Phaser-2.7 series changelog

From Phaserwiki
  • add feature to filter out data with low information content
  • bugfixes
solution history formatting
R-value test with multiple search components failing
limit number of permutations in amalgamation to stop combinatorial explosion
correct handling of effect of experimental error in fast rotation function
search method full & permutations on had corrupted search ensemble copy numbers
  • add option to allow input SAD atoms to change atom type during refinement
  • bugfixes
packing checked performed on input solution in auto mode
packing checked performed on final solution after refinement in auto mode
only top fast translation function peaks checked for packing, if not a helix
  • solution coodinates placed nearest to input coordinates if possible
  • logfile width extended
  • default number of processors for parallelization increased to 4
  • support for unicode
  • GIMBLE refinement, which splits ensembles by chain for rigid body refinement
  • automatic tNCS NMOL determination in presence of commensurate modulation
  • bugfixes
input solution list checked for pass of intrinic TF packing test
ensembles can refine to zero occupany in pruning
  • additional peak clustering selection/annotation functionality PEAK ... LEVEL
  • search deep only implemented for auto fast mode, not for FRF mode
  • packing test during FTF run by default with 50% pairwise packing cutoff
  • bugfixes
stored solution list does not carry phasing arrays excessive memory
occupancy refinement takes account of VRMS
  • option to run packing test during FTF so as to eliminate non-packing translations from top
  • packing test during FTF run by default for helices
  • (selected) obsoleted keywords don't throw error reported as warning in logfile
  • bugfixes
peak search percent for translation function
  • replace unweighted 2nd moments by inverse variance-weighted 2nd moments for twinning tests
  • pairwise packing by percent only, other packing modes obsoleted
  • PAK reported in annotiation is maximum percent overlap amongst solution set
  • bugfixes
use of input trace molecule
  • Speed up packing test to O(N)
  • bugfixes
correctly increase scattering with implausible composition
  • write tNCS and anisotropy parameters to binary files (on user request)
  • read TNCS/anisotropy binary files to avoid refinement, when running through scripts
  • account for measurement error in eLLG calculation
  • rigid body refinement can separate chains in ensembles with MACMR CHAINS ON
  • change Vm probabilities to Weichenberger & Rupp 2013 from Kantardjieff & Rupp 2003
  • CCA accounts for PROTEIN/NUCLEIC/MIXED compositions different Vm probabilities
  • packing trace molecule can be entered independently of coordinates and map
  • packing volume for maps calculated using Wang mask algorithm
  • packing test by PAIRWISE only considers pairwise overlaps, not global overlap
  • packing sampling can be forced to ALL, CALPHA, HEXGRID or AUTO (default AUTO)
  • obsoleted packing test by BEST
  • automatic search number for single search ensemble
  • bugfixes
RMS from ID estimator function parameters (A,B,C) did not match published parameters
negative variance when trying to amalgamate many solutions with negative B-factors
chain overlap incorrectly detected for nucleic acid
  • improve tests for RMS correlations between members of an ensemble
  • change from average scattering residues to polyalanine for target eLLG calculation
  • determine number of average scattering residues required to achieve target eLLG
  • bugfixes
header for Rotation Gyration Refinement mode corrected
  • selection by CHAIN and MODEL for PDB coordinate entry
  • concatenation of PDB files on ENSEMBLE keyword input obsoleted
  • occupancy refinement operates on single model, not ensembles
  • change to format of SOLUTION HISTORY
  • change to SOLUTION variance card, only parse/unparse DRMS
  • bugfixes
only use solutions that pack to determine top LLG for purge after refinement
occupancy refinement using ensembles failing various asserts in some circumstances
exclude solutions that fail to pack from duplicate test after refinement
incorrect selection of best model when CARD ON input used sequence identity
reporting negative contrast (Cntrst) for LLG near purge mean
bugs in FAST mode solution reversion with pruning
  • No functional changes from 2.6 series
  • new internal data structures for ensembles
ensemble parameters either input or worked out on input
Indexing of values in ensemble/model arrays consistent
  • introduce command line flag --changelog gives changelog and bugfixes from previous version
formal tracking of changes
  • bugfixes
Wilson B-factor not being carried internally or used correctly for output pdb files
ID being used as RMS is eLLG calculation in MODE MR_FRF