2010 Accelrys North American User Group Meeting Speaker Abstracts
- Plenary Session
- Pipeline Pilot Collaboration and Deployment Track
- Discovery Stuido Track
- Materials Stuido Track
- Chemistry Track
- Biology Track
- Workshops
Plenary Session
Professor Robert S. Langer
MIT
Novel Biomaterials and Tissue Engineering: New Approaches Using High Throughput Methodologies
We have been studying polymers in both controlled release systems and tissue engineering. Some recent efforts have involved the development of nanotechnology based systems for drug delivery. We are also developing new high throughput approaches for developing and studying materials for use in drug delivery and for controlling human embryonic stem cell differentiation for potential use in tissue engineering.
Jus Singh
Avila Therapeutics
Targeted Therapies in Cancer – the EGFR Story and Rational Drug Design of Covalent Drugs
Medicines that inhibit the EGFR pathway highlight progress made towards targeted therapeutics in cancer. Rational drug design and genotyping of patients who responded to drug treatment led to an important understanding of cancer biology and more focused use of EGFR inhibitors. Unfortunately, this area also serves as an example of drug resistance to targeted therapies and the need for further chemical and computational innovation to overcome these challenges. Rational drug design can deliver targeted, innovative therapies to solve these problems, including in the innovative areas of targeted covalent drugs.
Marian Brodney & Jacquelyn Klug-McLeod
Pfizer
Project-Focused Activity and Knowledge Tracker: A Unified Data Analysis, Collaboration, and Workflow Tool for Medicinal Chemistry Project Teams
Advances in the field of drug discovery have brought an explosion in the quantity of data available to medicinal chemists and other project team members. New strategies and systems are needed to help these scientists to efficiently gather, organize, analyze, annotate, and share data about potential new drug molecules of interest to their project teams. Herein we describe a suite of integrated services and end-user applications that facilitate these activities throughout the medicinal chemistry design cycle. The Automated Data Presentation (ADP) and Virtual Compound Profiler (VCP) processes automate the gathering, organization, and storage of real and virtual molecules, respectively, and associated data. The Project-Focused Activity and Knowledge Tracker (PFAKT) provides a unified data analysis and collaboration environment, enhancing decision-making, improving team communication, and increasing efficiency.
Lori Harmon
Merck
Biological Registration: A Case Study for Pre-Competitive Alliances
Pressures in the pharmaceutical industry are driving traditional competitors to new ways of working together. Pre-competitive alliances can decrease overall cost of investment and risk, while yielding benefits for all. In 2007, Accelrys, Merck, and Abbott formed a Special Interest Group, with the goal of creating of a flexible commercial application for registering and tracking the diverse set of biological samples used as both reagents and therapeutic agents, such as antibodies, proteins, vaccines, cell lines, plasmids, and RNAi's. The successes and challenges of this Special Interest Group model will be discussed, in the context of Biological Registration.
Anne Chaka
National Institute of Standards and Technology
Probing Mechanisms at the Nanoscale – from Gold Nanowires to Environmental Pollution
Density functional theory is invaluable for probing mechanisms in complicated systems at a resolution not obtainable by experiment. In this presentation we will highlight two recent successes using DMol3 to (1) determine the structural and electronic transitions during elongation of gold nanowires, and (2) understand the reactivity of metal oxide nanoparticles in the environment that play a key role in the transport of heavy metal pollutants.
(1) Gold nanowires undergoing tensile deformation can exhibit different properties than those at equilibrium. As tensile deformation is conducted under a wide range of conditions, large structural changes are observed, resulting in the formation of locally ordered intermediate structures with interesting electronic properties. A rich diversity of deformation pathways is also uncovered, that converge to only two final local configurations with reproducible breaking strengths, in agreement with experimental results.
