Cheminfostream

Virtual Screening Methods

Specific binding interactions are central to many biological processes and pathways. Similarly, most drugs act by binding specifically to a site on a target protein, thereby modulating protein activity. The quest for new drugs relies on many approaches, including computer-based virtual screening and docking. Over the past fifteen years, and in parallel with the exponential increase in the number of available high-resolution protein structures, many screening and docking methods and programs of use in the drug discovery process have emerged. Understanding the similarities and differences of different methods as well as their capabilities and limitations is both important and increasingly challenging.

The main objective of our Virtual Screening eCheminfo Community of Practice activity is to foster discussion amongst researchers working on both development of screening and docking methods and the application of such methods to drug discovery. This interaction is intended to lead to a better understanding of the current state-of-the-art, improved screening and docking tools in the future, and enhanced awareness of how to apply the current set of tools.

On Tuesday 17th October 2006 a number of leading screening experts and practitioners will meet at the joint eCheminfo and InnovationWell Community of Practice meeting at Bryn Mawr College, Philadelphia to discuss virtual screening and docking methods. 

On the afternoons of both Monday 16th and Tuesday 17th October we will also hold a number of workshops on latest virtual screening and docking methods and software.

On the afternoon of Tuesday 17th a forum will discuss current virtual screening and docking methods and software, results of existing validation and comparison studies, and procedures for useful independent comparative studies that could be undertaken by the community of practice.

The group of presenters and workshop leaders includes Stan Young (National Institute of Statistical Sciences), John Irwin (UCSF), William Douglas Figg (Nantional Cancer Institute), Daryll Reid (SimBioSys), Neysa Nevins (GlaxoSmithKline), Deepak Bandyopadhyay (Johnson & Johnson PRD), Paul Hawkins (OpenEye), Shashi Rao (Schrodinger) and Alex Clark (Chemical Computing Group).

I provide below a summary of the presentations and workshops. (Follow the continuation…)

Barry Hardy

(Follow the continuation…)

Open Seminar & Knowledge Cafe on ELNs

InnovationWell will hold an Open Seminar & Knowledge Cafe on Electronic Laboratory Notebooks (ELNs) at Bryn Mawr on Tuesday 17 October 2006 to discuss the emerging roles of Knowledge Management and Electronic Laboratory Notebook solutions in managing multidisciplinary research activities, innovation and collaboration.  The Open Event will take place alongside the InnovationWell and eCheminfo Autumn InterAction Meetings.  Attendance at the Open Event & Knowledge Café involves no registration fee. However seated places are limited and will be restricted to a confirmed guest list of 100. Please send an RSVP with your name and organisation via email to innovationwell [at] douglasconnect.com

More information is available at: http://barryhardy.blogs.com/theferryman/2006/09/electronic_lab_.html

Structure-based Drug Design

I provide below a summary of the presentations and workshops to take place in our eCheminfo Structure-based Drug Design session to be held Monday 16th October 2006 at the joint eCheminfo and InnovationWell Community of Practice meeting at Bryn Mawr College, Philadelphia

Barry Hardy

Structure-based Drug Design
eCheminfo InterAction Meeting Session, Bryn Mawr, Philadelphia, USA
http://www.innovationwell.net/COMTY_drugdesign/
Monday, 16 October 2006
chaired by Frank Hollinger (Locus Pharmaceuticals)

While there are a multitude of approaches being deployed in early stage drug discovery, few have undergone the multitude of changes that structure based drug design has over the past 10-20 years.  Structure based drug design combines science and technology from the fields of computational chemistry, informatics, medicinal chemistry, biology, biochemistry, structural (crystallography and NMR).  Collectively scientists involved in these disciplines form the teams to make a structure based design effort successful.

The convergence of improved structural capabilities (highly efficient protein generation and purification techniques, high throughput crystallography, SAR by NMR, etc.), improved computational algorithms combined with faster, economical compute capability to evaluate ligand/ protein interactions, and the evolution of higher throughput chemistry techniques (e.g. parallel synthesis, focused library design and synthesis, etc.) has resulted in an increase in successful applications of structure based drug design.

It is not uncommon in this generation of drug discovery to start a therapeutic program with a 3D structure (co-crystal or NMR) of your ligand/ protein complex.  This complex structure can significantly enhance a discovery team’s understanding of the ligand’s binding to the protein as well as provide insights in how to enhance those interactions. 

The main objective of the Structure Based Drug Design symposium is to present several case studies of structure based design highlighting some of the diverse approaches deployed to illustrate best practices for the next time they will repeat the process.  We will expect to hear how structural information is capable of identifying hits as well as optimization of those hits to incorporate the necessary drug-like properties for in vivo efficacy.  This session should foster discussion among researchers working in early stage drug discovery through the optimization process about the best practices teams should consider using to achieve success in SBDD focused projects.  This interaction will lead to a better understanding of the current state-of-the-art, improved structure based design approaches and processes and enhanced awareness of how to apply the current set of tools.

Presentations

Structure-Based Design of Estrogen Receptor-beta Selective Compounds

Michael S. Malamas, Wyeth

Michael S. Malamas, Heather A. Harris, James C. Keith, Jr., Robert McDevitt, Iwan Gunawan, Christopher P. Miller, Eric Manas and Richard Mewshaw

The discovery of a second subtype of the estrogen receptor (ER-beta) in 1996 prompted an intense discovery effort within the scientific community to identify ER-beta selective ligands in order to elucidate the receptor’s physiological role in mediating estrogen action. However, until very recently ER-beta’s physiological role has remained unclear. One approach that has not yet been fully exploited to date is the use of highly ER-beta selective ligands to elucidate the functional role of ER-beta. Toward this end, we have designed highly potent and selective agonists for ER-beta and have characterized their activity in several clinically relevant rodent models.

