry in MR model, use to identify equivalent
      solutions
    - updated atom names consistent with PDBv3
  - phenix.sculptor:
    - new tool to automate selection and editing of molecular replacement (MR)
      models
  - phenix.ensembler:
    - new multiple superposition tool to automate construction of ensembles
      for MR

- phenix.refine:
  - Torsion-angle dynamics for simulated annealing
  - Real-space refinement using "simple" and difference density map targets
  - Allow custom constrained occupancy groups in occupancy refinement
  - Make use of electron density correlation and electron density value at
    atomic center, to make water picking more reliable
  - Option to compute Average Kick Maps
  - Enhancements for neutron and joint XN refinement: initial implementation
    of map-based DOD placement and local fit, local real-space search to
    correctly orient rotationally ambiguous X-H/D positions
  - Automatic finding of best residue rotamers based on Rotamer library,
    combined with real-space refinement (in development)
  - Use better (faster, less memory expensive, with one bug fixed) code for
    mask calculation
  - Complete re-do of the code that determines bulk-solvent ksol&Bsol,
    anisotropic scale matrix and twin-fraction

- Wizards:
  - new GUIs for AutoSol, AutoBuild, and LigandFit
  - Simplified and general format for wizard parameters files
  - Any combination of MAD/SAD/MIR datasets in AutoSol supported through
    parameters files

- phenix.elbow:
  - restraints and geometry files can be output for all chiral combinations
  - faster generation of ligand topology (3-5x)
  - more robust graph matching
  - extensive testing of MOL, SDF input
  - save a 2D respresentation
  - load chemical components using key
  - send output directly to REEL

- phenix.reel:
  - R/S nomenclature for chiral centres displayed
  - switch between planar and tetrahedral geometry
  - ability to compare geometries with ideal restraints
  - load chemical components using key

- phenix.ready_set:
  - more robust for the edge cases
  - pass a ligand CIF file to use instead of eLBOW generated file
  - add exchangeable deuteriums (including HIS)
  - replace all hydrogens with deuteriums
  - allow better metal coordination via more options
  - add hydrogene to water
  - converts PDB LINKS to edits
  - can remove all waters

- phenix.ligand_identification:
  - command line version of old GUI ligand identification strategy
  - allow users to input biological information of macromolecule to generate
    custom-made libraries (e.g. function=kinase; or EC=1.1.1.1 …etc.)
  - allow custom library as list of 3-letter codes
    (e.g. ligand_list="ATP CTP APN FMN")
  - supports multi-thread computing

- Validation:
  - phenix.clashscore: all-atom contact analysis using Probe
  - phenix.real_space_correlation:
    - general tool to compute correlation between two maps
    - can work with ensembles of structures
  - phenix.validate_model (GUI only):
    - Shows all geometry outliers, with links to Coot/PyMOL
    - Includes Ramachandran, rotamer, C-beta, and bad contact analyses
  - phenix.validate (GUI only):
    - Extends phenix.validate_model with results from phenix.model_vs_data
    - Includes real-space correlation plots and outliers
  - phenix.polygon (GUI only):
    - Graphically ompare model statistics to similar structures in the PDB
    - reference: Acta Cryst. D65, 297-300 (2009)

- phenix.model_vs_data:
  - automatically detects and reports twinning (using Xtriage)
  - reports Molprobity-style set of numbers: Ramachandran statistics,
    clash-scores, CB-outliers
  - various bug fixes and enhancements.

- phenix.tls: tool for manipulating TLS records in PDB file headers

- phenix.grow_density:
  - a tool for local density improvement, as originally described in Acta
    Cryst. (1997). D53, 540-543 (in development)

- phenix.xtriage:
  - GUI rewritten with embedded graphs, tables, and text summaries
  - twin law selection integrated with phenix.refine GUI

- phenix.fit_loops: fast building of missing loops with RESOLVE

- phenix.find_helices_strands:
  - Rapid chain tracing (20 residues/sec) for evaluating map quality

- phenix.multi_crystal_average:
  - Multi-crystal averaging

- phenix.fobs_minus_fobs_map:
  - Utility for creating isomorphous difference maps (also in GUI)