(2) The complex nature of environmental interfaces mandates a thoughtful and layered approach to modeling. Interfaces between water and oxide surfaces play key roles in many applications and natural phenomena, but surfaces that exist under UHV conditions often do not persist under operational or hydrated conditions. The first step in investigating environmentally relevant surfaces is thus to employ ab initio thermodynamics to solve for the lowest-energy hydrated surface structures under relevant conditions. Subsequent study of heavy metal adsorption such as Pb(II) requires consideration of both surface adsorption sites and proton displacement patterns. Finally, delineation of structure-property relationships is achieved by systematic analysis of trends due to oxide composition, surface structure, exposed oxygen functional groups, surface hydrogen bonding, adsorption energies, directional Pb-O overlap, local and long-range adsorption-induced surface relaxations, Pb-cation repulsion, and the role of the partially filled hematite d-band.
“Elongation and breaking mechanisms of gold nanowires under a wide range of tensile conditions”, F. Tavazza, L.E. Levine, and A.M. Chaka, J. Appl. Phys. 106, 043522 (2009).
“Pb(II) Adsorption on Isostructural Hydrated Alumina and Hematite (0001) Surfaces: A DFT Study”, Sara E. Mason, Christopher R. Iceman, Kunaljeet S. Tanwar, Thomas P. Trainor, and Anne M. Chaka. J. Phys. Chem. C 113 (6), 2159-2170 (2009).
Pipeline Pilot Collaboration and Deployment Track
Jennifer Heymont
Eisai
Component Recursion: Implementation of Recursive WorkflowsUsing Components That Call Themselves
Recursion is a classic algorithmic technique that is invaluable in many areas, including workflows that deal with trees or other hierarchical data. Many recursive techniques that are applied to a single data record can be implemented simply in PilotScript. However, when the workflow requires the entire datastream to be processed recursively it is not possible to simply create a loop in the pipeline in order to call the recursive process. Examples of this kind of recursive workflow are discussed, with a demonstration of how we have created components that call themselves in order to create recursion on the entire datastream and implement these workflows.
Conrad Agramont
Accelrys
Cloud Computing Directions
During this session we’ll share our research and development projects surrounding Cloud Computing at Accelrys. We’ll also provide an overview of Cloud architectures, benefits, and concerns throughout the scientific community.
Christopher Smith
Pfizer
EHS Virtual Workspace that provides Scientific Knowledge Management and Operational Excellence
The development of an Environmental Health and Safety virtual work space has been accomplished by Pfizer, Accelrys and RedShift Technologies. This virtual workspace allows environmental scientists to dynamically store, curate, analyze and report on environment fate and effects data which in turn allows for electronic risk assessments which in turn supports Intelligent Testing Strategies. The resulting virtual workspace is utilizes the compute and reporting power of Pipeline Pilot , the storage intelligence of RedShift Technologies and the presentation of Microsoft SharePoint. With this dynamic best of breed combination Pfizer will be able to provide:
- Risk reduction – Improve the understanding of a chemicals risk potential through an integrated profile approach that incorporates data from multiple public and proprietary sources.
- Efficacy – Identify key data gaps that would better inform risk assessment and minimize the incidence of unnecessary and/or duplicative testing.
- Quality – Centralized data review, interpretation, and entry to ensure consistency and accuracy.
And in the near future, once data is populated in quantity and quality, exploratory analysis and Predictive and Suggestive data modeling will occur which will drive the reduction in laboratory testing and inadequate disposal methods.
Seung-Hoon Choi
Insilicotech
How to Manage Scientific Data in a Corporate R&D Center
Research Information Management System (iRIMS) is a Pipeline Pilot-based web solution designed for the manipulation and analysis of data produced by enterprise R&D centers, especially focused on the chemistry, material, electronic field. There are several components in iRIMS including Material Management System (MMS), Assay Management System (AMS), Workflow Management System (WMS), Formulation Management System (FMS) and external 3rd party Design of Experiment (DOE) solution. Among these components, WMS is the key element that can be used to draw the workflow of a R&D process with the help of Pipeline Pilot. WMS workflow is composed of several consecutive/split/ merged unit processes correlated to the actual R&D and manufacturing procedure. All kinds of R&D data from each unit process can be archived with detailed information such as input and output materials, process conditions and parameters, literature information and performance of that process. AMS is a simplified version of WMS for simple one unit process with no need to use workflow, for instance, efficacy data management during a drug discovery process. Both WMS and AMS can be linked with MMS; the major functions of MMS are the registration of material information and the management of inventory with or without chemical structures. Batch Registration Interface, in conjunction with DOE solution, allows users to register many experimental data in the FMS database simultaneously and to optimize the conditions and parameters as well as compositions and materials for the enhancement of system performance. In summary, iRIMS is a solution developed to preserve the R&D data together with the actual workflow and perform the trend analysis on the stored R&D asset. With the aid of iRIMS, users are expected to be able to make smarter decisions during a R&D process and, as a result, reduce lead time and development cost.
