Detailed chronological listing of some changes since MODELLER-4 release on 17 June, 1997

For an incomplete list of the changes from 17 March 1994 to the release of MODELLER-4, see file 'doc/modeller4-changes'.

• 6/23/97, AŠ: Added ALIGNMENT_FORMAT = 'FASTA' format to the WRITE_ALIGNMENT command (routines rdaln and rdpir).

• 6/24/97, AŠ: Corrected the bug in OPERATE which sometimes caused errors in using the result variable (e.g., with WRITE OBJECTS = RESULT) that has been defined by the user and not assigned a scalar value before the call to OPERATE (routine act13).

• 6/25/97, AŠ: Incorporated several minor code changes suggested by Andrzej Sawaryn to allow compilation with f90 on Ultra 2 running Solaris 2.5.1 (routine ran6, file 'Makefile').

• 6/27/97, AŠ: Changed BUILD BUILD_METHOD = '3D_INTERPOLATION' so that the new ``line'' of atoms is a little randomized to prevent numerical problems with energy calculations (linear impropers, angles, etc.) (routine insxyz).

• 6/27/97, AŠ: Upgraded TRANSFER_XYZ by adding CLUSTER_METHOD, which selects how are the template-template comparisons done at each residue position ('RMSD' or 'MAXIMAL_DISTANCE') (routine trfxyz).

• 7/11/97, AŠ: Changed the format of the representation of spline restraints in the .rsr files (MODELLER12 format changes to MODELLER5 format). The specs now contain only the $y$ values, the derivatives are calculated on the fly. The same for the non-bonded MODELLER restraints in the parameter file.

• 7/13/97, AŠ: Added arguments SIGNIF_CUTOFF and SEARCH_GROUP_LIST to the SEQUENCE_SEARCH command. All members of the groups that have SIGNIF larger than SIGNIF_CUTOFF[1] and are not more than SIGNIF_CUTOFF[2] units worse than the best hit are put into the alignment arrays.

• 7/13/97, AŠ: Added structure resolution to the code in the output pairwise distance matrix produced by the ID_TABLE and SEQUENCE_COMPARISON commands.

• 7/19/97, AŠ: Added a check for sufficient overlap in MALIGN3D; not a stop, but a return in that case now.

• 7/22/97, AŠ: Added the READ_RESTYP_LIB command for reading new residue type definitions, and the corresponding RESTYP_LIB_FILE variable for specifying the library name.

• 7/22/97, AŠ: Simplification of the 'mod5' script.

• 7/22/97, AŠ: Added automatic recognition of the R, E, and P entries to the ADD_RESTRAINT command (it was impossible to add the excluded pairs and pseudo atom definitions before this modification).

• 7/30/97, AŠ: Added keyword SWAP_ATOMS_IN_RES to SUPERPOSE. This allows swapping atoms in residues `DEFHLNQRVY' (as selected) to minimize sidechain RMS deviation between MODEL and MODEL2 that have exactly the same atoms in the same order (match.F).

• 8/8/97, AŠ: Changed the __model.top script so that the FINAL_MALIGN3D option works for alignments that use the . option for segment specification.

• 8/9/97, AŠ: Renamed the KEEP_RESTRAINTS string variable to the RESTRAINT_SEL_ATOMS integer variable and generalized its meaning (see below). Old KEEP_RESTRAINTS of ONE_ATOM (default) is equal to the new RESTRAINT_SEL_ATOMS of 1; and the old KEEP_RESTRAINTS of ALL_ATOMS is equal to the new RESTRAINT_SEL_ATOMS of 9999. Impacts MAKE_RESTRAINTS and PICK_RESTRAINTS.

• 8/9/97, AŠ: Renamed the DYNAMIC_FLAG variable to NONBONDED_SEL_ATOMS and modified its meaning in order to make it consistent with the new variable RESTRAINT_SEL_ATOMS. Old DYNAMIC_FLAG of 0 (default) is equal to the new NONBONDED_SEL_ATOMS of 2; and the old DYNAMIC_FLAG of 1 is equal to the new NONBONDED_SEL_ATOMS of 1. Impacts MAKE_RESTRAINTS and OPTIMIZE.