While only two subtle amino acids differences are present in the ligand binding domains of the two ER-alpha and ER-beta isoforms, our structure-based approach enabled us to rapidly advance our initial leads to successful candidates. X-ray crystallography data and molecular modeling tools allowed us to exploit a single amino acid difference between the two ERs (ER-betaIle421 to ER-alphaMet373). Several ligands were found to highly bind to ER-beta with IC50 values of 3-5 nM, which are similar to 17beta-estradiol. However, unlike estradiol, these ligands are >100 fold selective for ER-beta over ER-alpha.

Our ER-beta selective ligands exhibited little or no utility in hormone therapy or as contraceptive agents (e.g., uterotrophy, osteopenia, mammotrophy, thermoregulatory dysfunction, and ovulation) where nonselective estrogens (e.g. 17beta-estradiol) or ER-alpha selective agonists are known to have robust effects. However, we have found that ER-beta selective agonists have a dramatic beneficial effect in animal models of various chronic inflammatory conditions (inflammatory bowel disease, arthritis) and they are currently undergoing clinical trials as novel therapies to treat such conditions. The data suggest that one function of ER-beta may be to modulate the immune response, and that ER-beta selective agonists may offer a novel therapy to treat chronic inflammatory conditions.

Harnessing the power of Structure Based Drug Design using a Fragment Based Approach

Frank Hollinger, Locus Pharmaceuticals

Prospective structure based drug design is one of the more challenging aspects of drug discovery. A powerful design process has been developed which allows the identification and optimization of novel, diverse and potent ligands with druggable properties. This design process has been made possible by harnessing the power of fragment centered structure based drug design.

The development of a novel grand canonical Monte Carlo (GCMC) simulation paradigm permits the generation of binding free energies for fragments to a protein. The output of such a simulation provides all the information necessary to accurately identify high affinity interaction sites, binding sites and druggable binding sites on a protein surface. This computational approach also enables the design team to understand the role of water (just another fragment in our paradigm) and to take full advantage of that information. Using novel analysis tools we are able to combine fragments into synthetically accessible drug-like molecules for evaluation against a desired protein target. A case study will be presented describing the design of novel, potent and selective allosteric inhibitors for p38-alpha. Experimental data will be presented which validates our process of using fragment based ligand design approaches to identify leads and optimize their properties.

Structural Interactions of CCR5 with HIV-1 entry inhibitors

Debananda Das, National Cancer Institute

Debananda Das, Kenji Maeda, Philip Yin, Kiyoto Tsuchiya, Hiroaki Mitsuya

Affiliation: HIV & AIDS Malignancy Branch, National Cancer Institute, National Institute of Health; 10 Center Drive, Room 10S255 - MSC 1868; Bethesda, MD 20892-1868

Effective treatment of HIV continues to be a daunting challenge due to the emergence of drug resistant mutations in the target viral enzymes. CCR5 is a novel cellular target for the intervention of HIV replication. However, an X-ray or NMR structure of CCR5, a GPCR, does not exist. By combining the results of site directed mutagenesis experiments, homology modeling, and docking that accounted for the flexibility of the receptor side chains, we characterized the structural and molecular interactions of CCR5 with multiple CCR5 inhibitors active against R5 HIV-1 including a potent in vitro and in vivo CCR5 inhibitor aplaviroc. The quality of the structural model was evaluated by carrying out new saturation binding experiments by mutating CCR5 residues predicted to be important by the model. The structural model enabled us to precisely define the binding site of CCR5 inhibitors within CCR5 and elucidated the key binding site interactions responsible for the anti-viral activity of the inhibitors. We will discuss structure based drug design strategies that target specific residues of CCR5 to minimize toxic side effects.

Main points to be covered:
* Combination of site directed mutagenesis and molecular modeling can be a powerful tool for structure based drug design
* Iterative structure refinement that accounts for the flexibility of the receptor is important for generating a robust homology model.

Small Molecule Inhibitors of Protein-Protein Interactions

Max Cummings, Johnson & Johnson PR&D

Normal cell function is dependent upon many protein-protein interactions. Specific inhibitors of protein-protein interactions can serve as useful tools in the study of biochemical pathways, and may ultimately lead to the development of new drugs. Protein-protein interactions are perceived as a potential source of many new targets for drug action, but at the same time are thought to be particularly challenging targets for drug discovery. They are commonly characterized as a distinct target class with respect to small molecule drug discovery. Structures related to the HDM2-p53 protein-protein interaction will serve as an introduction, followed by a more general discussion of selected structural aspects of protein-protein and protein-small molecule binding interactions.

Using ab initio calculations as routine tools to help design CDK2 inhibitors

Jose Duca, Schering-Plough

Ab initio methods can be used systematically in drug design to provide insights when experimental data is not easily obtainable. In this talk we present three examples of Structure Based Drug Design where ab initio tools played a prominent role. In the first study ab initio methods were utilized to compute pKa values using a model of the catalytic site of TACE and to predict a proton transfer concomitant with binding. Second, the use of ab initio calculations to compute pKa values and tautomer properties of a series of substituted pyrazolopyridines (CDK2 inhibitors) is presented. Finally, a series of ab initio free energy calculations is used to identify determinants of binding affinity for some recently published pyrazolopyridine inhibitors of CDK2.

High Strain Energies of Bound Ligands

Paul Labute, Chemical Computing Group

The prediction of the bioactive bound conformation of a candidate ligand is important for computational methodologies such as pharmacophore search and docking. The strain energy of a conformation (relative to the global minimum energy) is often used as a criterion for rejection of a conformation from consideration. Recent molecular mechanics studies using ligand-receptor complexes from the PDB have suggested that high strain energies (> 10 kcal/mol) are not only possible but routinely observed. We present the results of computational experiments that attempt to explain these observations and determine their validity.