Version 1.4 (December 2008)
===========================

- General:
  - Command-line validation programs: phenix.ramalyze, phenix.rotalyze, and
    phenix.cbetadev
  - PyMOL now included as general linux and Mac binaries

- phenix.refine:
  - Missing F(obs) filled in with F(calc) in default maps
  - Phase-combined maps created when experimental phases used
  - Selection of constrained occupancy groups
  - Graphical interface: phenix.refine_gui

- phenix.xtriage:
  - Graphical interface: phenix.xtriage_gui

Version 1.3 (August 2008)
=========================

- General:
  - Release version 2008
  - Changes to installer to check for csh/tcsh
  - changes in svn_checkout command
  - incorporated the PDB chemical components dictionary to aid with the
    interpretation and processing of ligands
  - infrastructure for OpenMP compilation added

- Wizards:
  - AutoSol:
    - Scoring now uses Bayesian estimation of correlation coefficient
    - Building changed to helices and stands by default (faster)
    - Building will continue with more thorough algorithm if it
      fails to build with the fast algorithm
    - When testing both hands in building the best model from the first
      hand will not be overwritten with the current model from the
      second hand unless it is better
  - AutoBuild:
    - Parallelized for running on SMP architectures using threads

- phenix.refine
  - complete rewrite of PDB handling - new model hierarchy, faster structure 
    reading, and direct structure writing. 
  - constrained group occupancy refinement applied to alternate conformations
    by default
  - algorithms implemented for hydrogen and deuterium picking in neutron/X-ray
    maps
  - numerous changes to better handle joint refinement of X-ray/neutron data
  - map coefficients written for an anomalous difference Fourier map if
    anomalous data is used

- phenix.xtriage
  - analysis of systematic absences

- phenix.elbow
  - extended to more element types
  - bug fix to deal with close atoms recognized as bonds
  - Numeric SVD code used for larger structures (addresses bug in tntbx)

- phenix.reel
  - more model manipulation features added

- Phaser:
  - minor bug fixes

- phenix.reduce
  - minor bug fixes

Version 1.3b
============

- General:
  - support for 2 letter chain IDs.
  - support for Hybrid36 atom and residue numbers.
  - Mac OSX port (PowerPC and Intel)
  - GUI modifications to support the Mac
  - Conversion to latest wxPython (2.8.4.0) and associated libraries
  - Installer modifications to support Mac-OSX port
  - Removed C++ library test on linux platforms
  - New tools: phenix.xmanip, phenix.pdbtools, phenix.reel, 
               phenix.find_helices_strands, 
  - New document generation system
  - Full documentation

- Wizards:
  - Simple command-line versions of all Wizards
  - AutoBuild:
    - Automated nucleic acid model building
    - Simple-omit, sa-omit, and iterative-build-omit maps
    - Automatic treatment of ligands when rebuilding models
    - Simple commands for building multiple models compatible with data
    - Simple command for rebuilding to produce one very good model
    - Prime-and-switch maps
    - Iterative model-building of strands and helices at low resolution
  - AutoMR
    - now using Phaser-2.1.2:
    - Default packing criteria relaxed in AUTO_MR mode
    - Composition can be estimated from solvent content
    - Default composition corresponds to 50% solvent
    - If AUTO_MR search finds some, but not all, components, the partial 
      solution is output
  - AutoSol
    - SAD phasing starting from MR partial model using Phaser
  - Ligand identification
    - Improved ligand scoring and ranking algorithm.
    - Implemented additional analysis of ranking results
    - Improved task display options
    - Task now accepts user-provided ligand library

- phenix.find_helices_strands: 
  - rapid fitting of secondary structure to an electron density map

- phenix.probe
  - better support for alternate names.
  - numerous bug fixes.

- phenix.reduce
  - updated het dictionary.
  - numerous bug fixes.

