Molecular Diversity Assessment using Chemotypes.

BACKGROUND: Many techniques to design chemical libraries for screening have been put forward over time. General use libraries are still important when screening against novel targets, and their design has relied on the use of molecular descriptors. In contrast, chemotype or scaffold analysis has been used less often. OBJECTIVE: We describe a simple method to assess chemical diversity based on counts of the chemotypes that offers an alternative to model chemical diversity. We describe a simple method to assess chemical diversity based on counts of the chemotypes that offers an alternative to model chemical diversity based on computed molecular properties. We show how chemotype counts can be used to evaluate the diversity of a library and compare diversity selection algorithms. We demonstrate an efficient compound selection algorithm based on chemotype analysis. METHODS: We use automated chemotype perception algorithms and compare them to traditional techniques for diversity analysis to check their effectiveness in designing diverse libraries for screening. RESULTS: The best type of molecular fingerprints for diversity selection in our analysis are extended circular fingerprints, but they can be outperformed by the use of a chemotype diversity algorithm, which can be more intuitive than traditional techniques based on molecular descriptors. Chemotype- -based algorithms retrieve a larger share of the chemotypes contained in a library when picking a subset of the chemicals in a collection. CONCLUSIONS: Chemotype analysis offers an alternative for the generation of a general-purpose screening library as it maximizes the number of chemotypes present in a subset with the smallest number of compounds. The applications of methods based on chemotype analysis that does not resort to the use of molecular descriptors are a very promising but seldom explored area of chemoinformatics.

Development of an informatics platform for therapeutic protein and peptide analytics.

The momentum gained by research on biologics has not been met yet with equal thrust on the informatics side. There is a noticeable lack of software for data management that empowers the bench scientists working on the development of biologic therapeutics. SARvision|Biologics is a tool to analyze data associated with biopolymers, including peptides, antibodies, and protein therapeutics programs. The program brings under a single user interface tools to filter, mine, and visualize data as well as those algorithms needed to organize sequences. As part of the data-analysis tools, we introduce two new concepts: mutation cliffs and invariant maps. Invariant maps show the variability of properties when a monomer is maintained constant in a position of the biopolymer. Mutation cliff maps draw attention to pairs of sequences where a single or limited number of point mutations elicit a large change in a property of interest. We illustrate the program and its applications using a peptide data set collected from the literature.

Mining and visualizing the chemical content of large databases.

The shift in the drug discovery paradigm that has resulted in increased dependence on large volumes of information for decision making has created new challenges and opportunities in chemoinformatics research. The large volumes of data that are now available present an opportunity for the development of better predictive models, for the selection of more promising drug candidates, and for more comprehensive decisions to be made by researchers in chemistry and biology. The progress of this approach in drug discovery has also introduced challenges associated with the need to develop better tools to manage the large volumes of data. For chemistry-related research in particular, there is a need for more effective retrieval methods and better tools for data analysis that incorporate good interface design to facilitate the use of such systems by the non-expert scientist. This article reviews recent developments in the mining of large volumes of chemistry data, with a focus on chemoinformatics tools for small-molecule drug discovery and tools developed for analysis and in-depth mining. External sources of data, such as a variety of web services, further expand the volume of data to be analyzed during any small-molecule discovery project. The impact of the generation of large volumes of data and the related research factors is reshaping chemoinformatics work.

Design of chemical libraries for screening.

IMPORTANCE OF THE FIELD: The ultimate goal of discovery screening is to have a fast and cost-effective strategy to meet the demands of producing high-content lead series with improved prospects for clinical success. While high-throughput screening (HTS) dominates the drug discovery landscape, other processes and technologies have emerged, including high-content screening and fragment-based design to provide alternatives that may be more suitable for certain targets. There has been a growing interest in reducing the number of compounds to be screened to prevent the escalation in the costs, time and resources associated with HTS campaigns. Library design plays a central role in these efforts. AREAS COVERED IN THIS REVIEW: This opinion provides a survey of some recent developments in the diversity based library design process, but within a historical context. In particular, the importance of chemotyping and substructure analysis and the challenges presented by novel lead discovery technologies that require the design of libraries for screening are discussed. WHAT THE READER WILL GAIN: Readers will gain an appreciation of some developments in the field of library design and the factors that are driving the development of new library design technologies; specifically, challenges presented for chemoinformatics with the novel screening technologies in diversity based screening and compound filtering. TAKE HOME MESSAGE: Chemotyping and substrutural analysis are techniques that have been underutilized in the process of library design. However, they offer a direct way to evaluate libraries and have been successfully used to develop predictive methodologies. Tools are available to this end, but the full power of the approach has not been realized yet.

Computational techniques in fragment based drug discovery.

Fragment based drug discovery is gaining acceptance as a complement to other more established techniques to identify leads and optimize drug candidates. In this review we illustrate areas where fragment based drug discovery has had an impact and point to some examples that show how fragment based analysis is being applied to new arenas. The traditional uses of computational methods in fragment based for lead discovery and optimization and for risk assessment are briefly summarized. The application of fragment analysis for the definition of bioisosteric replacements are discussed together with techniques to characterize the diversity of chemical libraries based on fragment distribution.

