Burton A. Leland

Managing the Combinatorial Explosion

Computer software designed to deal with the large amounts of data generated by chemical and biological research programs is described. The software includes methods for representation of the structural components of combinatorial libraries, enumeration of structures and structure-based searches, and comparisons within combinatorial libraries.

Over 20 years of reaction access systems from MDL: a novel reaction substructure search algorithm.

From REACCS, to MDL ISIS/Host Reaction Gateway, and most recently to MDL Relational Chemistry Server, a new product based on Oracle data cartridge technology, MDLs reaction database management and retrieval systems have undergone great changes. The evolution of the system architecture is briefly discussed. The evolution of MDL reaction substructure search (RSS) algorithms is detailed. This article mainly describes a novel RSS algorithm. This algorithm is based on a depth-first search approach and is able to fully and prospectively use reaction specific information, such as reacting center and atom-atom mapping (AAM) information. The new algorithm has been used in the recently released MDL Relational Chemistry Server and allows the user to precisely find reaction instances in databases while minimizing unrelated hits. Finally, the existing and new RSS algorithms are compared with several examples.

The substance module: the representation, storage, and searching of complex structures

Chemical structures are typically represented in computer programs as simple graphs, where atoms are represented by a list of nodes and the bonds by a list of nondirectional edges. While this convention allows for the representation of a large variety of chemical structures, it does not lend itself toward the representation of many common substances such as polymers, nonstoichiometric mixtures, and formulations. An extension to this convention has been developed which allows properties to be identified with a defined subgraph in a structure, an Sgroup. This extension has been implemented in the Substance Module, a new MACCS-11 module, and is used to represent and search a much broader class of chemical substances. A description of the new representation and searching capabilities is given as well as examples of its use.

Self-contained sequence representation: bridging the gap between bioinformatics and cheminformatics.

The wide application of next-generation sequencing has presented a new hurdle to bioinformatics for managing the fast-growing sequence data. The management of biomacromolecules at the chemistry level imposes an even greater challenge in cheminformatics because of the lack of a good chemical representation of biopolymers. Here we introduce the self-contained sequence representation (SCSR). SCSR combines the best features of bioinformatics and cheminformatics notations. SCSR is the first general, extensible, and comprehensive representation of biopolymers in a compressed format that retains chemistry detail. The SCSR-based high-performance exact structure and substructure searching methods (NEMA key and SSS) offer new ways to search biopolymers that complement bioinformatics approaches. The widely used chemical structure file format (molfile) has been enhanced to support SCSR. SCSR offers a solid framework for future development of new methods and systems for managing and handling sequences at the chemistry level. SCSR lays the foundation for the integration of bioinformatics and cheminformatics.

Computer Representation and Searching of Chemical Substances

Chemical structures are typically represented in computer programs as simple graphs, where the atoms are represented by a list of nodes and the bonds by a list of non-directional edges. We will describe an extension to such a representation which allows properties to be identified with a defined subgraph in a structure; these properties are fully searchable at a substructure level. A description of the representation will be given and examples of searching capabilities will be illustrated. An implementation of these techniques has been applied to homopolymers, copolymers, non-stoichiometric mixtures, formulations, and ‘superatoms’. Other potential uses will also be discussed. These extensions to chemical representation allow us to represent and search a much broader class of chemical substances than the set of discrete chemical structures which have previously been handled.

VET: A Tool for Reaction Plausibility Checking

Production of chemical reaction databases is a multistep process, with the possibility of errors at each of these steps. VET is a tool developed to trap errors in the chemical reactions identified as a part of this process. VET has been designed to minimize the acceptance of incorrect reactions, while still supporting various common practices in reaction depiction, including unbalanced reactions, suppressed components, and reactions with alternative products. We discuss the assumptions made in its construction, a general overview of its structure, and some performance characteristics.

Representation and searching of biomolecules

Biomolecules present challenges to chemical information systems designed for small molecules. Their sizes, up to tens of thousands of atoms, overwhelm representation/storage/searching solutions built on explicit chemical representation of the structures. But biomolecules are largely made up of many repeats of a limited number of building-block molecules, a fact which has been used to provide a compressed representation for biomolecules using templates for the building blocks. We have adopted a modified template-based representation for biomolecules. Our primary interest is in the chemically modified portions of biomolecules, for which we choose to use explicit chemistry. These areas of explicit chemistry are then embedded in the template-compressed, unmodified portions of the full biomolecule. The regions containing explicit chemistry are indexed, and thus can be structure searched with good performance. A limited number of residues surrounding explicit chemistry regions are included in the index for searching the context of these explicit regions. By using explicit chemistry to represent modified regions we can search across classes of modifications for common features. For example a single substructure search query will find green fluorescent protein, and its histidine, phenylalanine and tryptophan analogs. Templates are stored with the structure providing a self-contained file format. The use of NEMA keys allows templates from different structures to be compared, and allows storage of structures containing a canonical list of templates. The residues have defined attachment points, allowing automated traversal of a protein backbone, or location of non-backbone bonds to residues. We will present example structures and structural queries highlighting capabilities of our representation.

Unlocking stereochemical diversity

A little more than 50 years ago Pfeiffer [1] noted that the more active a chiral drug was, the less active its enantiomer was. Recent years have seen increased success in screening compounds in silico. While the majority of 2D screening approaches are insensitive to stereochemistry, this is not the case in 3D docking, which is typically used to refine compound lists. Unfortunately, a number of common workflows result in generation of only one member of the number of structures actually registered in a 2D database [2]. This results in a fraction of candidate structures being ignored in silico, even though they are present in the real compound libraries being considered.