WORKSHOPS

Quantum Biochemistry Workflows

Lance Westerhoff, QuantumBio

QuantumBio, Inc. in conjunction with Discovery Machine, Inc. has been working to apply computer learning and knowledge management to fully automate the multi-step processes required to characterize biomolecular interactions at a quantum mechanical level within in silico drug discovery workflows. Such workflows involve database searches, structure preparation, molecular mechanics-based cleanup, and finally quantum mechanical treatment in order to fully characterize these interactions. Each of these steps can include any number of subtasks. During the in silico drug discovery process, these complicated workflows are coupled with simulations that involve the characterization of hundreds if not thousands of biomolecular structures at a time. In addition to simulation parameters themselves, quantum mechanics methodologies are notoriously sensitive to structural defects in which the convergence of the calculation will be adversely affected. This leads to longer calculation times and other problems. Therefore, these simulations often require that the user understand not only the chemistry of the structure, but also the theory involved in the computational methodologies so that problem structures can be filtered early in the process.

With this in mind, an intelligent and adaptive system for quantum mechanics-based, in silico drug discovery has been developed to encapsulate these workflows to describe the types and strengths of enzyme-inhibitor interactions that play an important roll in drug discovery efforts. This system is built on the synergy between QuantumBio’s CHEMIX molecular modeling user interface and Discovery Machine’s workflow management solution. The goal of this workshop is to introduce the community to an early version of this system in order to demonstrate its usefulness, and to gain feedback for continued development.

Fragment- and Structure-Based Drug Design

Zenon Konteatis, Jennifer L. Ludington & Frank Hollinger, Locus Pharmaceuticals

Structure-based drug design (SBDD) has evolved over the years and has seen many technological advances which have enabled significant discovery team successes. One of the more recent advances to SBDD has been the development of fragment based techniques.

This workshop will focus on a novel fragment based design process which uses a novel computational approach to calculate predicted binding free energies for a collection of fragments binding to a protein. The design process employed uses highly evolved analysis software to assemble potent, synthetically accessible lead molecules. This process has the natural ability to provide insights into how to further optimize the affinity and physical properties of the designed molecules.

The workshop will consist of two parts, the first part will present the steps in the fragment focused design process using real world examples. The second part will demonstrate the computational software designed specifically to analyze the fragment binding free energy information to yield novel, selective and diverse (physical property) molecules.

A key objective of the Workshop is to foster discussion around fragment focused structure based drug design approaches and illustrate a best practices example which has led to potent, selective, efficacious molecules.

Participants will have ample opportunity to discuss their perspectives and criticisms of the methods studied and should take-away key nuggets of understanding from this intensive session. Participants should return to their labs with new ideas, best practices and software experiences to maximize productivity in their own drug discovery research activities.

Advanced Techniques in Pharmacophore Perception and Successful Applications in Drug Design

Osman F. Güner, Turquoise Consulting

Pharmacophore perception and use of pharmacophores in drug discovery and design has evolved from a specialist activity in early 90s into an essential aspect of modern computer-aided drug design. The significance of pharmacophore technology has been continuously increasing together with the increasing availability of protein targets. In this workshop we will briefly cover the historical evolution of the pharmacophore concept and its successful use and application in drug discovery.

The bulk of the lecture focuses on various techniques for pharmacophore modeling and database searching. Methods for perceiving a pharmacophore are presented, starting from the simplest method of visual pattern recognition. The significance of training set selection is covered with an example for PDE IV inhibitors. Techniques for pharmacophore model refinement are then presented for 5-HT3 inhibitors, and finally predictive model development is covered with two examples: FTP inhibitors and antimalarial agents.

The second half of the workshop starts with a review of hit list and pharmacophore analysis metrics. It then provides details for various three-dimensional database-searching techniques. Vector-based queries are exemplified with an application to HIV-1 protease inhibitors and endothelin antagonists. Methotrexate bound conformation within DHFR is used to demonstrate the utility of ligand-shape in 3D searching, as well as combined shape and pharmacophore search. This section ends with a discussion of the conformational flexibility issue in 3D databases and provides a comparison of current solutions.

The availability of new targets and protein structures is now generating a renewed interest in structure-based drug design. In a similar fashion, there is now more interest in the development of receptor-based pharmacophores, as opposed to the traditional ways of using a set of known active molecules to derive a putative pharmacophore. Development of pharmacophore models entirely from the receptor active site provides new challenges. Typically, an active site will have more potential binding sites than the ones that are utilized by a given set of active compounds. Hence, automated pharmacophore model generation from receptor active sites becomes a combinatorial problem and the hundreds to thousands of pharmacophore models that are generated in this manner need to be evaluated and scored. We will discuss these problems and propose solutions based on some recent work. The workshop will close with several examples from literature and success stories.

Further reading:
1. Güner OF: The impact of pharmacophore modeling in drug design. IDrugs (2005) 8(7):567-572.
2. Van Drie J: Pharmacophore-based virtual screening: A practical perspective, In Virtual Screening in Drug Discovery. Alvarez J, Shoichet B (Eds), CRC Press, Boca Raton, FL, USA (2005):157-205.
3. Güner OF: History and evolution of the pharmacophore concept in computer-aided drug design. Curr Top Med Chem (2002) 2(12):1321-1332.
4. Mason JS, Good AC, Martin EJ: 3D pharmacophores in drug discovery. Curr Pharm Design (2001) 7(7):567-597.
5. Güner, OF (ed): Pharmacophore Perception, Development, and Use in Drug Design. IUL Biotechnology Series, La Jolla, California, USA (2000).

Hypothesis generation from docking results using activity measurements, interaction fingerprints, clustering and 2D visualization methods

Alex Clark, Chemical Computing Group

Given the availability of crystallography data for a drug target, it is possible to generate a large number of reasonable docked poses using modern software. This workshop will address the use of protein:ligand interaction fingerprints, combined with activity data, to reduce the noise which is inherent in docking results. A combination of clustering methods and 2D visualization can be used to produce model hypotheses, which can be applied to subsequent screening of compound databases.

