Title: Quantum chemical data storage needs, structures, and directions in industrial applications

Speaker: Lance M. Westerhoff, General Manager, QuantumBio Inc. 200 Innovation Boulevard, State College, PA, USA.

Abstract: Traditionally, in the pharmaceutical world, quantum mechanics (QM) is applied to molecular systems such as ligands and ligand fragments made up of a few 10’s of atoms. However, advances in the development of fast, accurate, semi-empirical QM Hamiltonians and more recently in the implementation of linear scaling algorithms(1,2) have made possible the use of QM to routinely characterize protein/ligand(3) and even protein/protein(4) complexes made up of hundreds or thousands of atoms. Today, the computational chemist armed with modern linear scaling QM methods can generate megabytes of primary data – including Fock and density matrices – and megabytes of secondary or derived data – including final energies, atom-by-atom pairwise energy terms, eigenvectors, and so on. As these volumes are measured on a per-biomolecular complex basis, one can imagine a single routine screening simulation involving tens or hundreds of complexes generating many gigabytes of data. When one couples the copious volume requirements of these sorts of simulations along with the fact that many of the data types that are generated from QM methods do not have analogues in the conventional, molecular mechanics regime, it soon becomes apparent that new ways must be devised to store and access QM biochemical data.

As part of this discussion, experiences with SQL-based relational storage, look up tables, flat files, hierarchal data stores, and proprietary data formats will be compared. Various QM-based biomolecular data types will be presented along with an outline of decisions QuantumBio has made concerning this important topic. At the same time, some of the realities of working with industrial pharmaceutical research will be noted.

(1) Dixon and Merz Jr. “Fast, accurate semiempirical molecular orbital calculations for macromolecules.” The Journal of Chemical Physics (1997).

(2) Dixon and Merz Jr. “Semiempirical molecular orbital calculations with linear system size scaling.” The Journal of Chemical Physics (1996).

(3) Zhang X, Gibbs AC, Reynolds CH, Peters MB, Westerhoff LM. “Quantum Mechanical Pairwise Decomposition Analysis of Protein Kinase B Inhibitors: Validating a New Tool for Guiding Drug Design.” Journal of Chemical Information and Modeling (2010).

(4) Diller DJ, Humblet C, Zhang X, Westerhoff LM. “Computational alanine scanning with linear scaling semiempirical quantum mechanical methods.” Proteins (2010) vol. 78 (10) pp. 2329-37.