- phenix.refine:
  - Updated to handle the new atom names from the remediated
    PDB-V3. Currently protein and RNA/DNA only. For ligands with new names,
    use elbow.builder to generate CIF files with matching restraints.
  - Writing of a PDB deposition header
  - Constraints for occupancy refinement added (used in refinement of atoms 
    in alternative conformations);
  - Default behavior change: now phenix.refine always refines occupancies
    for atoms in alternative conformations or for atoms with partial 
    occupancies;
  - Water picking procedure improved: bug fix to remove waters located too 
    far from the molecule, H-bond criteria added;
  - Significant improvement in handling of hydrogens in refinement;
  - Bulk-solvent and scaling procedure made even more robust (B_cart 
    optimization is done a few times in-between grid search steps in 
    ksol and Bsol);
  - Documentationa reworked
  - Vdw_1_4_factor changed from 2/3 to 0.8 (to improve
    geometry and MolProbity clash scores in particular)
  - Obsolete remove_hydrogens option removed
  - Site symmetry enforced for ADP shaking (weight calculation) in 
    anisotropic ADP refinement (no more eventual crashes due to corresponding
    assertion error)
  - Complete set of tools for refinement at ultra-high resolution (higher 
    than 0.9A; modeling of bond electron density with interatomic scatterers,
    IAS)
  - New support for custom angle restraints
  - Determination of nonbonded distances distinguishes between refinement
    with and without hydrogens (based on remove_hydrogens=True/False);
  - Numerous enhancements for joint Xray+Neutron refinement (clearer handling
    of input data files, better targets weight calculation protocol)
  - MZ (multiple zone) rigid body refinement protocol is further optimized to 
    decrease the runtime, accommodate a provision for multiple refinable 
    rigid bodies.
  - A set of examples is added.
  - Bug fixes:
    - Proper handling of chain breaks, to avoid incorrect angle,
      dihedral, chirality and planarity restraints. This bug
      was likely to lead to bad restraints for structures with
      alternative conformations.
    - Avoid Wilson B crash for no data if low resolution omitted
    - Proper handling of negative residue numbers
    - Avoid "Segmentation fault" on Macs in twin refinement
    - Twin refinement (fix for wrong weight for X-ray/ADP term => refinement
      unstable)
    - Anisotropic ADP refinement for a model with hydrogens
      (riding refinement of H)
    - Fix inconsistency in group B refinement due to deep_copy method
      problem in twin_f_model.py

- phenix.pdbtools (new command-line tool for performing various manipulations
  on a model in a PDB file and general model statistics):
  - shaking of coordinates (random coordinate shifts)
  - rotation-translation shift of coordinates
  - shaking of occupancies
  - shaking of ADP
  - shifting of ADP (addition of a constant value)
  - scaling of ADP (multiplication by a constant value)
  - setting ADP to a given value
  - conversion to isotropic ADP
  - conversion to anisotropic ADP
  - removal of selected parts of a model
  - generates complete statistics for model geometry and Atomic Displacement
    Parameters.
  - tools for structure factors calculation from PDB model.
  - A documentation page added.

- phenix.cif_as_mtz (tool for conversion of reflection files in CIF format 
  into MTZ format). Main features include:
  - distinguishes between Iobs and Fobs;
  - distinguishes between Xray and Neutron data sets;
  - analysis and picks up valid R-free flags;
  - simple sanity check of data.
  - A documentation page added.

- phenix.elbow
  - Perform simple eLBOW jobs in COOT
  - Automatically perform the appropriate calculations on all the unknown
    ligands in a PDB file and combine the CIF results into one file
  - Covalent bonding of ligands to macromolecules (phase I implementation)
  - Print the sequence of PDB file (elbow.print_sequence)
  - User control of addition and writing of Hydrogens
  - Output of TRIPOS mol2 files
  - User control over naming of output files
  - elbow.join_cif_files has re-ordered arguments (target as last argument)
  - elbow.link_edits will generate "edits" from PDB LINK records for input
    to phenix.refine
  - elbow.metal_coordination will generate phenix.refine "edits"
    for metal coordination spheres including angles
  - improved chiral centre handling
  - access to all MSDChem SMILES strings
  - converted to PDBv3.0

- phenix.reel
  - alpha release of restraints editor
  - read, edit, optimise and write CIF files
  - run eLBOW from a GUI window

