Abstract

The work is aimed at modifying an algorithm called the Activation-Relaxation Technique (ART). The algorithm was initially envisioned by the original authors to be applied to the study of materials. This work, however, aims to take the original vision forward by applying it to the exploration of potential energy surfaces (PESs) of chemical and enzymatic reactions. To make ART more efficient in terms of the force evaluations needed and, consequently, to reduce the computational time and cost, it has been modified to include additional algorithms that significantly speed up the exploration of the PES. Each such algorithm is designed such that it allows for a strategic displacement of the atoms and subsequent sharing of atomic displacement vector information. Such a strategic displacement of atoms, for example, involves exploiting an a priori knowledge of the reactants in order to avoid atomic displacement moves that would not contribute to the tracing of their PES. A molecule like N-methylacetamide, for instance, which is a good model system for proteins, only undergoes a conformational change (a gradual dihedral rotation) which acts as the sole reaction coordinate for its PES and thus algorithms are included in ART that disallow for any other atomic displacement moves except for a gradual rotation of the dihedral. The sharing of atomic displacement vector information can prove helpful, for example, in enzymatic reactions where the enzyme mutants (which are similar in structure to the wild-type) can be made to follow the displacement vector information of the wild-type in order to avoid going astray along the PES and to have a good starting point from the beginning. The goal, ultimately, is to apply ART to the study of complex chemical reactions and enzymatic reactions involving large-scale rearrangements of the protein environment. As validation steps in the development of this algorithm, we illustrate the modified ART method on selected small molecules -- ethane, propane, methyl acetate, N-methylacetamide, etc. -- highlighting pitfalls and successes.