Skip to content

Releases: keitaroyam/servalcat

0.4.99

11 Dec 07:45
Compare
Choose a tag to compare
  • refinement and map calculation related changes
    • "No Refmac" refinement in general
      • vdw exclusion condition revisited
      • mixed derivatives between occupancy and ADPs
      • use obs data resolution
      • set up nonbonded between angle-related atoms when angle is not defined
      • angle restraints: special case for ideal angle of 180 degrees
      • made restraint related keywords common
      • write stats to mmcif
      • print hydrogen related options
    • metals.json
      • cap sigma at 0.02, and factor of 1.5 was needed for mad
    • refmacat
      • now entity_poly is written
  • new command: util conf
    • to compare torsion angles between chains or files
  • util seq
    • scoring method optimised (hopefully)
  • fsc
    • fix bug when using --model and --map together
  • fofc
    • hide --omit_proton/--omit_h_electron options, which are experimental and not ready to use
  • pdb/mmcif writing
    • mmcif: _atom_site.auth_comp_id was requested
    • pdb: shorten long ccd codes
  • switched to nanobind
  • requires gemmi 0.7.0

Full Changelog: v0.4.88...v0.4.99

0.4.88

14 Sep 01:50
Compare
Choose a tag to compare
  • refmacat: --keep_entities copies entities from input to output
  • geometry related changes
    • show sigmas in outliers
    • support symmetry in angles (only through external restraints)
    • ignore hydrogen chem_type from dictionary (not to fail with unknown hydrogen atom types)
    • bug fix: external restraints should be read after loading topology
    • NCS restraint related bug fixes
  • included in CCP4 9.0.004
  • requires gemmi 0.6.7

Full Changelog: v0.4.82...v0.4.88

0.4.82

19 Aug 22:03
Compare
Choose a tag to compare
  • refmacat: bug fix for unrestrained refinement
  • refine: bug fix for adp restraint with ncs restraint
  • refine_xtal_norefmac
    • experimental implementation of twin refinement
    • in intensity refinement, R1 is calculated instead of "R2", using I/sigma>2
    • fixed broken reading of shelx hkl file
  • included in CCP4 9.0.003

Full Changelog: v0.4.77...v0.4.82

0.4.77

25 Jul 01:38
Compare
Choose a tag to compare
  • refmacat
    • quick workaround for "given" NCS matrices
    • fix label_seq_id in output mmcif
    • fix erroneous handling of unknown external restraint keywords
    • bug fix for unrestrained refinement
    • write servalcat version to mmcif
  • NoRefmac refinement
    • add --newligand_continue option
  • refine_xtal_norefmac
    • don't allow negative or more-than-unity occupancies
  • fsc
    • made model masking optional (--mask_model)
  • included in CCP4 9.0.002

Full Changelog: v0.4.74...v0.4.77

0.4.74

20 Jul 08:50
Compare
Choose a tag to compare
  • refmac_keywords: fixed infinite loop occurring when "symm" was specified in an atom spec
  • refine/refmacat: add metal links only when requested
  • fw: support anomalous intensity input
  • included in CCP4 9.0.001

Full Changelog: v0.4.72...v0.4.74

0.4.72

12 Jun 02:20
Compare
Choose a tag to compare
  • refinement and map calculation related changes
    • no longer discriminates between ATOM/HETATM records
    • refmacat
      • cap occupancy at 1 for refmac compatibility
      • fix elements in the model using monlib
      • prevent duplicate block names in output mmcif
      • refmac compatibility fix for strict ncs
      • bug fix for missing chain IDs
    • geometry
      • exclude plane atoms within 1-4 distance from vdw
    • refine_spa_norefmac
      • reduced data stats printing
      • new option: --target_bond_rmsz_range
    • refine_spa / refine_spa_norefmac
      • no longer include F(000)
      • lower correlation cutoff for helical symmetry test
      • soft mask creation made much faster
    • "No Refmac" refinement in general
      • now understands refmac keywords "restr tors incl" and excl
      • don't write hydrogen to the output model by default
      • refmac way of weight adjustment
      • write stats json every cycle (for ccp4i2)
      • (experimental) occupancy-only refinement with restraints
      • fixed wrong second derivative calculation wrt occupancy
      • add outliers to stats.json
      • always normalise space group name in the structure
    • refmac keywords:
      • "end" to stop in each file
      • fixed reading "exte weig scal" lines
  • new command: util add_op3
  • new command: util dnarna
  • util geom: add --nucleus option
  • requires gemmi 0.6.6
  • included in CCP4 9.0.000