- phenix.xtriage
  - now handles unmerged data

Version 1.26.1b
===============

- Conversion to latest wxPython (2.8.1.1) and associated libraries
- Withdrawn support for SGI-Irix and Alpha-Tru64 (GUI libraries are 
  no longer supportable on these platforms)

Version 1.26b
=============

- General
  - Reduce (from Jane and Dave Richardson, Duke University) integrated into 
    PHENIX, available as phenix.reduce

- Wizards:
  - improved model building in AutoSol
  - improved model building in AutoBuild
  - Automated detection and application of NCS
  - Automated loop building, crossovers between chains in different models 
    of a structure, and side-chain optimization
  - SA-composite omit maps
  - Simple omit maps
  - Iterative-build composite omit maps
  - Omit around atoms in PDB file
  - Model rebuilding in place 

- Phaser:
  - Updated to version 2.1
  - Molecular replacement changes:
    - Clash test is more forgiving
    - Partial solution is output if full solution fails
  - SAD phasing includes partial structure option to start from molecular 
    replacement solution

- Textal:
  - Minor bug fixes

- Resolve ligand identification:
  - Improved ligand ranking algorithm 

- phenix.xtriage:
  - If pseudo translation peaks implicate a higher symmetry and/or smaller 
    unit cell, the appropriate unit cell and spacegroup are listed.
  - Optional detwining and treatment of data, including protocols for 
    anisotropy correction, outliers detection and twinning or detwinning 
    of data
  - xtriage code reorganized in such a way that it can be easily used in other
    scripts and the wizards  

- phenix.refine:
  - Bug fixes (major and minor)
  - Testing and bug fixes in TLS refinement
  - Final f_model exported as proper MTZ or CNS file (was ad-hoc before)
  - Group_anomalous refinement (for anomalous atoms such as Se)
  - Refinment of hemihedrally twinned data using a lsq target function
  - Output of 'detwinned' and gradient maps for twinned data
  - Support for refinement with hydrogens (not fully tested)
  - Simultaneous use of NCS restraints and automatic water picking
  - Individual anisotropic ADP restraints (similar to the SIMU restraint in 
    SHELX)
  - Flexible model parameterization (refinement can use rigid body, isoADP,
    anisoADP, etc simultaneously)
  - Simultaneous refinement of structures against X-ray and neutron data
  - Bug fix in processing of multiple apply_cif_link blocks applied to the 
    same residue
  - Changes to allow user specified covalent bonds: apply_cif_modification, 
    apply_cif_link
  - Fixes to binning to deal with sparsely distributed free-R set (e.g. thin 
    shells)
  - Individual occupancy refinement
  - Group occupancy refinement
  - Automatic NCS detection
  - Reorganization of phenix.refine parameters

- eLBOW
  - Perform simple eLBOW jobs in COOT
  - Automatically perform the appropriate calculations on all the unknown 
    ligands in a PDB file and combine the CIF results into one file
  - Covalent bonding of ligands to macromolecules (phase I implementation)
  - Print the sequence of PDB file
  - Added classes to monitor peptide linkages, nitrogen planarity and acid 
    groups to maintain geometry during coordinate calculations
  - Greatly improved molecule matching procedures
  - User control of addition and writing of Hydrogens
  - User control of bond cutoff for auto bonding
  - User control of number optimisation steps and convergence tolerance
  - Output of TRIPOS mol2 files
  - User control over naming of output files
  - User control of total memory usage 
  - More tests added
  - AM1 optimisation includes peptide linkage correction and more efficient 
    search procedures
  - elbow.join_cif_files has re-ordered arguments (target as last argument)

- phenix.explore_metric_symmetry:
  - Renamed from iotbx.ehms
  - Generation of point group graphs given a unit cell and space group 
    combination
  - Comparison of unit cells and their sublattices

- phenix.twin_map_utils:
  - Refinement of twin fractions and map parameters for all possible twin 
    laws given a fixed model and a given dataset

- mmtbx.remove_outliers
  - Outlier removal using Wilson or model based statistics