InnovationWell eCheminfo cheminformatics chemoinformatics bioinformatics Computational Chemistry Molecular Modelling Molecular Modeling pharmaceutical pharma meeting workshop conference management executive Bryn Mawr Philadelphia Critical Pathbiomarkers metabolomics toxicology personalised medicine KM Life Sciences Pharma Drug DiscoveryResearch and Development Drug Development Healthcare Innovation Knowledge Management events

Press Release:  http://www.prweb.com/releases/2006/9/prweb443727.htm

Appyling Predictive Toxicology in Drug Discovery & Development

On Thursday 19th October 2006 a number of leading experts and practitioners will meet at the joint eCheminfo and InnovationWell Community of Practice meeting at Bryn Mawr College, Philadelphia to discuss new methods and challenges in predictive toxicology so as to enable the development of safer drugs through early-stage computational predictions of toxic properties of potential lead compounds.  The group includes Tudor Oprea (Univ. New Mexico), Navita Mallalieu (Roche Pharmaceuticals), Alex Tropsha (UNC), Curt Breneman (RPI), Sanji Bhal & David Adams (ACD/Labs), Michael B. Bolger (Simulations Plus and USC School of Pharmacy), Bob Clark (Tripos), and Gilles Klopman (Multicase).

On the preceding afternoon on Wednesday 18th October 2006 we will hold a number of workshops in this same area followed by an evening poster session and bbq.  I hope you can join the activities and conversations.

I provide below a summary of the presentations and workshops.

Barry Hardy

Predictive Toxicology
eCheminfo InterAction Meeting Session, Bryn Mawr, Philadelphia, USA
http://www.innovationwell.net/COMTY_predtox/
Thursday, 19 October 2006
chaired by Curt Breneman (RPI)

The ability to make informed decisions during the early phases of drug discovery is the key to decreasing hit-to-lead and lead optimization cycle times.  The motto “fail early, and fail cheap” represents the need to identify problematic chemotypes early in the development process so that more productive lines of inquiry may be followed.  The field of predictive modeling has reached a point where there is a realistic expectation that troublesome moieties can be flagged through computational virtual screening.  The next major step in the development of predictive methods is to be able to also use these modeling techniques to suggest productive courses of action to identify and correct ADME and PK problems during lead compound optimization.  Thus, the use of validated, interpretable models may serve both as a way of identifying ADME/Tox and PK failures, and also provide a means for correcting them.

The Physical basis for the Rule of Five
Tudor I. Oprea, University of New Mexico School of Medicine

Tudor I. Oprea (1), Scott Boyer (2), and Igor Tetko (3)
(1) Division of Biocomputing, UNM School of Medicine, MSC 11 6145, Albuquerque NM 87131-0001, USA
(2) Safety Assessment, Astrazeneca R&D Molndal, SE-431 83 Mölndal, Sweden
(3) München Information Center for Protein Sequences, GSF, Germany

Our understanding of the quality of leads for drug discovery rests on mining the known bioactivity and medchem spaces. Such sources are the WOMBAT database [1], which contains over 136,000 unique chemicals and 307,000 biological activities. A derivative database related to clinical pharmacokinetics is WOMBAT-PK (WB-PK) [2]. WOMBAT-PK 2006 contains 935 drugs with multiple human ADME/Tox endpoints: > 750 oral bioavailability and half-life data, > 700 plasma protein binding and volume of distribution (steady state) values, > 500 total clearance, non-renal clearance and maximum recommended therapeutic daily dose values, etc. Matching clinical data with calculated properties, one can gain better insights for lead discovery. The relationship between the Maximum Recommended Therapeutic (daily) Dose, MRTD, its value corrected for the fraction unbound, MRTD_U, and the partition coefficients (clogP and LogD74) will be discussed. These two parameters provide a physical basis to the Lipinski "Rule of five".
* the talk will explain the importance of property filtering in lead discovery
* it provides simple guidelines for computed property cut-off values
* it will discuss the importance of multi-parametric predictions to assist decision-makers in virtual and biomolecular screening
[1] Olah M, et al. WOMBAT: World of Molecular Bioactivity, in Chemoinformatics in Drug Discovery. Oprea TI (Ed), Wiley-VCH, New York, 2005, pp. 223-239.
[2] WOMBAT and WOMBAT-PK are available from http://www.sunsetmolecular.com

A Roadmap for Integrating Modelling & Simulation in Pre-Clinical DMPK Research
Navita Mallalieu, Roche Pharmaceuticals

Integration of modeling and simulation in preclinical research is still a dream for some organizations and in its in infancy for others. Bringing a new twist to existing pathways is both a challenge and an exciting opportunity. But most importantly, bringing in new blood into an entrenched system must deliver on its promise of a better outcome than the current paradigm. I’d like to share with you my experience with incorporation of modeling and simulation into the world of preclinical DMPK at Roche. The emphasis of this presentation is on development and utilization of a PBPK based PK model approach. We have tapped into the vast amount of preclinical data that is generated as part of routine drug development at Roche and tied it together using the PBPK model to show management an alternative approach for decision-making in discovery. The presentation will cover the following topics:
* Results of validation of the rat PK model using 45 compounds and a retrospective prediction of human pharmacokinetics using 10 compounds.
* Contrast of the PBPK approach with currently utilized approaches like allometric scaling and the FDA recommended Maximum Recommended Starting Dose (MRSD) for the first dose in humans.
* A proposal for incorporating M&S into the discovery screening strategy. I will share a road-map of how best to utilize the latest tools and expertise in M&S to complement the existing information and generate knowledge.

Our emphasis was to improve decision-making by showing an integrated view of the preclinical data. This approach also offers the potential to reduce experimentation by replacing it with simulations, where appropriate, thereby decreasing cycle-times.

The statistical significance vs. mechanistic interpretation of ADME/tox models
Alex Tropsha, UNC

Alexander Tropsha (1), Ann Richard (2), Kun Wang(1), Maritja Wolf (2), Clarlynda Williams (2), Jamie Burch (2)

(1) Laboratory for Molecular Modeling, School of Pharmacy, UNC-Chapel Hill, Chapel Hill, NC 27599
(2) Mail Drop D343-03, National Center for Computational Toxicology (NCCT), Office of Research & Development, US Environmental Protection Agency, Research Triangle Park, NC 27711

Several major trends affecting public toxicity information resources have the potential to significantly alter the future of predictive toxicology. Standardized chemical structure annotation of toxicity databases and integration of diverse biological activity data afford a mine-able chemical semantic Internet. Formalized toxicity data models and public toxicity data schemas allow for flexible data mining and relational data searching across layers of chemical and biological information. Curated, systematically organized, and web-accessible toxicity and biological activity data in association with chemical structures is clearly the next frontier of advancement for QSPR and data mining technologies. The examples of such systems are provided by the DSSTox database and affiliated projects such as the Carcinogenic Potency Database (CPDB), PubChem, Leadscope ToxML, and the National Toxicology Program.