Andrew LeBeau
Accelrys
Updates for the Pro Client and Reporting Collection
Pipeline Pilot 8.0 sees significant changes to the Pipeline Pilot Professional Client and the Reporting Collection. This presentation will highlight key enhancements such as, for the Pro Client, Design Mode, Watch Windows to get a better view of your data along the pipeline, an advanced component/protocol search tool, and protocol profiling to find bottlenecks in your protocols. For the Reporting Collection, we cover Styles, several new components, and much more interactivity in your reports.
Integration of Pipeline Pilot and HEOS®, a SaaS Platform, for Drug Discovery
HEOS® is a SaaS software platform for comprehensive drug discovery information management that facilitates the collaboration of geographically distributed project scientists. It allows complex data (biology, chemistry, analytical, safety, PK, etc.) to be exchanged, stored and shared rapidly and securely across multiple sites worldwide. HEOS® has been seamlessly integrated with Pipeline Pilot to allow partners to perform advanced data analysis and to provide decision tools to scientist and management. We will include in our presentation a study case of how HEOS® has quickly become the standard knowledge-sharing resource in the area of neglected diseases research.
Pipeline Pilot as a Rapid Prototyping Tool for Agility Support: Competitor Drug Information Analysis for Gene Target and News Aggregation and Delivery using RSS feeds
Business drives the need to adapt rapidly and cost effectively in response to changes in customer demands and the business environment. Prototyping as a design deliverable is on the rise. For proof-of-concept purposes, prototyping plays an important role in facilitating the communication between users and developers, in responding to the fast changes in business requirements and in continuously delivering added business value.
As part of the agility support team, Pipeline Pilot is a tool utilized for iterative agile development of prototypes and effective communication, focusing on delivery of core business functionality to Roche’s Pharma Research & Early Development Informatics (pREDi) teams.
In this presentation, two prototype applications developed utilizing Pipeline Pilot will be demonstrated. Both prototypes utilize Pipeline Pilot forms, dashboards and general reporting capability. The Competitor Information Analysis Tool enables researchers/project managers to review, assess and analyze competitor drug pipeline data/information based on gene searches. The UI, consolidated tables, statistics and visualization charts generated via Pipeline Pilot help to identify key competitors and associated information that can facilitate decision-making in regards to current or future projects.
The user-friendly News Aggregation, Posting and Reporting tool utilizes RSS feeds to help Information Scientists automate pre-selection and the posting of relevant and up-to-date scientific news for distribution to Roche internal end-users, including scientists and information support groups.
Discovery Studio Track
Francisco Hernandez-Guzman
Accelrys
Best practices in the design of Antibody homology models
As the pharmaceutical and biotech industries continue to invest in antibodies as alternative therapeutic agents to classical small molecule therapy, the need for structural information has become instrumental towards understanding molecular interactions, as well as various biophysical properties. By leveraging the high homology within the antibody family, one can successfully build reliable models with a high degree of molecular detail. In this presentation, we will explore the process of building a homology for a Fab domain. We will highlight a process that has shown to give high quality models and we will discuss some of the challenges that one has to watch for during the model building process. We will also emphasize methodologies used to refine the Complimentary Determining Regions (CDRs) of the antibody, and in particularly explore some approaches that can be used to properly model the challenging H3 loop.