- PHENIX GUI:
  - Minor bug fixes
  
- Installer:
  - New solve/resolve CVS installation
  - Updates to libtbx package installation
  - Improved developer tools for updating PHENIX components (cvs_checkout,
    svn_checkout)
  - Backwards compatibility checks in phenix_env to detect an installation 
    from an earlier OS and use it if necessary

Version 1.25b
=============

- Minor bug fixes

Version 1.24b
=============

- Wizards:
  - LigandFit Wizard made more robust and tested against 9000 
    ligands in PDB
  - IterativeBuild and ModelBuild Wizards combined into single 
    AutoBuild Wizard
  - AutoBuild Wizard 
    - Introduced "rebuild_in_place" for rebuilding MR models
    - Full-omit map and multiple-model generation available
  - AutoMR Wizard
    - Automatic rebuilding of MR solutions using AutoBuild 
      rebuild_in_place
  - AutoSol Wizard decision-making improved using model completeness 
    as a quality measure
  - All Wizards can be fully run with scripts 
  - Sample scripts available for all Wizards with data in $PHENIX/examples

- Phaser:
  - Integrated with AutoMr Wizard
  - Sphericity restraint on overall anisotropy
  - Improved pruning of list of plausible solutions
  - Various bug fixes for unusual space groups and non-Linux architectures

- Textal:
  - created a single integrated command-line script: textal.build
    options: build backbone only (C-alpha chains)            [--capra_only]
             build complete model (backbone and side-chains) [default]
             build side-chains for user-supplied C-alpha's   [--input_model]
  - reads reflection files or XPLOR maps for building
    options for region of space to build:
      - runs FINDMOL to automatically create map centered over a contiguous,
        symmetry-unique molecular region using symmetry ops and clustering
      - build in ASU [--asu]
      - build in user-supplied XPLOR map [--input_map]
  - will do sequence alignment to improve amino acid identity if
    amino-acid sequence file is provided [--sequence]
  - user can provide coodinates of Selenium sites (if available) to potentially
    improve accuracy of sequence alignment by identifying methionine 
    residues [--se_sites]
  - now runs simulated annealing (phenix.refine) automatically as final
    step to estimate R-factor (as an indicator of model quality)
  - new unified Textal task/strategy in Phenix GUI
    - graphical front-end to textal.build script; same options, all on
      one dialog box

- eLBOW
  - Documentation available in installation and on-line
    - computer resources page
  - Uses bond length defaults from HF/6-31G(d,p) quantum calculations
  - Uses GDIIS to improve geometry optimisation convergence and DIIS in 
    quantum convergence
  - Restart mechanism begins from previous best geometry
  - Expert/Novice modes
  - Can use GAMESS for geometry optimisation
  - Provides estimates of the computer resources needed for a calculation
  - Provides suggestions upon completion of a run
  - eLBOW will extract a specified ligand from a PDB file containing both 
    protein and ligand
  - eLBOW will dynamically update geometry visualised in PyMOL
  - eLBOW can construct a short protein sequence
  - eLBOW can use a file as a template to name the ligand atoms
  - Database of SMILES and PDB files included
  - Input can be piped to eLBOW 

- iotbx.ehms:
  - Minor updates to functionality and output
  - Centring type is inputable rather than just a spacegroup 

- mmtbx.xtriage:
  - Maximum likelihood based estimation of twin fraction
  - Maximum likelihood based estimation of twin fraction 
    while taking into account possible NCS parallel to 
    the twin axis.
  - PHIL Interface rationalisation:
    - Detwinned data can be written out
    - various parameters can be set by the user if desired
  - Help/Manual updated (mmtbx.xtriage --h)
  - various bug fixes

- phenix.refine:
  - NCS restraints (coordinates and ADPs)
  - TLS refinement
    - Recalculation of TLS parameters from overall ADPs values
    - Parameter modification to remove positive definite atomic aniso-Us
    - TLS parameters written in output PDB file
    - Options:
      - TLS alone
      - TLS plus ADPs
      - TLS plus ADPs and coordinates
  - Group B-factor refinement
  - Rigid body refinement
  - Simple structure factor calculations (with or without bulk solvent 
    and scaling)
  - Output MTZ coefficients for use in COOT
  - Increased checking of input PDB files and reflection data
  - Updated documentation