The importance of the combined toxico-cheminformatics and QSPR modeling is illustrated by the analysis of the CPDB. It contains TD50 data resulting from animal cancer tests of over 1000 chemicals as well as the information on species, strain, sex, shape of the dose-response, etc. Such extensively structured and mine-able database provides unique opportunities for developing QSPR models for subsets of compounds selected on the basis of biologically meaningful parameters. Rigorous QSAR analysis was applied to a subset of 693 compounds tested for their mutagenicity. 70 compounds were selected randomly for external validation. The remaining 623 compounds were split into diverse training set and test sets. Classification kNN QSPR approach afforded modes with the prediction accuracy for training, test and external validation sets as high as 0.917, 0.847, and 0.893, respectively. The analysis of chemical descriptors that afforded statistically significant and predictive QSAR models allowed for the mechanistic interpretation of such models in terms of chemical properties important for mutagenicity. We stress that only those QSPR models that have been rigorously validated both internally and externally should be considered for the mechanistic interpretation.

Predictive ADME: How do I know if my predictions will be useful?
Curt Breneman, RPI Chemistry

Given the current capabilities of academic, commercial and semi-commercial statistical modeling tools and descriptor generators, there is a clear need to establish “best practice” methods and criteria for model validation - particularly in an area as complex as Predictive Toxicology. Different philosophies exist regarding the number and type of descriptors that can be used for generating predictive models, and the number of cases necessary for creating linear or non-linear models that can reliably cover important regions of chemical structure/activity space. To what extent does a model need to be interpretable in order to produce trusted predictions? What are the lower limits of training set size to address given types of problems? Is there a way to determine whether a given set of unknowns are within the domain of a particular model? These and other related questions will be discussed, along with examples of classification, regression and non-linear regression models of molecular ADME properties.

An In-Silico Approach to Reduce the Burdens of Lead Discovery and Optimization
Sanji Bhal and David Adams, Advanced Chemistry Development, Inc. (ACD/Labs)

Lead discovery and optimization are the challenging endeavors of balancing the efficacy of potential drugs with their pharmacokinetic properties. Although potency is essential, so is the ability of a compound to penetrate through various biological barriers and act upon the intended target site.

Use of physicochemical property predictors has long been known to help filter possible hits for desired ADME characteristics. However, the prediction quality is frequently dependent on the particular chemistry used in the training set. ACD/Labs' predictions are based on an additive-constitutive fragmental approach and therefore offer some unique advantages. This approach offers the capability to easily increase the prediction accuracy of specific, often novel, classes of compounds by training the algorithms with experimentally measured pKa, logD, and now solubility data. Moreover, this approach has allowed the development of a unique software tool utilized in lead optimization to rapidly identify structural modifications. These modifications are expected to give analogs with improved physicochemical properties for better bioavailability, absorption through the GI tract, and penetration (or lack of penetration) through the blood brain barrier. In this presentation, we will introduce examples of how this software can be applied to adjust physicochemical properties that directly impact the in-vivo behavior of drugs.

Integration of Early ADME using Property Estimation and PBPK Simulation
Michael B. Bolger, Simulations Plus and USC School of Pharmacy

Purpose: To introduce methods and describe validation results for the integration of in silico and in vitro data in early discovery.

Introduction: A wide variety of software tools are available for biopharmaceutical property estimation related to absorption, distribution, metabolism, elimination, and toxicity (ADMET). These in silico methods are supplemented by in vitro measurement of log P, solubility, Caco-2 Papp, fraction unbound in plasma, and metabolic stability. We will present methods and validation results for integration of all of this data in a database that is linked to whole body physiologically-based (PBPK) simulation of ADME.

Methods: By using estimated and experimental values for biopharmaceutical properties, fraction absorbed (Fa) and fecal excretion were simulated in silico using GastroPlus™. Bioavailability (Fb) and Cp vs. time profiles were simulated with knowledge of: intrinsic clearance, volume of distribution (Vd) and plasma protein binding (fup). Vd was estimated by using physiologically-based pharmacokinetics linked to a tissue composition model parameterized in silico using just log P and fup. In vitro experiments were required for estimation of clearance and bioavailability.

Results: We also found that the simulation model was able to correctly predict the dissolution and plasma concentration vs. time profile for poorly soluble drugs. We were able to predict the bioavailability for drugs with high first pass extraction using PBPK. Finally, we were able to correctly predict the non-linear dose dependence for substrates of influx transporters and for talinolol which is an effluxed substrate for Pgp.

Conclusions: Integration of biopharmaceutical properties from in silico estimation and in vitro experiments with a mechanistic, physiologically-based gastrointestinal simulation is a valuable predictive tool in pharmaceutical discovery and development.

The "Structures" in Structure-Activity Relationships
Bob Clark, Tripos, Inc.

Most in silico ADME/Tox work is based, in one way or another, on analyses of structure-activity relationships (SARs). People carrying out such work are often very sensitive to the limitations of their assay data, as they should be. They rarely consider uncertainties involving molecular structure, however. Such studies generally presume that a single structure adequately reflects each particular compound under consideration, but this assumption is dangerous when there is potential for protonation, deprotonation, tautomerization or stereochemical ambiguity. Even bond isomerization can be problematic in some circumstances, particularly where molecular similarity or fragment analysis is involved. Moreover, in many cases where a single structure can adequately represent a compound, the appropriate structure to consider is different in different contexts. Hence there is a real need to consider structure as a very high-dimensional problem that extends well beyond simple molecular connectivity.