0.4.60

11 Mar 03:08
Compare
Choose a tag to compare
  • now python -mservalcat works instead of -mservalcat.command_line
  • refinement and map calculation related changes
    • group occupancy refinement implemented
    • update old atom names in the model (if available as _chem_comp_atom.alt_atom_id)
    • check duplicated atoms
    • show polymer gaps
    • geometry
      • local NCS restraints implemented
      • nonbonded interaction now doesn't rely on altlocs, but checks sum of occupancies and occupancy group (Refmac way)
      • use all hydrogen torsion angles
      • remove duplicated links
      • fix atom order in bonds: duplication wasn't resolved properly
      • fix link wrongly considered Asu::Different if symmetry code not given in the model file
      • limit ideal angle to [0,180]; e.g. an ideal value of -180 caused a problem
      • fix wrong reporting of torsion angle values
      • ignore bad (NaN) ideal chirals
    • refine_geom --update_dictionary
      • support pdbx_model_Cartn_x_ideal
      • write json file (requested by acedrg)
      • made more robust. there was a risk for H position to be NaN when the geometry was very poor
    • refmacat
      • ignore entities in input
      • refmac workaround for long residue names
      • fix for 3 letter chain IDs
      • fix for Refmac bug (DNA became non-polymer entity after Refmac)
      • don't prepare crd if unrestrained refinement requested
      • fix MAKE keyword reading
      • write H/D expanded mmcif file as well
      • fix issues arisen when used from acedrg: empty chains, wrong date format, space in _entry.id
    • crystallographic refinement
      • refine_xtal_norefmac & refine_cx: fail if model and data symmetry mismatch
      • allow anomalous data input and calculate anomalous difference map
      • remove FP<=0
  • util seq: check unknown seq (wrong model)
  • util geom: --selection added

requires gemmi 0.6.5.

0.4.39

02 Nov 13:42
Compare
Choose a tag to compare
  • refinement and map calculation related changes
    • show bin stats in cycle 0
    • fix ADP constraints
    • harmonic restraints implemented
    • metals: user's keywords have priority
    • refine_xtal_norefmac:
      • intensity: added an FOM proxy
      • solvent mask: use Refmac radii and ignore H/zero-occupancy atoms
      • smoothen ML parameters (D and Sigma)
      • amplitude target and derivatives in C++
    • refmacat
      • add --version option
      • fix buffered output in refmacat
  • Refmac pipeline
    • print raw output when failed to check the version #13
  • symmetry: bug fix for negative helical rise
  • util geom: support exte keywords
  • requires gemmi 0.6.3

0.4.32

09 Aug 07:15
Compare
Choose a tag to compare
  • refinement and map calculation related changes
    • bug fix: it failed when the link header refers to non existent atoms (including 'gap' links)
    • don't add new links automatically. --no_link_check was obsoleted, and --find_links was added.
    • "no refmac" refinement in general:
      • ADP restraints: don't put distance dependence for bond and angle related pairs.
      • report ADP rms.
      • check -LL difference after ml parameter optimization
      • bug fix in bond angle rmsd reporting
      • bug fix in second derivative smoothing
    • refine_xtal_norefmac:
      • least square scaling significantly improved
      • ADP restraints now uses KL divergence as Refmac does. more controllable parameters
      • automatic weight decision
      • write *_stats.json
      • change --use_all_in_est to --use_work_in_est
      • amplitude target: fixed an anisotropy bug introduced in the previous version
      • intensity: fix wrong derivatives
      • intensity: use all reflection to ML refine B_aniso
      • binning-independent second derivative calculation
      • use previous result as initial for ML scaling and least square scaling
      • guess free flag number automatically
      • ignore nan in CCavg calculation
      • stop extending to lowest resolution
      • bug fix: --ligand didn't work
    • refine_spa_norefmac:
      • fix target function (only affects printed -LL)
      • add --keep_entities
      • bug fix: expand ncs keeping links in Fc calculation
  • fsc
    • model based mask and the sharpen-mask-unsharpen procedure are back
  • trim
    • fix for symmetry code in the link headers. was problematic when symmetry-related links exist
  • fix for numpy/pandas rc* version number
  • new command: util adp
    • show ADP stats and histogram
  • new command: util seq
    • print sequences of the model and align using input sequence to detect mutations.
  • tests fixed for Windows
  • included in CCP4 8.0.016

0.4.24

31 May 20:42
Compare
Choose a tag to compare
  • refinement and map calculation related changes
    • fixed accidental TER problem (TER in between polymer residues caused breakage of polymer links)
    • use of metals.json in restraints MonomerLibrary/monomers#28
    • write SSBOND header in pdb (_struct_conn.conn_type_id disulf in mmcif)
    • support chemcomp block in input model
    • refine_xtal_norefmac:
      • ML parameter refinement stabilised
      • do not load monomer library when --unrestrained
      • report CCavg, the averaged correlation over resolution bins weighted by the number of reflections
      • now --source is required
      • add --no_link_check
      • bug fix and many improvements in intensity target
      • bug fix in space group decision
      • fixed negative diagonal bug in log likelihood part
      • bug fix in same_asu decision in link finding (crystallographic translation was overlooked)
      • add neutron support (but no hydrogen refine)
      • do not overwrite model space group unless needed (as COOT does not seem to support "R -3 m:H" style)
    • refine_spa_norefmac:
      • support symmetry-related links
    • "no refmac" refinement in general:
      • print data stats by resolution
      • add max_vdw_radius to cap van der waals radius as a quick fix for K+
    • per-atom geometry violation score
    • SPA refinement: Fupdate is now calculated as <F> + F_diff
    • refine_geom: --update_dictionary re-adds hydrogen before 2nd round
  • refmacat
    • fix vdwr/vand/nonb keyword parsing
    • fix links for DOD when changing to HOH
    • write crd file in output dir
    • delete crd file unless --keep_original_output is given
  • fw: find intensity columns automatically and keep test set
  • new command: util map2mtz
  • CMakeLists.txt added, thanks @wojdyr
  • MANIFEST.in added (previous package missed some files for compilation)
  • requires gemmi 0.6.2
  • included in CCP4 8.0.012