Ring systems in mutagenicity databases.

The distribution of ring systems in public mutagenicity databases is analyzed. An automated enumeration of substructures permits determination of the occurrence of different scaffolds in data sets. The counts are used to perform population analysis via proportions and odds ratios of mutagenic compounds. Pairwise calculations of odds ratios between scaffolds allow comparison of ring systems for isostere replacement studies. These findings are presented in tables that readily show which scaffold is likely to occur in mutagenic compounds. Also, rings identified in public domain mutagenicity data sets are compared to rings in drugs data sets; unfortunately, public mutagenicity data sets do not reflect the types of scaffolds in drugs and those typically used in medicinal chemistry. The findings bring into question the utility of predictive models that were derived from public domain data sets. The automated ring identification and statistical approaches used here can be applied to other pharmacological properties to yield information about chemical scaffolds.

Novel cyclooxygenase-1 inhibitors discovered using affinity fingerprints.

We used protein affinity fingerprints to discover structurally novel inhibitors of cyclooxygenase-1 (COX-1) by screening a selected number of compounds, thus providing an alternative to extensive screening. From the affinity fingerprints of 19 known COX-1 inhibitors, a computational model for COX-1 inhibition was constructed and used to select candidate inhibitors from our compound library to be tested in the COX-1 assay. Subsequent refinement of the model by including affinity fingerprints of inactive compounds identified three molecules that were more potent than ibuprofen, a commonly used COX-1 inhibitor. These compounds are structurally distinct from those used to build the model and were discovered by testing only 62 library compounds. The discovery of these leads demonstrates the efficiency with which affinity fingerprints can identify novel bioactive chemotypes from known drugs.

Using NMR for ligand discovery and optimization.

Several recent technology-driven advances in the area of NMR have rekindled an interest in the application of the technology to problems in drug discovery and development. A unique aspect of NMR is that it has applicability in broadly different areas of the drug discovery and optimization processes. NMR techniques for screening aimed at the discovery of novel ligands or low molecular weight structures for fragment-based build up procedures are being applied commonly in the industry. Application of NMR in structure-guided drug design and metabonomics are also becoming routine. We present an overview of some of the most recent NMR developments in these areas.

Genome-wide profile of oxidoreductases in viruses, prokaryotes, and eukaryotes.

Enzymes that utilize nicotinamide adenine dinucleotide (NAD) or its 2'-phosphate derivative (NADP) are found throughout the kingdoms of life. These enzymes are fundamental to many biochemical pathways, including central intermediary metabolism and mechanisms for cell survival and defense. The complete genomes of 25 organisms representing bacteria, protists, fungi, plants, and animals, and 811 viruses, were mined to identify and classify NAD(P)-dependent enzymes. An average of 3.4% of the proteins in these genomes was categorized as NAD(P)-utilizing proteins, with highest prevalence in the medium-chain oxidoreductase and short-chain oxidoreductase families. In general, the distribution of these enzymes by oxidoreductase family was correlated to the number of different catalytic mechanisms in each family. Organisms with smaller genomes encoded a larger proportion of NAD(P)-dependent enzymes in their proteome (approximately 6%) as compared to the larger genomes of eukaryotes (approximately 3%). Among viruses, those with large, double-strand DNA genomes were shown to encode oxidoreductases. Gram-positive and gram-negative bacteria showed some differences in the distribution of NAD(P)-dependent proteins. Several organisms such as M. tuberculosis, P. falciparum, and A. thaliana showed unique distributions of oxidoreductases corresponding to some phenotypic features.

A path from primary protein sequence to ligand recognition.

A novel method to organize protein structural information based solely on sequence is presented. The method clusters proteins into families that correlate with the three-dimensional protein structure and the conformation of the bound ligands. This procedure was applied to nicotinamide adenine dinucleotide [NAD(P)]-utilizing enzymes to identify a total of 94 sequence families, 53 of which are structurally characterized. Each of the structurally characterized proteins within a sequence family correlates to a single protein fold and to a common bound conformation of NAD(P). A wide range of structural folds is identified that recognize NAD(P), including Rossmann folds and beta/alpha barrels. The defined sequence families can be used to identify the type and prevalence of NAD(P)-utilizing enzymes in the proteomes of sequenced organisms. The proteome of Mycobacterium tuberculosis was mined to generate a proteome-wide profile of NAD(P)-utilizing enzymes coded by this organism. This enzyme family comprises approximately 6% of the open reading frames, with the largest subgroup being the Rossmann fold, short-chain dehydrogenases. The preponderance of short-chain dehydrogenases correlates strongly with the phenotype of M. tuberculosis, which is characterized as having one of the most complex prokaryotic cell walls.

Chemoproteomics as a basis for post-genomic drug discovery.

The large number of small organic compounds now available for drug-lead screening has led to numerous methods for classifying molecular similarity and diversity, the aim being to restore a balance between the quantity and drug-like quality of compounds in small-molecule libraries. Whereas structural and physicochemical attributes continue to be emphasized in compound selection for drug-lead screening, chemoproteomics--the use of biological information to guide chemistry--offers a highly efficient alternative to small-molecule characterization that can accelerate drug discovery in the post-genomic era.