Release Notes¶

Platform Support¶

• Support has been added for Ubuntu 16, OSX 10.10, OSX 10.11, and OSX 10.12.
• OSX 10.7, OSX 10.8, OSX 10.9, SuSe 11, and Redhat 5 x86 are no longer supported.

Coot Support¶

• Support has been added for Coot 0.8.
• Coot versions 0.6 and earlier are no longer supported.

Bug Fixes¶

• A bug in WRITEDICT that caused inappropriate bond lengths to be written for some cyclic systems has been fixed. Originally, WRITEDICT outputted the MMFF bond potential minimum as the bond length. However, for aromatic rings this value is inappropriately large since MMFFs angle and torsion potentials also affect the bond length significantly. In these cases, the equilibrium bond lengths are used by WRITEDICT.
• A bug that caused PDB remark data that was more than 72 characters long to be clipped from the refined output has been fixed. Practically, this only occurred when the input and output file formats were both PDB.
• On Windows, the default “save” location has been changed from C:Program Files (x86)AFITT 2.4.0 to the Users Home directory.
• Previously, if an older version of AFITT was installed on Windows, users could not overwrite the old preferences. This issue has been fixed.
• A crash that occurred on Windows 10 after selecting the Refmac option without a locally installed copy of Refmac has been fixed.
• The AFITT suite ships with a simple Python-based do_gold.py test script, to make sure AFITT-CL (FLYNN) is functioning as intended. This script analyzes the Gold dataset (protein+ligand complexes) available from the Electron Density Server (EDS) and included with the distribution. The do_gold.py test script has been modified to run correctly with either Python 2 or Python 3.

Platform Support¶

• Added support for RedHat 7, Ubuntu 14 and OSX 10.9
• Dropped support for RedHat 5 32 bit, Ubuntu 10 and OSX 10.6
• This will be the last release for SuSe 11

FLYNN¶

• Alter the protein masking to not remove density that may be created by hydrogen or covalent bonds. This was causing density very close to the protein to be truncated and hard to find automatically.
• -box is now called -manualSearch (-box still works)
• -blobsThenBox is now called -autoThenManualSearch for clarity (-blobsThenBox still works)
• -ciftype now follows the new writedict convention (accepts both numeric and more readable names such as “refmac”, “xplor”)
• -ciftype parameter is now in the correct parameter section (Advanced not Fragment)
• FLYNN always fixes serial numbers when writing CIF output to better enable use in workflow situations. These can be caused when hydrogens are added to ligands outside the context of the protein for instance.

WRITEDICT¶

• AFITT now Post-processes equilibrium angles to see if they are compatible with geometric forcefields. If they are not, replace them with fully MMFF optimized angles. This mostly affects 6-5 aromatic systems, but is a general fix ensuring the geometric force-field will work properly given the constraints in the CIF dictionary.
• Writedict now properly fail if the input protein has duplicate serial numbers. Most refinement programs cannot handle this state.
• Prevent duplication of CRYST1 records in the writedict output.
• Always output disulfide bridges if input
• Never rename standard residues hydrogens, this was causing Markly naming violations.
• Fix phenix covalent link output when covalently linking two residues of the same residue type.
• WRITEDICT no longers fails if input standard residues have alternate conformations with different graphs.
• WRITEDICT now defaults to MMFF94s
• WRITEDICT now understands human readable names for the dictionary type, i.e. -type refmac as opposed to -type 0
• WRITEDICT now has a -fix_serial_numbers flag that corrects duplicate serial numbers in the resulting pdb output

ROTFIT¶

• Added rotfit rotamer driving command line program. This used to live entirely in the GUI.

Other¶

Note: Last release for SUSE Linux Enterprise 10.

WRITEDICT¶

• Hydrogen naming is now consistent with RSCB standards.
• added -prefix flag for controlling report and logging generation

FLYNN¶

• Optimization fixes for molecules with high strain.
• Better default ligand density location.
• Optional PM3 and AM1 ligand minimization (see the -ff flag).
• Fixed symmetry operations on cubic lattices.
• added -prefix flag for controlling report and logging generation.

COOT Integration¶

• Better error message when the ligand and protein are set to the same coordinates
• Added ability to choose both chain id and residue name for the fit ligand.

WRITEDICT¶

• Updated to support REFMAC 5.6 and 5.7 (CCP4 6.3.0)
• Monomers updated to current RCSB and CCP4 releases.
• PDB LINK records are validated to be in the correct order. This validation is required for REFMAC 5.6 and up.

FLYNN¶

• Bugfixes and general speedups applied to internal optimizer.
• A memory leak in the optimzer was fixed.

WRITEDICT¶

• Added PHENIX suppoert
• Monomers in CIF files are given IUPAC names when possible
• Updated dictionaries to support REFMAC5 in CCP4 6.2.0

FLYNN¶

• Added PHENIX suppoert
• Fixed issues where ligands fit in poor density could be overstrained.

WRITEDICT¶

• Updated internal REFMAC and RCSB to current versions.
• When remapping dictionaries, WRITEDICT now indicates where the matching monomer was sourced.
• Moved “lookup” out of the -strict settings to a new flag “-lookup” adding this switch will search for a matching monomer from the internal libraries using the current -strict settings.
• -force options added. WRITEDICT will output the residue name even if it doesn’t match the internal list of known monomers.
• Various corrections to aniline nitrogen dictionaries.

FLYNN¶

• FLYNN now sorts the output fits based on real space correlation coefficient if applicable (RSCC)
• FLYNN now can automatically fit fragment cocktails. (use -fragment flag)
• FLYNN adds additional scores to all ligands besides the default MMFF Tanimoto/Shape, these are RSCC (Real space correlation coefficient), PLP (Piecewise-linear potential, requires protein) and Chemscore (requires protein).
• -resname now forces the ligand to use the given residue name even if it conflicts with the internal monomer library.
• -chainid Output the ligands with the given chain id.

WRITEDICT¶

• Added correct dictionary creation including connection table. This is required for newer versions of coot
• CIF dictionaries properly encode all LINK records.
• CIF dictionaries now encode proper planarity for guanine and no longer have duplicate planar atoms when using MMFF94s.
• CIF Dictionaries now look up known non-peptide residues. If the residue is not consistent with the known dictionary, the residue name is changed and a new MMFF94 derived dictionary is generated.
• Updated COOT integration to 0.6+

FLYNN¶

• Enhanced automatic density location. Blobs are now slightly larger than before making better targets for fitting.
• Added experimental fragment fitting. See the documentation for more information about this command line option. For version 2.0.1 this requires the fragment feature of the FLYNN license. This requirement will be removed when this feature is considered non-experimental. If you are interested in helping us test this feature, please contact business@eyesopen.com for more details.
• Refinement dictionaries are always written out for the ligand. Please note that these are written as if the -split flag is set to true. Refer to the documentation or type “flynn –help” at the command line for more details.
• Updated COOT integration to 0.6+