Machine Intelligence in the Design of New Biologically Active Chemicals
Gilles Klopman, Multicase

The design of new biologically active chemicals, whether pharmaceuticals or agricultural requires a good understanding of their potential beneficial activity, their ability to reach and act upon their intended target, and lack of detrimental or toxic effects.

Enormous resources go into the development of new chemicals and it is extremely desirable to find ways to assess these properties at the earliest possible stage so as to minimize the number of expensive failures. In this lecture, several such techniques, developed in the lecturer’s laboratory will be described and demonstrated. These techniques are now in use at various sites of the US FDA for the purpose of assessing the safety of new pharmaceuticals and other chemicals.

WORKSHOPS (Note – these are running on Wednesday 18th October 2006)

in silico technology in drug discovery and development
Michael B. Bolger, Simulations Plus and USC School of Pharmacy

Introduction:
in silico models of biopharmaceutical, pharmacokinetic, and physiological properties related to absorption, distribution, metabolism, excretion, and toxicity (ADMET) have become a valuable tool for reducing the number of experiments that need to be conducted in order to find drug candidates with less chance of failure during development. Application of the estimated properties and/or measured in vitro properties in a physiologically-based gastrointestinal simulation to predict fraction absorbed, bioavailability, and Cp vs. time profiles in discovery, development, pre-clinical, and clinical phases of pharmaceutical development can be considered to be a means of integrating the various calculated and measured properties to enhance decision making.

This workshop is intended to provide the attendees with a lecture and demonstration of in silico tools for drug discovery and development. The workshop will present general information about publicly available sources of data for model building, concepts and schematic outlines of model building methods, and knowledgeable interpretation and analysis of results.

Workshop Outline:
* Introduction to properties and models of interest in early ADMET (log P, log D, pKa, native solubility, solubility in buffer, solubility in bio-relevant media, pH dependence of solubility, effective permeability, apparent permeability, diffusivity, dissolution rate, fraction absorbed, bioavailability, volume of distribution, clearance, plasma protein binding, carcinogenicity, mutagenicity, maximal recommended therapeutic dose, and hERG K+ channel inhibition.
* Overview of QSPR model building methods.
* Calculation of properties for a single molecule and databases of drugs with known fraction absorbed or human effective permeability.
* Integration of biopharmaceutical properties and formulation factors by using gastrointestinal simulation and physiologically-based pharmacokinetics. (parameter sensitivity analysis and virtual clinical trials).

Using Physicochemical Property Predictions to Overcome ADME Concerns at Lead Optimization
Sanji Bhal and David Adams, Advanced Chemistry Development, Inc. (ACD/Labs)

In this workshop we will be focusing on the application of ACD/Structure Design Suite—a software tool that uses a combination of our physicochemical predictors and a critically evaluated substituent database. With this software, the medicinal chemist can quickly evaluate the biological effect of structural modifications, and design a selection of analogs with enhanced physicochemical properties. This software-aided approach allows the chemist to retain the pharmacaphore and proposes a diverse range of substituents to adjust selected parameters such as solubility, logP or pKa. We will be demonstrating the software and discussing its capabilities, as well as the PhysChem property predictors that are an integral part of this module.

We invite attendees to submit ADME-related lead optimization problems they are experiencing in their laboratory due to physicochemical liabilities, such as low solubility, three weeks before the workshop. Results from the software will be discussed for select examples (please indicate the pharmacaphore to be retained and any other limitations to structural modification).

Machine Intelligence in the Design of New Biologically Active Chemicals
Gilles Klopman, Multicase

The workshop will illustrate:
1. The use of MC4PS to evaluate the potential safety of chemicals
2. Demonstration of the creation of a new model to be used to predict the outcome of mouse lymphoma assay of new chemicals.
3. Demonstration of the use of ”In Silico technique” to identify new safe anti-HIV agents.

The Challenges of ADME/Tox Prediction
Bob Clark, Tripos, Inc.

Absorption, distribution and excretion problems with drug candidates have been greatly reduced by the wide-spread adoption of Lipinski’s “Rule of 5” and similar rules of thumb for blood-brain barrier penetration. Problems not identified by such relatively simple filters continue to contribute to disappointing clinical trials for drug candidates from smaller companies, but their importance to large pharmaceutical companies has gone down as relevant in vitro and ex vivo screens have become more widely available and more reliable. As a result, issues involving metabolism and toxicology – with toxicity often being associated with one or more metabolites – have become all the more important. Such concerns are apt to be exacerbated by two trends in drug development: the tendency to look for blockbuster drugs that ameliorate but do not cure chronic disorders; and the recognition that many drugs – particularly inhibitors of regulatory kinases and those targeting mental disorders – must be somewhat promiscuous to be effective.

Broadly speaking, predictive metabotox approaches fall into two classes: those based on molecular similarity and those based on mechanism. In a sense, similarity based methods are non-parametric, in that they make minimal assumptions about how and where problems might occur. Mechanistic models, in contrast, are parametric, in that conformity to the model assumptions is critical to predictive performance and reliability. Each class has generic strengths and weaknesses, particularly as regards scope of applicability and the kind of errors that are likely to dominate performance. These will be discussed in some detail, as will limitations shared by both.

