VERSION

==========     VERSION 1.76    ==========

* No more outputting of the PREFIX.cad cluster information file
  The file generated is too large.

==========     VERSION 1.75    ==========

* The whole population is saved during the simulation
  The structure is an array of chromosomes: chrom_snapshot

* A new function was created called cluster. The function
  clusters the whole genetic algorithm individuals.

* A new parameter allows to set the RMSD between poses in
  the clustering step at the end of the GA
  The parameter is included in the docking input file

  CLRMSD <float_value>

==========     VERSION 1.74    ==========

* allocation of box now using malloc rather than calloc,
  due to errors on windows with calloc.

* Next to improve: box could be allocated only once 
  at the start of the program (dimension-issue)

* Fixed roulette such in cases where a chromosome with index -1
  is selected (read access violation)

==========     VERSION 1.73    ==========

* altfdih is now freed (memory) in read_lig function.

* now using boost headers to generate random numbers
  also using boost::int32_t to avoid using RAND_MAX;

==========     VERSION 1.72    ==========

* Fixed a bug when reallocation opt_par structures.
  map_opt_par was not being reallocated like others.

==========     VERSION 1.71    ==========

* floats have been changed to doubles in genetic algorithms genes
  parameters and in CF calculations.

* Genetic algorithm now encodes all genes as 32-bits.
  The integral type saved is divided by RAND_MAX to gives a value between
  [0,1[ and this value is used to select the correct bin
  for a given parameter

==========     VERSION 1.69    ==========

* Modifies_pdb now lists ligands/modified a.a. as HETATM by default,
  if they are listed as ATOM
  However, they cannot be filtered out not including HET (INCHET)

* Optimization of vcfunction using double casts
  More precision when multiplying floats

* When an atom has type 0, we ensure complementarity is 0.000
  otherwise it would use block of memory[-1] (Type0 - 1 in matrix)

==========     VERSION 1.68    ==========

* Accounts for cases where residue names of RNA molecules are reversed
  e.g. '  C' --> 'C  '

* residue_conect is no longer executed when a RNA molecule is target

* Atom types are being assigned properly with RNA molecules

==========     VERSION 1.67    ==========

* Accounts for cases where ligands are listed as ATOM in PDB format
  A new function ensures the residue is a natural amino acids before
  reordering its atoms

==========     VERSION 1.66    ==========

* RNA molecules also go through modify_pdb.

* GLOBAL does not exist anymore.
  To do a GLOBAL simulation, you need to combine multiple clefts
  together. A grid will be built in all clefts.

==========     VERSION 1.65    ==========

* ori_atoms and rmsd_structures were removed.
  New member fields called coor_ori and coor_ref in atoms structure
  are now used to replace those. The new input format is now:
  
  RMSDST <RMSD_STRUCTURE.pdb>

  Only the matching atoms are used to calculate RMSD.
  (Matching residue number, name, chain and atom number)

* The code is optimized

* More than 3 GPA atoms are now possible.
  e.g.     0    0    0
        GPA1    0    0
        GPA2    0    0
        GPA2    0    0

* A new member field in atoms structure called graph is used to
  specify in which subgraph of the molecule the atom belongs to

==========     VERSION 1.62a    ==========

* The code is optimized

==========     VERSION 1.61     ==========

* Bugfix: Fixed the bug when generating the grid from the ensemble
  of spheres

==========     VERSION 1.60     ==========

* LOCCLF now takes as input a sphere file (cleft spheres)
  instead of a pre-built grid. The grid is now built automatically
  at the start of each run using C++ code (maps) within the 
  function generate_grid

* LOCCEN now has the same effect as LOCCLF as a grid is also
  generated within the boundaries of the sphere.

* SPACER can be used in CONFIG data file in order the set the spacing
  length between 2 vertices of the grid

* OUTRNG now prints a PDB atoms for the grid rather than PDB lines

* DELTAP is now equivalent to VARDIS, VARANG, VARDIH, VARFLX for
  respectively: Distance, Angle, Dihedrals, Flexible Dihedrals

* NORMAL is now equivalent to NMAMOD, NMAAMP, NMAEIG for respectively:
  Number of modes, amplitudes and eigenvectors

* Files partition_grid and slice_grid now uses C++ code.
  

==========     VERSION 1.58     ==========

* Default interaction matrix is now set to M6_cons_3.dat
  The interaction matrix allows to select the right 
  definition of types.
  e.g. when 12 types are read --> AMINO12.def will be used

* DEPSPA can be used in CONFIG data file in order to search
  for dependencies files elsewhere

* radii.dat file is no longer necessary for 
  overriding FlexAID's default radii

* Bugfix: Fixed the bug when deallocating memory
  at the end of simulations when using NMA

==========     VERSION 1.57     ==========

* Bugfix: Fixed the bug when reading normal grid files when
  using normal mode analysis (NMA) - Line buffering

* Bugfix: Fixed the alteration of coordinates of individual
  atoms using NMA

* Bugfix: A default value is now given to force_interaction
  when using constraints

==========     VERSION 1.55     ==========

* New feature: 'METTYP %d' can be used in CONFIGURATION file
  that allows to change the type of metals (HETATM)

==========     VERSION 1.54     ==========

* Bugfix: in flexible side-chain file, as well as 
  in constraints file, now automatically
  replaces '-' sign (no chain identifier) by ' '

* Changed the rotamer outputting format to match
  was is read by the NRGsuite

==========     VERSION 1.53     ==========

* New feature: 'REFPCG %f %f %f' can be used in the input 
  ligand file to force a reference protein center of geometry
  It allows to use the same one as the one used in the
  extraction process.