• 8/9/97, AŠ: Added RESTRAINT_SEL_ATOMS and modified the MAKE_RESTRAINTS command so that the stereochemical, MRFP stereochemical, dihedral angle, and bi-dihedral angle restraints are calculated only when a sufficiently large number of atoms in a possible restraint are selected (set 1): For a given possible restraint, only when all restraint atoms or at least RESTRAINT_SEL_ATOMS are selected (set 1), is the restraint calculated. An exactly equivalent role for the `static' non-bonded restraints (and also for the `dynamic' non-bonded restraints calculated during OPTIMIZE) is played by NONBONDED_SEL_ATOMS (the new TOP variable that replaces and modifies the meaning of the old, deleted TOP variable DYNAMIC_FLAG). This now makes it possible to restrict the calculation of all new restraints to the selected atoms only (atom-atom distance restraints already depended on selected atom sets 2 and 3; secondary structure restraints are calculated for explicitly defined chain segments only).

• 8/9/97, AŠ: Modified the MAKE_RESTRAINTS command so that the calculation of the `static' non-bonded restraints now takes into account RESIDUE_SPAN_RANGE and RESIDUE_SPAN_SIGN.

• 8/9/97, AŠ: Corrected a bug in MAKE_RESTRAINTS ('mkdyn.F') that resulted in an incorrect use of DYNAMIC_FLAG during calculation of the `static' non-bonded restraints.

• 8/9/97, AŠ: Optimized speed of MAKE_RESTRAINTS for non-bonded restraints by considering the selected atoms (set 1) instead of all atoms.

• 8/9/97, AŠ: Optimized speed of PICK_RESTRAINTS by using the 'picatm' array instead of the 'cnsin1' and 'cnsin2' functions.

• 8/9/97, AŠ: Corrected a bug in 'preppdf.F' that sometimes resulted in a loss of explicitly excluded non-bonded pairs when these were not defined or read in before each call to 'preppdf.F'.

• 9/7/97, AŠ: Initialized the 'ialn()' array for the PAP, INSIGHT, and QUANTA formats ('rdaln.F'). This bug may have caused some problems when a short protein is added to the alignment that contained a longer protein (framework) in its place (e.g., with READ_ALIGNMENT after MALIGN3D).

• 9/13/97, AŠ: Renamed TOP variable MODELLER_STATUS into ERROR_STATUS and moved it within the top.ini file.

• 9/16/97, AŠ: Modified protsize.cst and Makefile to allow more flexibility in defining the sizes of the most important arrays.

• 9/20/97, AŠ: Introduced the ALIGN_BLOCK argument for the MALIGN command.

• 9/28/97, AB: New parameters:ASGL_2D_TYPE - controls the way 2D arrays are output by the ENERGY_PROFILE command. PROFILE_2D_PHYS - selects physical type to be presented as a 2D energy profile. NORMALIZE - selects normalization for ENERGY_PROFILE. ASGL_2D_OUTPUT - whether or not to write 2D output for ASGL. Command names changed: PATCH_DISULFIDES to PATCH_SS_MODEL. INHERIT_DISULFIDES - to PATCH_SS_TEMPLATES. ENERGY_PROFILE now outputs 2D profiles. OUTPUT can contain the type of energy you output by ENERGY_PROFILE: ENERGY or IDEAL_ENERGY1, and NORMALIZE specifies the type of normalization. Removed the error in calculation of the spline non-bonded terms.

• 10/28/97, AB: the bug in rsr9 (tmps size) fixed. Restored output of the objective function to the output PDB file.

• 1/2/98, AB: version 5f-ab: Introduced 32-nd physical restraint type and 7-th feature type (accessible surface in $\mbox{\AA}^2$). Libraries 'sched.lib', 'feats.lib', and 'top.ini' are changed to reflect this. New logical variable DYNAMIC_ACCESS was introduced. Limitation, each time before you want to use the 32-nd feature in ENERGY etc., you have to calculate solvent accessibility by WRITE_DATA OUTPUT=PSA.

• 1/5/98, AŠ: A bug (causing a crash) in seqsearch.F is fixed.

• 1/20/98, AB: version5g = merge (5f-ab, 5f-as). To use solvent accessibility: READ_PARAMETERS FILE = $LIB/par-32.lib

• 2/6/98, AB: version5g/doc: Improved the html version of the manual. Changes in Makefile. latex2html was patched to properly handle the macros containing sectioning commands.