- PHENIX GUI:
  - color indicators added to show when a strategy/wizard icon has been
    double-clicked
  - strategies and tasks rationalized
  - improved default map and coordinate viewing options when displaying
    results in PyMOL
  
- Installer:
  - updated to 0.99rc6 version of PyMol
  - updated to more recent versions of freetype, fontconfig and pango on linux.
    This fixes a memory allocation problem on Fedora Core 5.
  - added --alias_mtype option to installer to force installation of
    a different machine type (binaries only). May be necessary on some
    SuSE systems. 
  - now using freeglut and latest 8.4.x versions of tcl/tk
  - solve/resolve source code now included in the distribution
  - examples directory restructured and more examples/tests added

Version 1.23a
=============

- New tool: iotbx.ehms
  Explore Higher Metric Symmetry. Allows one to easely compute/list
  all possible point and spacegroups allowed by a certain point or
  spacegroup. A graph is constructed and is plotted if graphiv is
  installed.

- mmtbx.xtriage updates:
  - Data strength analysis
  - Low resolution completeness reporting
  - Wilson plot sanity analyses
  - Outlier detection based on Wilson statistics
  - Ice ring detection
  - Anomalous difference analyses
  - More elaborate reporting on type of twin laws
  - If calculated data is supplied next to experimental data,
    R.vs.R statistics is reported (Lebedev et al, Acta Cryst.
   (2006). D62, 83-95)
  - Missing rotational symmetry analyses. If symmetry is too low,
    new spacegroup is suggested on basis of R-value.
  - various bug fixes

- New tool: mmtbx.fest
  - Simple delta F and Fa construction from 2-lambda MAD, SAD SIR or RIP
    data.
  - Various scalings protocol can be chosen. Most (if not all)
    critical parameters can be set/modified by the user.

- New tool: eLBOW

Version 1.22a
=============

- Updated version of SOLVE/RESOLVE (2.11)
- Generation of environment file for sh/bash

Version 1.21a
=============

- Updated wizard infrastructure:
  - Wizards can call other wizards
  - Bug fixes and changes to AutoSol and AutoMR
  - New loop-fitting algorithm implemented in RESOLVE
  - New Iterative, TextalBuild and ResolveBuild Wizards
- New automatic ligand geometry and restraint building (ELBOW):
  - semi-empirical method for determination of coordinates
  - generation of coordinates from SMILES
  - generation of monomer library (cif) files for refinement 
- Updated phenix.refine:
  - improved performance (1.5 to 2 times faster)
  - automated water picking
  - local-neighborhood individual ADP restraints
  - improved bulk-solvent/anisotropic scaling calculations
  - new automated weight generation scheme
  - refinement against neutron diffraction data
- New data analysis tool: mmtbx.xtriage
  - automated detection of anisotropy, twinning, and 
    non-crystallographic translational symmetry
  - maximum likelihood determination of anisotropic and absolute
    scaling parameters
  - automatic determination of possible twin laws
- Infrastructure:
  - latest versions of thrid-party GUI libraries (glib, gtk+, 
    wxWidgets and wxPython)
  - removed BLT from installation
  - linux installation modified to account for new x86_64 architectures 
    and new linux 2.6 kernel
  - SGI installer

Version 1.2a
============

- Updated wizards

Version 1.1a
============

- Release version based on 1.08a
- Minor bug fixes for TEXTAL, SOLVE/RESOLVE strategies
- Minor bug fixes for PHENIX GUI
- Additional documentation - HySS and phenix.refine