InnovationWell eCheminfo cheminformatics chemoinformatics bioinformatics Computational Chemistry Molecular Modelling Molecular Modeling pharmaceutical pharma meeting workshop conference management executive Bryn Mawr Philadelphia Critical Pathbiomarkers metabolomics toxicology personalised medicine KM Life Sciences Pharma Drug DiscoveryResearch and Development Drug Development Healthcare Innovation Knowledge Management events

Press Release:  http://www.prweb.com/releases/2006/9/prweb443727.htm

Latest Advances in Drug Discovery & Development - Program & Agenda

The final program and agenda for our 4 Day joint InnovationWell and eCheminfo Community of Practice Meeting on the themes of Innovation in Life Science & Healthcare Research & Product Development and Latest Advances in Drug Discovery & Development is now available and provided below.  The meetings will take place at Bryn Mawr College, Philadelphia, USA, 16-19 October 2006. Program brochures may also be downloaded here:

InnovationWell Program Brochure on Innovation in Life Science & Healthcare Research & Product Development (Autumn 2006):

Download innwprogrambrynmawr06final_web3.pdf

eCheminfo Program Brochure on Latest Advances in Drug Discovery & Development (Autumn 2006):

Download eChemProgramBrynMawr06-web1a4.PDF

Updates & Abstracts will be located at: http://www.innovationwell.net/COMTY_conferenceprogram/
and http://www.echeminfo.com/COMTY_conferenceprogram/

PROGRAM AND AGENDA

Monday 16 October

07.30 Registration & Welcome Coffee Opens, Thomas Great Hall

InnovationWell InterAction Meeting Session, 09.00 – 13.00
Utilising Knowledge Management to increase R&D Productivity along Critical Paths
Chair: Michael Liebman (Windber Research Institute)

Delia Y. Wolf (Harvard Medical School), What a Quality Assurance Program can do to Facilitate Clinical Research and Development Process
Peter Gates (Johnson & Johnson PR&D), A Framework for Research Informatics
Jonathan Sheldon (InforSense), Building an Informatics Infrastructure for Translational Research
Duane Shugars (Concentia Digital), Why the “Top-Down Approach” to Knowledge and Content Management has failed the United Sates Intelligence Community – Implications for Healthcare Research
Jian Wang (Biofortis), Knowledge Management for Translational Research
Jeff Spitzner (Rescentris), Applying Knowledge Assessment Techniques to improving Productivity in Life Science Research

eCheminfo InterAction Meeting Session, 08.30 – 13.00
Structure-based Drug Design
Chair: Frank Hollinger (Locus Pharmaceuticals)

Erin Duffy (Rib-X), Structure-Based Drug Design Targeting Infectious Disease
Mike Malamas (Wyeth), Structure-Based Design of Estrogen Receptor-beta Selective Compounds
Frank Hollinger (Locus Pharmaceuticals), Harnessing the power of Structure Based Drug Design using a Fragment Based Approach
Debananda Das (National Cancer Institute), Structural Interactions of CCR5 with HIV-1 entry inhibitors
Max Cummings (Johnson & Johnson PR&D), Small Molecule Inhibitors of Protein-Protein Interactions
Jose Duca (Schering-Plough), Using ab initio calculations as routine tools to help design CDK2 inhibitors
Paul Labute (Chemical Computing Group), High Strain Energies of Bound Ligands

13.00 Lunch

InnovationWell Workshop Activity, 14.00-17.30
14.00-15.30
Carrying out an Onsite Audit and Self-assessment in Clinical Trial Management
Delia Y. Wolf (Harvard Medical School)

16.00-17.30
Knowledge Assessment in R&D – Impact on Project Management & Research Productivity
Barry Hardy (Douglas Connect) & Jeff Spitzner (Rescentris)

eCheminfo Workshop Activity, 14.00 – 18.00
14.00-15.30
Quantum Biochemistry Workflows, Lance Westerhoff (QuantumBio)

Fragment- and Structure-Based Drug Design, Zenon Konteatis and Jennifer L. Ludington (Locus Pharmaceuticals)

16.00-18.00
Advanced Techniques in Pharmacophore Perception and Successful Applications in Drug Design, Osman F. Güner (Turquoise Consulting)

Advances in Virtual Screening and Structure-based Drug Design
Hege Beard and Shashi Rao (Schrodinger)

Hypothesis generation from docking results using activity measurements, interaction fingerprints, clustering and 2D visualization methods
Alex Clark (Chemical Computing Group)

18.00 Drinks & BBQ

Tuesday 17 October

InnovationWell InterAction Meeting Session, 09.00 – 13.00
Decision Support for Research & Development
Peter Henstock (Pfizer), The Role of Systems Biology and Knowledge Management in Advancing Toxicology Knowledge in Big Pharma
Craig Liddell (Realtime Science), Advanced Technology in Support of Analytics in the Life Sciences
David Mosenkis (Spotfire), Case Studies in Using Interactive Visual Analytics to Accelerate Drug Development
Joel Hoffman (Insightful), Management Reporting of Clinical Trial Programs, Portfolios, and Studies: Managing Risks / Managing Projects
Dennis Underwood (Praxeon), Searching for Answers in Drug Development: The Game of Twenty Questions

eCheminfo InterAction Meeting Session, 09.00 – 13.00
Screening
Stan Young (National Institute of Statistical Sciences), Analysis of HTS data using Recursive Partitioning
John Irwin (UCSF), Investigating bias in Docking Screens with Target, Ligand and Decoy Benchmarking Sets
Deepak Bandyopadhyay (Johnson & Johnson PR&D), A new Self-organizing Algorithm for Molecular Alignment and Pharmacophore Development
Daryll Reid (SimBioSys), Virtual Ligand Screening with eHiTS
Neysa Nevins (GlaxoSmithKline Pharmaceuticals), A Critical Assessment of Docking Programs and Scoring Functions
William Douglas Figg (National Cancer Institute), Development of Angiogenesis Inhibitors - from Screening to the Clinic

InnovationWell Workshop Activity, 14.00-17.30
14.00 – 15.30
Electronic Laboratory Notebook Workshops

16.00 – 17.30
Applying Roadmap processes to the Clinical Trials Project Process, Joel Hoffman (Insightful)

Innovation Management in R&D – an Enterprizer Briefing and Case Study, Joseph Bitran (Enterprizer)

Using Interactive Visual Analytics to Accelerate Drug Development, David Mosenkis (Spotfire)

16.00-21.00 Open Event on Knowledge Management in R&D & ELNs

16.00 Electronic Laboratory Notebook Demonstrations

17.30 Open Seminars & Panel Discussion

19.30 Knowledge Café on Knowledge Management in R&D

20.30 Reception

eCheminfo Workshop Activity, 14.00-17.30
14.00 – 15.30
Applications of Filtering and Similarity in Virtual Screening
Paul Hawkins (OpenEye)

Docking and Screening
Darryl Reid (SimBioSys)

16.00 – 17.30
Roundtable Discussion on Virtual Screening & Docking Study
This session will discuss current virtual screening and docking methods and software, results of existing validation and comparison studies, and procedures for community of practice studies to be undertaken.