Sean Ekins
Collaborative Drug Discovery
A Collaborative Database And Computational Models For Tuberculosis Drug Discovery
The search for molecules with activity against Mycobacterium tuberculosis (Mtb) is employing many approaches in parallel including high throughput screening and computational methods. We have developed a database (CDD TB) to capture public and private Mtb data while enabling data mining and collaborations with other researchers. We have used the public data along with several cheminformatics approaches (in Discovery Studio) to produce models that describe active and inactive compounds. We have compared these datasets to those for known FDA approved drugs and between Mtb active and inactive compounds. The distribution of polar surface area and pKa of active compounds was found to be a statistically significant determinant of activity against Mtb. Hydrophobicity was not always statistically significant. Bayesian classification models for 220,463 molecules were generated and tested with external molecules, and enabled the discrimination of active or inactive substructures from other datasets in the CDD TB. Computational pharmacophores based on known Mtb drugs were able to map to and retrieve a small subset of some of the Mtb datasets, including a high percentage of Mtb actives. The combination of the database, dataset analysis, Bayesian and pharmacophore models provides new insights into molecular properties and features that are determinants of activity in whole cells. This study provides novel insights into the key 1D molecular descriptors, 2D chemical substructures and 3D pharmacophores which can be used to mine the chemistry space, prioritizing those molecules with a higher probability of activity against Mtb.
Lingling Shen
Harvard University
Estimation of Drug Thermostability in RNAi Therapeutics
For RNAi therapeutics, understanding the thermodynamic properties of the oligonucleotides can lead to a more stable and improved drug. For unmodified siRNA and AON, the use of nearest neighbor methods can be used to fast and accurately predict its thermodynamic properties. But when chemically modified nucleotides need to be considered, the problem is much more challenging with no accurate methods available. Here we describe a MD method through Pipeline Pilot with charmm27 component in Discovery Studio2.5 to predict the melting temperature for both unmodified and chemically-modified siRNA and AON. Simulation results show the similar performance as current nearest neighbor methods restricted to unmodified oligonucleotides. Moreover it has a broader application allowing incorporation of chemical modifications and detailed structure profiling.
Modeling Protein Degradation Processes and the Development of Rational Approaches to Stabilization and Protein-Cosolute Interactions
We describe a new strategic approach to the formulation and stabilization of biotherapeutics. The approach is based on applying both molecular and macroscopic modeling tools in order to gain an understanding of degradation processes with unprecedented detail and accuracy. The microscopic modeling approach can be used to visualize key regions of proteins, including antibodies, that lead to degradation or are otherwise important for a variety of technological applications. Such applications include screening molecules for developability and manufacturability during discovery, identifying key sites that are responsible for degradation for the purpose of removing them, identifying sites for conjugation of payloads, and identifying binding regions. The macroscopic modeling approach can be used to screen formulations and generally understand better degradation phenomena. We also show how molecular-level simulations can lead to an understanding of protein-cosolute interactions with unprecedented detail and therein the rational design of new formulations. Our group works on a variety of degradation processes, such as aggregation, oxidation, deamidation, and hydrolysis. We combine modeling with a experimental approaches, always with the objective of aiding in the development of biopharmaceuticals. Emphasis for this talk is on protein/antibody aggregation.
Iwona Weidlich
NIH
Inhibitors of Human Tyrosyl-DNA Phosphodiesterase Developed by Virtual Screening Using Ligand-Based Pharmacophores
Human tyrosyl-DNA phosphodiesterase (hTdp1) inhibitors have become a major area of drug research and structure-based design since they have been shown to work synergistically and selectively in cancer cells.1 The pharmacophore features of fourteen hTdp1 inhibitors were used as a filter to screen the ChemNavigator iResearch Library of about 27 million purchasable samples. Docking of the inhibitors and hits obtained from virtual screening was performed into a structural model of hTdp1 based on a high resolution X-ray crystal structure of human Tdp1 in complex with vanadate, DNA and a human topoisomerase I (TopI)-derived peptide (PDB code: 1NOP). A total of 46 compounds matching the three-dimensional arrangement of the pharmacophoric features were assayed. Using a high-throughput screening assay (HTS),2 we have identified an 1H-Indol-3-yl-acetic acid derivative as a potent Tdp1 inhibitor with an IC50 value of 7.94 µM.3 The obtained novel chemotype may provide a new scaffold for developing inhibitors of Tdp1.
