(604c) Systematic Incorporation of 2D, 3D and Electrostatic Characterization in Molecular Design

Computer-aided molecular design (CAMD) is emerging as a very powerful tool in chemical and pharmaceutical industries today. The ability to design products, which meet a set of specified target properties, exemplifies the applicability of CAMD. With the advent of new computing technologies, a more detailed and comprehensive description of molecular structures is possible. This growth in the ability to model and predict a wide range of molecular properties should parallel the increased stringency of regulations on environmental, health and safety standards. Thus, more tools must be developed which take advantage of relationships between molecular structure and the macro scale properties realized for these systems. It is well understood that most properties of interest are not solely a function of the topological arrangement and connectivity of the molecules under investigation. The spatial arrangement, e.g. steric hindrance or ligand binding, and electrostatic nature of these structures are also necessary to reveal more optimal candidate molecules. It has been suggested that searching the chemical space with a combination of 2D and 3D descriptors could prove more effective than employing single methods [1].

This contribution details a systematic methodology for doing just that, and it is based on the signature molecular descriptor [2]. One difficulty in the application of reverse molecular design using geometric based descriptors is that the minimum energy molecular conformation is a function of the complete structure. The most stable conformation is ultimately determined by interactions among subatomic particles through the force fields they generate, which can be accurately modeled by quantum mechanics. The effect of these force fields diminishes as a function of increasing inter-particle distance. Thus, a reasonable estimation of local geometry can be made by considering atoms extending radially, up to fixed distance, from a central atom. The signature descriptor is ideal for this approach as it details the extended valence sequence of an atom up to a pre-defined height or distance away from the central atom. Previous studies have proven the applicability and effectiveness of atomic signature descriptors utilized in molecular design with topological and topographical considerations [3-5]. These signatures can be geometry optimized through various quantum mechanical approaches, and this information is conveniently stored in the form of a geometry matrix. Also, the electron densities of these signature “fragments” can be generated through ab initio calculations, given the previously obtained geometry information, and recombined to produce complete molecular electron distributions. This approach, known as MEDLA (Molecular Electron Density Lego Approach) [6], utilizes a “fuzzy” electron density description so that boundary adjustments upon fragment recombination are unnecessary. The nature of these signature descriptors and boundary-less electron cloud fragments allow for a linear optimization approach and are consequently much less computationally demanding than other ab initio quality calculations. The framework and mathematical techniques necessary for formulation and solution of property or activity based molecular design problems, with 2D, 3D and electrostatic characterization, are presented here along with a case study which illustrates the benefits of such an approach.

[1]  Nettles J.H., Jenkins J.L., Bender A., Deng Z., Davies J.W., and Glick M. (2006) Bridging Chemical and Biological Space: “Target Fishing” Using 2D and 3D Molecular Descriptors. J. Med. Chem., 49, pp. 6802-6810

[2]  Faulon J.-L., Visco D.P. Jr., and Pophale, R.S. (2003) The Signature Molecular Descriptor. 1. Using Extended Valence Sequences in QSAR and QSPR Studies. Journal of Chemical Information and Computer Sciences, 43, pp. 707-720.

[3]  Herring R.H., Rudolfs N., Chemmangattuvalappil N.G., Roberts C., and Eden M. (2012). Molecular Design Using Three-Dimensional Signature Descriptors. Computer Aided Chemical Engineering, 31, (In Press).

[4]  Chemmangattuvalappil N.G., Solvason C.C., Bommareddy S. and Eden M.R. (2009) Incorporating Molecular Signature Descriptors in Reverse Problem Formulations. Computer aided Chemical Engineering, 27(1), pp. 73-78.

[5]  Chemmangattuvalappil N.G., Solvason C.C., Bommareddy S. and Eden M.R. (2009) Novel Molecular Design Technique using Property Operators based on Signature Descriptors. Computer aided Chemical Engineering, 27(2), pp.897-902

[6]  Walker P. D., Mezey P. G. (1993) Molecular Electron Density Lego Approach to Molecule Building. J. Am. Chem. Soc., 115, pp. 12423-12430.