Wednesday 18 October

InnovationWell InterAction Meeting Session, 08.30 – 16.00
Application of Metabolomics to Drug Discovery & Development
Chair: George G. Harrigan (Monsanto)

George Harrigan (Monsanto), An Overview of Developments in Metabolomics Approaches
Rick Beger (NCTR, FDA), FDA's Critical Path Initiative: Opportunities for Metabolomics
Alvin Berger (Metabolon), Application of Metabolomics to Biomarker and Off-Target Side Effect Identification in Marketed Drugs and New Chemical Entities
Don Robertson (Pfizer), Uses and Abuses of Metabonomics in Pharmaceutical Preclinical Safety Assessment
Gregory Banik (Bio-RAD), Toward Diagnosis of Diabetes by NMR-based Metabolomics
Teresa Garret (Duke University), Identification of novel, minor lipids in total lipid extracts of Eschericia coli using Electrospray ionization mass spectrometry
Nick Haan (BlueGnome), Analysis of Metabolic Profiling Data - Combining the Strengths of NMR and MS
Susan Connor (Glaxo SmithKline), A Pharma Perspective on Metabolomics - the Opportunities and Realities
Laszlo Boros (SIDMAP), Tracer Substrate-based Metabolomics and the 2005 Nobel Prize award in Physiology & Medicine
Bruce Kristal (Cornell University), Serum Markers of Caloric Restriction
Oliver Fiehn (UC Davis Genome Center), Standards in Reporting Initiative
Eric Nemec (Leco Corporation), Studies of Drug-Induced Liver Injury using Comprehensive 2D Gas Chromatography with Time-of-Flight Detection

eCheminfo InterAction Meeting Session, 09.00 – 13.00
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.

Panel: Chris Cooper (BMS), James Arnold (AstraZeneca), Phil Edwards (AstraZeneca), Pete Connolly (Johnson & Johnson PRD), Victor Lobanov (Johnson & Johnson PRD), Jim Wikel (Coalesix)

InnovationWell Workshop Activity, 16.30-18.00
NMR-based Positional Isotopomer Analysis in Metabolomics
Andrew N. Lane (JG Brown Cancer Center, U. Louisville)

Linking Metabolic Profiles to Biological Outcome
S. Stanley Young (National Institute of Statistical Sciences)

Understanding Metabolomics  Mixtures with Principal Components Analysis
Gregory Banik (Bio-RAD)

eCheminfo Workshop Activity, 14.00-17.30
14.00 – 15.30
in silico Technology in Drug Discovery and Development
Michael B. Bolger (Simulations Plus and USC School of Pharmacy)

Using Physicochemical Property Predictions to Overcome ADME Concerns at Lead Optimization
Sanji Bhal and Karim Kassam (ACD/Labs)

16.00 – 17.30
Machine Intelligence in the Design of New Biologically Active Chemicals
Gilles Klopman (Multicase)

Challenges of ADME/Tox Prediction
Bob Clark (Tripos)

18.00 Poster Session, Drinks and BBQ


Thursday 19 October

InnovationWell InterAction Meeting Session, 09.00 – 13.00
Biomarker Discovery & Applications in Drug Development

Keith Elliston (Genstruct), Harnessing the Power of Systems Biology – Delivering Mechanism-of-Action and Biomarkers in Drug Development
Zentam Tsuchihashi (Bristol-MyersSquibb), Many Layers of Biomarker Roles in Tumor Immunotherapy
Darius Dziuda (Central Connecticut State University), Multivariate Biomarkers Discovery
Michael Jones (Novartis), Application of Proteomics to Biomarker Discovery
Bernd Bonnekoh, (Otto-von-Guericke-University), Perspectives for Multi Epitope Ligand Kartography (MELK) for Detection of Diagnostic and Therapeutic Biomarkers in Skin Diseases, Allergology and Beyond

eCheminfo InterAction Meeting Session, 09.00 – 16.00
Predictive Toxicology
Chair: Curt Breneman (Rensselaer Polytechnic Institute)

KEYNOTE: Tudor Oprea (Univ. New Mexico), The Physical basis for the Rule of Five
Navita Mallalieu (Roche Pharmaceuticals), A Roadmap for Integrating Modelling & Simulation in Pre-Clinical DMPK Research
Alex Tropsha (UNC), The Statistical Significance vs. Mechanistic Interpretation of ADME/tox models
Curt Breneman (Rensselaer Polytechnic Institute), Predictive ADME : How Do I Know if my Predictions will be Useful?
Sanji Bhal & Karim Kassam (ACD/Labs), An in silico Approach to Reduce the Burdens of Lead Discovery and Optimization
Michael B. Bolger (Simulations Plus and USC School of Pharmacy), Integration of Early ADME using Property Estimation and PBPK Simulation
Bob Clark (Tripos), The "Structures" in Structure-Activity Relationships
Gilles Klopman (Multicase), Machine Intelligence in the Design of New Biologically Active Chemicals

InnovationWell Workshop Activity, 14.00-16.00
Ansgar J. Pommer (SkinSysTec), Analyzing in-situ Proteomics by Multi Epitope Ligand Kartography (MELK) for Detection of Diagnostic and Therapeutic Biomarkers

eCheminfo cheminformatics meeting workshop conference management science Bryn Mawr Philadelphia Critical Pathbiomarkers metabolomics toxicology virtual screening KM Life Sciences Pharma Drug DesignDrug DiscoveryResearch and Development Drug Development Healthcare Innovation Molecular ModelingComputational ChemistryMedicinal Chemistry events

Press Release:  http://www.prweb.com/releases/2006/9/prweb443727.htm