SEAM with Amber

• F. Jensen "Locating Minima on Seams of Intersecting Potential Energy Surfaces. An Application to Transition Structure Modeling." J. Am. Chem. Soc. 114 (1992) 1596-1603.
• F. Jensen "Transition Structure Modeling by Intersecting Potential Energy Surfaces." J. Comp. Chem. 15 (1994) 1199-1216.
• P. T. Olsen, F. Jensen "Modelling Chemical Reactions for Conformationally Mobile Systems with Force Field Methods." J. Chem. Phys. 118 (2003) 3523-3531.
• F. Jensen "Using Force Field Methods for Locating Transition Structures." J. Chem. Phys. 119 (2003) 8804-8808.

Preparing the input

1. Create a force field for the start and end structures of the reaction.
   Q2MM (PON parametrisation) is a very useful method if you have non-standard structures.
   
   
2. Create two input files for tleap, one for the start and for the end force fields, both with the same pdb coordinates. Name these X.leap and Y.leap (X and Y can be anything as long as they are the same at all occasions in this document). See the sample file below.
   The two files should generate prmtop and prmcrd files, called X.prmtop and X.prmcrd, and similar for Y.
   The numbering of the atoms in both leap outputs (i.e. the prmtop and prmcrd files) have to be exactly the same. Otherwise SEAM will not work.
   The following two commands gives you output pdb-files to check.
   tleap -f X.leap
   ambpdb -p X.prmtop <X.mdrest >test.pdb
   
   
3. Copy the nmode input files from /home/bio/SEAM to your working directory. These are named nmode.in and nmode_seam.in.
   cp /home/bio/SEAM/nmode*in .
   
   
4. Copy the seam input file (seam.dat) from /home/bio/SEAM to your working directory.
   cp /home/bio/SEAM/seam.dat .
   
   
5. Edit the seam.dat file for your needs.
   ! at the beginning of a line indicates a comment that will be ignored.
   All available options are in the file but only those needed are un-commented in the default file.
   
   The most important options are described here (those that you should consider to modify). Some more are described in the default seam.dat  file, whereas the other ones are only described in the source code, rddata.f
   
   • MAXSTEP - the number of iterations to run before giving finding a minima
   • INPFILE - should be AMBER if you are not using another program and then you are in the wrong web page
   • OPTALGO - NRTS is best, it is a Newton-Raphson Transition State optimiser.
   • FUNCTION - LAGRANGE is default, other options are availabe if this doesn't work.
   • VDVMOD - exchange the Lennard-Jones non-bonded function with a Buckingham potential. This is recommended. It should only be used together with STRMOD
   • STRMOD - exchange the regular energy function for bonds with a Morse-type  potential. This is recommended. It should only be used together with VDWMOD
   • HESS - ANAL is the option to choose to get the best results. It reads the Hessian from nmodeh.
   • PRINT - The amount of output data, 0 to 4 is available. 0 is only what you really need, 4 is huge amounts of data.
   • GEOMETRY - The geometry of your starting input. PRODUCT, REACTANT or AVERAGE.
   
   
6. If you want to fix the coordinates of some atoms, create the file "fixatoms" and list the atom numbers, one on each row.
   
   
7. To get better energies, you can replace the bond dissociation energies of the Morse bond function (used when STRMOD is ON) by creating the file "morsepar" in the following way:
   It should be in four column format with variables
   atomtype1 atomtype2 bond_order energy
   where atom types are Macromodel atom type ids (see extern.f in the source code for transfer from amber types)
   bond_order is 1, 2, or 3
   energy is bond-dissociation energy in kJ/mole.
   
   
8. Using non-standard atom types:
   Atom types in the program are handled using Macromodel types and all the standard atom types from the amber FF99 and gaff force fields have been implemented. To include new atom types, you have to edit the source code in extern.f
   You find the functions atmnamechange (atom type change) and ambbondtype (bond type determination) at the bottom of the file.
   Preferably, an external file could be created from which one could import this data, but this has not been implemented yet.

Running the program

Sample leap input file for X

Modifications of Amber nmode