Where do we currently stand with cheminformatics-driven medicinal chemistry?

On Wednesday 18 October 2006 at the eCheminfo Autumn InterAction Meeting we are convening a forum on Bench Scientists’ & Modellers’ Discussions on Discovery Tools & Modeling.

In this session a panel of experimental and computational chemists will discuss their experiences in using computational modeling methods in drug discovery. They will discuss where the methods and software are having success, and where current methods are not yet meeting their needs, are failing or have challenges or complications. Short presentations on drug discovery experiences will be used to seed discussion of cheminformatics-driven medicinal chemistry and lead optimization and conversations on where new developments could aid improvement in practice and tools.

The Panel consists of a group of experienced practitioners in both medicinal and computational chemistry: Osman F. Güner (Turquoise Consulting), James Arnold (AstraZeneca), Phil Edwards (AstraZeneca), Pete Connolly (Johnson & Johnson PRD), Michael Farnum (Johnson & Johnson PRD), and panel chair Jim Wikel (Coalesix).  This panel will seed discussions with presentations of the following experiences in drug discovery:

James Arnold (AstraZeneca), Use of computational techniques in the discovery of low molecular weight non-peptidic beta-secretase inhibitors: Application of fragment screening and structure-based design to identify and progress millimolar affinity hits to sub-micromolar leads

Phil Edwards (AstraZeneca), The Discovery of low molecular weight non-peptidic beta-secretase inhibitors: Application of fragment screening and structure-based design to identify and progress millimolar affinity hits to nanomolar leads

Pete Connolly (Johnson & Johnson PRD), New Tools for Structure-Activity Relationship Visualization: Applications to Drug Discovery

Michael Farnum (Johnson & Johnson PRD), Applying Chemical Intelligence to support Drug Discovery Researchers

Jim Wikel (Coalesix), Man, Machine & Drug Design

For abstracts of these presentations, please follow the continuation below.

Barry Hardy

(Continued….)

Bench Scientists’ & Modellers’ Discussions on Discovery Tools & Modeling
eCheminfo InterAction Meeting Session, Bryn Mawr, Philadelphia, USA
http://www.innovationwell.net/COMTY_medchemistry/
Wednesday, 18 October 2006
chaired by Jim Wikel (Coalesix)

Presentations

Use of computational techniques in the discovery of low molecular weight non-peptidic beta-secretase inhibitors: Application of fragment screening and structure-based design to identify and progress millimolar affinity hits to sub-micromolar leads
James R. Arnold, AstraZeneca

James R. Arnold, Phillip D. Edwards, Jeffrey S. Albert, Gerard M. Koether, Christopher R. Holmquist, Donald W. Andisik, Russell C. Mauger, Mark A. Sylvester, Nathan Spear and James B. Campbell

Alzheimer’s disease (AD) is characterized by the progressive formation in the brain of amyloid plaques and vascular deposits composed of the beta-amyloid peptide (Abeta). Abeta is generated in vivo through the proteolytic cleavage of the membrane-anchored beta-amyloid precursor protein (APP) by beta- and gamma-secretases. Cleavage of APP by beta-secretase is the rate-limiting step of Abeta production, and beta-secretase is a target for the treatment of AD.

Molecular scaffolds targeting Beta Secretase (BACE) were identified using a combination of high concentration fragment screening and large-scale crystallography. In a series shown, a fragment with weak micromolar affinity afforded compounds with single digit micromolar affinity after a small number of compounds were synthesized. We describe how the team then rapidly produced potent compounds using a combination of techniques from Medicinal Chemistry, Computational Chemistry and Structural Biology.

The Discovery of low molecular weight non-peptidic beta-secretase inhibitors: Application of fragment screening and structure-based design to identify and progress millimolar affinity hits to nanomolar leads
Philip D. Edwards, AstraZeneca

Alzheimer's Disease (AD) afflicts millions of the elderly worldwide, resulting in a tremendous social and economic impact. The most widely accepted hypothesis for AD is that mis-metabolism of amyloid precursor protein by beta- and gamma-secretases generates beta-amyloid, which forms neurotoxic plaques. Beta-secretase is a membrane-associated aspartyl protease responsible for the initial step in the processing of APP to beta-amyloid. Despite intense efforts for over a decade in both academics and the pharmaceutical industry, it has only been within the last year that reports of the first low-molecular weight beta-secretase inhibitors have begun to emerge. In this presentation, we will describe how we have used multiple affinity screening techniques and a close collaboration between medicinal, computational and physical chemists to identify and characterize fragments with weak affinity (IC50 >100 uM), and evolve them into nanomolar, non-peptidic inhibitors of beta-secretase.

