View source for Relaxing existing minima with new potential and creating new database

Sometimes it may happen that you need to change potential. All minima that you found using previous potential can be relaxed in new potential and you do not need to start your job from the very beginning!

There is quite simple way to do it. First you need to extract coordinates of all minima from binary file ''points.min'' and minimize each structure using new potential, and then you have to create new ''points.min'' file.


Here is a step-by-step recipe:

'''0'''. update potential - in case of [[AMBER]] users it means new topology file

'''1'''. create subdirectory ''relax_min''

  mkdir relax_min

'''2'''. copy some necessary files to that subdirectory:

[[AMBER]] users:

  cp points.min min.data odata.start perm.allow coords.prmtop coords.inpcrd min.in relax_min

[[CHARMM]] users:

  cp points.min min.data odata.start perm.allow input.crd relax_min

'''3'''. change name of ''odata.start'' into ''odata'' (the same file is used in BHINTERP calculations, see [http://wwmm.ch.cam.ac.uk/wikis/cuc3/index.php/Using_BHINTERP_to_find_minima_between_two_end_points here] . Basically it should give you a minimization with the normal mode analysis performed at the end, with DUMPDATA also present to give the output ''min.data.info'' file).

  mv odata.start odata

'''4'''. [[AMBER]] users: create script ''relax'' as follows:

  #!/bin/csh
  
  set count = 1
  
  while ($count < 621)
  
     echo NATOMS 3002 > pathdata
     echo COPYFILES coords.prmtop coords.inpcrd min.in >> pathdata
     echo AMBER9 >> pathdata
     echo EXTRACTMIN $count >> pathdata
     /home/wales/bin/pathsample.2.1 >& pathsample.out
     mv extractedmin start
     (time /home/wales/bin/A9OPTIM.4.0) >& OPTIM.$count
     cp min.data.info min.data.info.$count
     head -1 min.data.info >> min.data.check
     @ count +=1
  
  end

[[CHARMM]] users: create script ''relax'' as follows:

  #!/bin/csh
  
  set count = 1
  
  while ($count < 621)
  
     echo NATOMS 3002 > pathdata
     echo COPYFILES input.crd >> pathdata
     echo CHARMM >> pathdata
     echo EXTRACTMIN $count >> pathdata
     /home/wales/bin/pathsample.2.1 >& pathsample.out
     (time /home/wales/bin/COPTIM.4.0.31_64) >& OPTIM.$count
     cp min.data.info min.data.info.$count
     head -1 min.data.info >> min.data.check
  
     @ count +=1
  
 end

Also make sure that in the [[CHARMM]] section of odata the file name specifying where the coordinates are read in from is ''./extractedmin.crd'', to match the output from PATHSAMPLE.

'''NOTE''': some values in this script are relevant to my system, like NATOMS=3002, so you need to change it according to atom number of your system. The number of count is equal to number of minima in your database + 1 (number of lines in ''min.data'' file, where is blank line on the end) therefore this value should be also changed according to your ''min.data'' file. Paths to executables should be also updated if necessary.

'''5'''. submit this script to queue on cluster:

  #PBS -q s1
  
  cd $PBS_O_WORKDIR
  ./relax

When finished you get a set of ''min.data.info.*'' files with coordinates that are optimized using new potential (if you didn't forget to update the force field :)). 

'''6'''. next step: creation new binary file ''points.min''; you will need new ''pathdata'' file:

  READMIN min.data.info.all
  NATOMS        3002
  TEMPERATURE   0.592
  GEOMDIFFTOL   0.1
  TRIPLES
  STARTTRIPLES
  ADDTRIPLES

[[AMBER]] users: add to ''pathdata'' file these three lines:

  EXEC /home/wales/bin/A9OPTIM.4.0
  COPYFILES   perm.allow coords.inpcrd coords.prmtop min.in
  AMBER9

[[CHARMM]] users: add to ''pathdata'' file these three lines:

  EXEC /home/wales/bin/OPTIM.4.0.31_64
  COPYFILES   perm.allow input.crd
  CHARMM

File ''min.data.info.all'' contains all ''min.data.info.*'' files and you can use cat command to make it easily (as is shown in next step).

'''7'''. remove old ''points.min'' and ''min.data'' files as well as ''min.A'' and ''min.B'', and submit this script to the queue on cluster:

  #PBS -q t1
  
  cd $PBS_O_WORKDIR
  
  for i in `find . -name "min.data.info*"  `; do cat $i>> min.data.info.all; done
  /home/wales/bin/pathsample.2.1 >& pathsample2.out

That's all! Now you have new ''min.data'' and ''points.min'' files. Flags -t -r ensure the same order of minima as in previous ''points.min'' file. It is useful because both endpoints should be placed again on the first and second positions and without these flags files can be sorted in a bit different order. Still it is worthy to check where in fact both endpoints are placed in new database. Minimize both endpoints in new potential to get their energies. If energies on first and second positions in new ''min.data'' file correspond to energies of minimized endpoints - it's OK! If not, you need to edit ''min.data'' and ''points.min'' files as is described [http://wwmm.ch.cam.ac.uk/wikis/cuc3/index.php?title=Removing_minima_and_transition_states_from_the_database here].