• 1/5/98, AŠ: Added keyword DIH_LIB_ONLY to MAKE_RESTRAINTS, to allow forcing the calculation of the ``homology-derived'' mainchain and sidechain dihedral angle restraints (single and binormal) from the residue type dependent libraries only (templates ignored).

• 1/25/98, AŠ: changed default BUILD_METHOD back to INTERNAL_COORDINATES.

• 2/19/98, AŠ: corrected a bug in distdp3.inc which sometimes caused incorrect C-terminal overhang treatment, possibly incorrect alignment as a whole.

• 2/27/98, AŠ: MNCH2_LIB moved to MNCH_LIB.

• 3/05/98, AŠ: Changed the algorithm for calculating the mainchain curvature. Now using least-squares lines instead of simple ${C}_\alpha$- ${C}_\alpha$vectors.

• 3/12/98, AŠ: added option SIGNIF_CUTOFF to SEQUENCE_SEARCH.

• 3/23/98, AŠ: changed ALIGN_CODES to ALIGN_CODES2 for READ_ALIGNMENT2.

• found a workaround for a nasty compiler bug in mod2ali (for SGI, make opts) that caused naln=1.

• 4/28/98, AŠ: added MNCH?_LIB libraries to allow for i-1,i,i+1 mainchain conformation restraints.

• 4/30/98, AB: in 'model/build.F': all DISU stuff was deleted. atom classes are assigned based on the iss array, resnam does not change.

• 4/30/98, AB: In energy calculations, "slow" character parameters were replaced by "fast" integer ones.

• 6/6/98, AŠ: Created new command SEGMENT_MATCHING for enumerating the alignments.

• 6/6/98, AŠ: Removed SELECTION_STEP from MODELLER.

• 6/6/98, AŠ: Expanded options to PICK_ATOMS by allowing for automated selection of loops; a special value for SELECTION_SEGMENT of 'LOOPS' '' and new variable GAP_EXTENSION were introduced.

• 6/6/98, AŠ: Modified the loop modeling scripts and other scripts so that robust loop modeling can be done automatically.

  There are at least 43 such arrays with a dozen diffrent character length.

• 6/6/98, AŠ: Added the single mutant modeling script.

• 7/1/98, AŠ: Added the missing ALIGN_ALIGNMENT top 'top.ini'.

• 7/1/98, AŠ: Removed undefined entries from top.ini. Changed the meaning of RENUMBER_RESIDUES so that it corresponds to the first residue index in renumbering.

• 7/1/98, AŠ: removed sdch.lib for sidechain classes 1-2 from all parts, including COMPARE.

• 7/1/98, AŠ: changed format of mnch.lib, removed class weights from this library because they occur in mnch?.lib.

• 7/7/98, AB: the multimple binormal restraints were debugged: 1) detection of most relevant phase of phi/psi angles 2) removal of the bug in calculation of the derivatives

• 7/8/98, AŠ: increased flexibility of ALIGN2D so that it needs only at least one structure for the first block of sequences.

• 7/9/98, AB: CA-only warning in READ_MODEL and other routines using rdpdb

• 7/11/98, AB: New kind of error message is introduced: _B - to inform about bugs. Now used for former "Increase MAX..." and similar messages only. Scan_errors modified respectively.

• 7/18/98, AŠ and AB: BUILD_METHOD for BUILD_MODEL can continue to be 3D_INTERPOLATION, INTERNAL_COORDINATES, but can now also assume ONE_STICK and TWO_STICKS values to generate different initial structures for inserted atoms.

• 7/20/98, AŠ: biperiodic is now used

• 7/24/98, AŠ: not using schedule in PICK_RESTRAINTS at all.

• 8/24/98, AŠ: introduced EXPAND_CONTROL to allow more control for the EXPAND_ALIGNMENT command.

• 8/24/98, AŠ: made a vector out of RENUMBER_SEGMENTS.

• 8/24/98, AŠ: introduced NO_TER.

• 9/10/98, AŠ: updated COMPARE_SEQUENCES so that variability of the first sequence in the alignment is written to the Biso column of MODEL.

• 9/15/98, AŠ: changed default NONBONDED_SEL_ATOMS from 2 to 1, in top.ini.

• 98-00, AŠ: too numerous changes to list here, conversion from f77 to f90.