New Tools for SAR Visualization: Applications to Drug Discovery
Peter Connolly, Johnson & Johnson PRD

Traditionally, analysis of structure-activity relationships (SAR) is a time-consuming process that involves manual dissection of chemical structures into their component parts and correlation of scaffolds and substituents with activity or potency at a biological target. Although medicinal chemists are comfortable with hands-on, visual representations of data, too often SAR analysis relies on static numerical correlations at the expense of flexibility and clarity. There is a pressing need for SAR analysis tools that provide a adaptable, intuitive, and highly interactive link between chemical structure and biological activity. By bringing together pictorial representations of chemistry and graphical visualization of biological and property data, such tools would establish new paradigms for SAR studies and facilitate the drug discovery process. In this presentation, new interactive SAR visualization tools developed at Johnson & Johnson PRD will be discussed. Several examples demonstrating their use in analysis of data sets from recent research projects will be shown.

Applying Chemical Intelligence to support Drug Discovery Researchers
Michael Farnum, Johnson & Johnson PR&D

ABCD is an informatics platform developed in-house at Johnson & Johnson that bridges multiple continents, data systems and cultures using modern information technology, and provides researchers with an environment that allows them to make better decisions. The system is build upon a data warehouse, which combines data from multiple chemical and pharmacological transactional databases, organized to support flexible querying at high performance. The main interface to ABCD is Third Dimension Explorer, which facilitates data upload, retrieval, mining and reporting. A central goal of ABCD is to provide users with the means to retrieve, view and analyze multifactorial SAR data. Key to the success of this effort is the ability to combine fast substructure and similarity searching with conventional relational queries, and deliver the results in an expedient and visually compelling format. In this presentation, we give an overview of ABCD, and focus on a few core components that represent the system's "chemical intelligence", including the chemical cartridge, sketcher, molecular spreadsheet and interactive data mining components.

Man, Machine & Drug Design
James H Wikel, Colaesix

Computational methods have become an important tool in many medicinal chemistry drug discovery and development efforts. Significant contributions have been made to several marketed drugs. However, the complete utilization of these technologies has not been realized by the medicinal chemists. The application of computational methods remains primarily in the hands of expert computational chemists. Lipinski popularized a few simple property calculation methods in the “Rule of Five” which were later added to by Veber using two more computationally intense properties. The question remains as to why computational methods are not in more widespread use by medicinal chemists. We will examine potential challenges and introduce a new approach to leverage both the power of computational methods and the experience of medicinal scientists. The method, known as Interactive Evolutionary Computing (IEC), has been applied successfully in other industries in which a complex optimization is achieved using both man and machine. Drug candidate design and optimization is a complex conundrum with two key characteristics, both of which are naturally addressed by the strengths of IEC. First, a successful drug candidate must satisfy a battery of complex and interrelated in vitro/in vivo potency and ADMET requirements that are difficult, if not impossible, to describe in a simple computer based model fitness function. While a skilled medicinal chemist can navigate these requirements, progress is slow and inefficient. Second, there are several components of the problem - such as synthetic tractability or IP status – that are subjective or inherently implicit and so do not lend themselves to a purely computational approach. IEC leverages both the expertise of the medicinal chemist and the computational skills of the computational chemist as key elements of this evolutionary search process.

Computational Tools for Medicinal Chemists
Osman F. Güner, Turquoise Consulting

For over a decade now, many software companies have attempted to provide computational productivity tools for medicinal chemists with limited success. In this presentation we will review these past attempts and evaluate the reasons for limited success.

Early attempts used the “platform” as a differentiating factor. Those days the hard-core computational tools were available mainly on mainframe computers and later for UNIX workstations. The thinking by the vendors was that if these tools were made available on a PC, then the medicinal chemists would use them. Of course, since the user interface for these applications maintained the full complexities of the original applications, they remained unusable by the medicinal chemists who are not trained in computational chemistry. The second wave of attacks attempted to use the Web browsers to eliminate the complexity barrier. Since the medicinal chemists were already familiar with Web browsers, the entry barrier would be lower. In this case, the vendors were challenged to make complex user options available through browsers. Whereas, this approach worked fairly well for “batch” type operations (like database searching, 3D structure generation, 3D or structural similarity screening) this became a daunting task for more interactive tasks such as molecular docking, manual alignments, and pharmacophore perception). The further challenge that was in front of the vendors was the users’ “fault tolerance”. The expert computational chemists were fairly fault tolerant. Even if the applications crash every other hour, they are still intrigued if the unique algorithm provided them the scientific advantage. This is not the case with the medicinal chemists who, unlike the computational chemists that spend 90-100% of their time in front of a computer, spend perhaps only 20% of their time with their computers. This means that usability becomes a main concern for the latter group. Also they would expect the application to be much less buggy.

Have we learned from the past experiences yet? What is the appropriate platform for developing computational productivity tools for the medicinal chemists? And if we started our design concept from the perspective of the medicinal chemist, making sure that the application addresses their critical needs and is going to be used by them and be useful to them, what kind of a system can we envision? This presentation will propose answers to the above questions.

InnovationWell eCheminfo cheminformatics chemoinformatics bioinformatics Medicinal Chemistry Computational Chemistry Virtual Screening Docking Molecular Modelling Molecular Modeling pharmaceutical pharma meeting workshop conference management Bryn Mawr Philadelphia Critical Path toxicology personalised medicine Life Sciences Pharma Drug DiscoveryResearch and Development Drug Development Healthcare Innovation Knowledge Management events