==========     VERSION 1.52     ==========

* Bugfix: nan happened when dividing by 0.0 when
  rebuilding atom coordinates. Perturbed atom
  coordinates.

* Contact areas between atoms now use the minimal area
  between in the 2 atoms in contact (min_areas function)
  	e.g. ( A --> B [30.0A^2] )
               B --> B [35.0A^2] )
	The area considered is 30.0

* Contact areas now use 'Radical Plane (R)' as opposed to
  'Extended Radical Plane (X)' in the Vcontacts function
  This allows better accuracy in contacting areas

* SAS calculations are now done within the vcfunction
  Previously, it was calculated within Vcontacts

* Bugfix: Intramolecular interactions of optimizable
  residues are now only calculated once

==========     VERSION 1.51     ==========

* Bugfix: order_pdb renamed to modify_pdb
  The function now controls filters applied
  such as removal of water/HETATM/H as well as 
  for the reordering of the PDB to match AMINO.def
  'A' conformations are saved by default  

==========     VERSION 1.50     ==========

* Bugfix: order_pdb can now account for inserted residues

* Bugfix: Fixed the bug for which BOINC users cannot end a task
  because the application is stalled at 100% because of memory deallocation
  Error: Vcontacts negative subscript when NC = 0

==========     VERSION 1.48     ==========

* Code optimization

* Bugfix: Vcontacts structures are allocated dynamically

==========     VERSION 1.47     ==========

* Bugfix: Initialization of some variables in release

* Bugfix: Fixed a bug when reading types definition

* Added a WARNING when atom types assigned exceed the number
  of atom types defined in the energy matrix

* Added safety when changing the solvent type using SLVTYP

==========     VERSION 1.46     ==========

* Bugfix: CONECT lines are fixed when atom indexing exceeds 9999

* Added error handling BOINC functions to check heartbeat

==========     VERSION 1.45     ==========

* ICDATA field is obsolete in .inp file for ligand

* NTYPES is obsolete in preprocessor defines
  Automatically assigned when reading Ematrix file

* SLVENT is redefined to SLVPEN

* SLVTYP '\d+' assigns a type to the solvent
  rather than using a fixed value against all atom types

* DEFTYP can be used to override the default
  atom definition files (AMINO.def/NUCLEOTIDES.def)

==========     VERSION 1.43     ==========

* Redirected all errors to STDERR to capture errors
  in the NRGsuite

==========     VERSION 1.42     ==========

* Dihedrals now have their own delta parameter value
  the field is DELTAF %f

==========     VERSION 1.41     ==========

* Fixed the bug when docking using RNA structures because
  of the re-ordering of PDB structures.

==========     VERSION 1.36     ==========

* Bugfix: Fixed the freeing of memory that stopped BOINC from
  giving valid results.

* Now fully Windows-compatible with BOINC
  Stack overflow because of huge structures.
  
* Bugfix: Fixed a bug in order_pdb to print out lines.

==========     VERSION 1.31     ==========

* Bugfix: Writing of the temporary PDB file.
  BOINC functions must not be used for temporary files
  but only for files writing in templates

==========     VERSION 1.30     ==========

* Fixed a bug when adding twice the same residue in flexible side-chains

* Constraints can now be as interaction between 2 atoms.
  A factor must be given, which multiplies the complementarity between those 2 atoms.
  Complementarity can be overriden (always be positive) by adding 'f' after the factor

     e.g.  INTERACT ...:10.0f 

* Added the SUPRNO in CONFIG.inp to explicitly assign SuperNodes Mode when using normal modes.

==========     VERSION 1.27     ==========

* In Flexible Side-Chains List file (in docking input file), ROTOBS is
now used as flag to the constant filename rotobs.lst.

* Default outputting was altered in order to account for NRGsuite interface.

==========     VERSION 1.26     ==========

* Fixed the error while free-ing up memory for data structures at 
  the end simulations.

==========     VERSION 1.25     ==========

* Compatibility with Windows XP platform.

* Changed the random seed number generator to srand() function.
  Also changed the random number generator to rand() in in-house function GenerateRandomNum().

==========     VERSION 1.22     ==========

* Fixed the error while generating the list of bonded atoms for side-chains.
  The list still does not include extra-residue bonds (peptide bonds).

==========     VERSION 1.21     ==========

* Added a new field 'MAXRES' in order to cap the number of results generated at the end of the GA. The param
  has to be filled in CONFIG.inp. The number of results (%d) has to follow 'MAXRES'.
  By default, the number is always set to 10.

==========     VERSION 1.20     ==========

* Changed the bonded structure for optimizable residues. Instead of having a list of neighbours for each atom
  in the optimizable residue, the whole residue is shown as a matrix N x N, in which N is the number of atoms
  of the optimizable residue.

* Added the new constraints to the function 'vcfunction'.

* Added the details on CF for each optimizable residue(s) in result file(s).

==========                      ===========