Version 1.08a
=============

- Structure refinement (phenix.refine) including:
  - LBFGS minimization
  - Cartesian dynamics simulated annealing
  - Maximum likelihood with amplitudes
  - Maximum likelihood with amplitudes and phases
  - Restrained individual B-factor refinement
  - CCP4 monomer library for geometry restraints generation
  - Automatic input PDB interpretation
  - Automatic input data interpretation
- Latest version of PHASER (1.3)
  - Bug fixes from Phaser-1.2
  - Improved automated MR
  - Final phasing and refinement to full resolution
  - Improved packing
    - RNA/DNA
    - close-packed oligomers
    - only most homologous model in ensemble used for analysis
  - Reduced memory requirements
- Latest version of TEXTAL
  - Builds models from structure factors
    - users no longer required to prepare a centered density map
    - will automatically adjust resolution as necessary, and can
      now build models for datasets at 3.0A resolution or higher
  - Automated detection, centering, and tracing of protein molecules
    using the FINDMOL algorithm
  - Simplex optimized model building for better CA positions and side-chain
    identities
  - Task re-organization into hierarchical groups:
    - New high-level tasks to simplify model building
    - utilities and primitive tasks (for experts)
  - Invocation of many Textal routines from command line
- Wizards
  - Updated AutoSol and AutoBuild wizards
  - New AutoMR wizard for automated molecular replacement
  - New AutoLig wizard for automated ligand fitting
- Latest solve/resolve version - 2.09
- Updated solve/resolve strategies
- Infrastructure
  - PHENIX installation co-exists with other cctbx-based packages
  - PHENIX co-exists with Linux GNOME environment
  - Now using Python 2.4.1
  - Now using version 2.0+ of GTK+
  - Now using latest versions of wxGTK (wxWidgets) and wxPython
  - Now using PyMOL 0.97

Version 1.07a
=============

- Added Wizard infrastructure to PHENIX
- Wizards for automatic structure solution and model building
- Improved use of UNIX domain sockets
- Fixes in PHASER code for SGI and Alpha platforms
- Updated PHASER strategies
- Updated TEXTAL strategies
- New TEXTAL stitch algorithm for gap closure
- Updated SOLVE/RESOLVE strategies
- New strategies added for running phase improvement with RESOLVE
- Updates to the PHENIX GUI infrastructure
- Splash-page documentation added
- SOLVE/RESOLVE incorporated in the PHENIX distribution
- Test for appropriate version of make for source installation
- Fixes to glib configure script for Alpha
- Fixes to wxGTK configure and source code for Alpha
- Installer compiles Python with C++ compiler if required 
  for wxPython compilation
- RedHat 7.3 removed from supported platforms
- Additional command line options for HySS
- Improved the killing of strategies
- Added ability to run strategies in queuing systems
- Sun Grid Engine currently supported for queuing (this may work
  with other queuing systems, but is not tested)

Version 1.06a
=============

- Minor bug fixes in GUI
- Added option to use rsh instead of ssh to run remote jobs
- New detection of network status on startup
- Use UNIX domain sockets if network is not functioning
- Updated tasks and strategies for PHASER
- Improved map reading speed in TEXTAL tasks

Version 1.05a
=============

- New tasks and strategies for PHASER
- New tasks and strategies for TEXTAL
- New map database for TEXTAL
- Bitmaps for task and strategy canvas in PHENIX GUI

Version 1.04a
=============

- New tasks and strategies for SOLVE/RESOLVE
- Improvements to hostname resolution to deal with incorrectly 
  configured systems

Version 1.03a
=============

- New installation scripts
- End-user installation from complete binary bundle
- Added SGI Irix support (remote job execution with rsh only)
- Added Redhat 9.0 support - to deal with NPL problems
- Generic Linux installation for platforms other than Redhat 7.3
- Bug fixes in GUI
- Improved support for remote job execution
- Moved to hostname only resolution, no use of explicit IPs
- Improvements to HySS performance

Version 1.02a
=============

- Fixed bug with making connections between tasks in GUI
- Fixed file ownership problems for binary bundles in installation
- Fixed problem with absolute links in binary bundle (all linked files removed)
- Changed install script back to using csh rather than tcsh

Version 1.01a
=============

- Numerous bug fixes in GUI
- Added text to better document PHASER and TEXTAL tasks
- Fixed file permission problems for TEXTAL PERL scripts
- Improvements to HySS procedure

Version 1.0a
============

- First alpha test release