A qHTS discovery process using more than 300,000 screening samples has been started for this target at NIH to generate additional small molecule lead compounds and better understanding of this system. This effort, a trans-institute collaboration between the Division of Cancer Treatment and Diagnosis (DCTD) and the Center for Cancer Research (CCR) at the NCI with the NIH Chemical Genomics Center (NCGC) and other NIH departments in the context of the newly formed Chemical Biology Consortium at the NCI.4 We will discuss the nature of the activity in the assay, as well as our docking results for a second, large set of ~300,000 compounds assayed in HTS.
The presentation is the results of research supported, in whole or in part, by direct costs funded by NIH.
Campbell McInnes
University of South Carolina
Structural and functional analysis of CDK4/cyclin D binding to p27 and substrate competitive inhibitors
An alternative strategy for inhibition of the cyclin dependent kinases in anti-tumor drug discovery is afforded through the substrate recruitment site on the cyclin positive regulatory subunit. While highly potent peptide and small molecule inhibitors of CDK2/cyclin A, E substrate recruitment have been reported, little information has been generated on the determinants of inhibitor binding to the cyclin groove of the CDK4/cyclin D1 complex. CDK4/cyclin D is a validated anti-cancer drug target and is continues to be widely pursued in the development of new therapeutics based on cell cycle blockade. Peptidic inhibitors of CDK4/cyclin D of pRb phosphorylation have been synthesized, and their complexes with CDK4/cyclin D1 crystal structures have been generated. Based on available structural information, comparisons of the cyclin grooves of cyclin A2 and D1 are presented and provide insights in the determinants for peptide binding and the basis for differential binding and inhibition. In addition, a complex structure has been generated in order to model the interactions of the CDKI, p27KIP1, with cyclin D1. This information has been used to identify unique aspects of cyclin D1, that have a significant impact on peptide interaction, and which can be exploited in the design of cyclin groove based CDK inhibitors. Peptidic and non-peptidic compounds have been synthesized in order to explore structure-activity relationship for binding to the cyclin D1 groove which to date has not been carried out in a systematic fashion. Such compounds will be useful chemical biology probes to determine the cellular and anti-tumor effects of CDK inhibitors that are cell cycle specific and do not inhibit the transcriptional regulatory effects of other cyclin dependent kinases.
Materials Studio Track
Stephen Todd
Accelrys
Recent Advances in Materials Science Modelling
This presentation will cover an overview of some of the on-going development efforts in Materials Studio, focusing on the quantum mechanics, classical simulation, and visualization tools. Recently, there has also been effort in delivering some of the key Materials Studio functionality such as CASTEP, DMol3, Amorphous Cell, and Forcite Plus within Pipeline Pilot. This presents the opportunity for users to improve the efficiency of their resource usage, create workflows using a simple graphical environment, and use reporting tools to share results with other members of your team. An overview of the new Materials Studio Collection will be presented with a focus on some of the new property prediction workflows that are provided with the collection.
Istvan Halasz
PQ Corp
Experimental and Theoretical IR and Raman Spectra of Silicates
Silica is the most copious mineral in the Earth’s crust. No wonder, that humans have utilized it since prehistoric times. Today, in addition to the obvious glass, ceramic, and construction material applications, a massive amount of silica is consumed in such diverse industrial processes like coal mining, crude oil drilling and manufacturing of paper, rubber, paint, beer, detergents, toothpaste, cosmetics, medical supplies, and many more. Silica is also a mainstay in catalysis which affects about 90% of all chemical processes. The majority of these applications use aqueous alkaline silicate solutions, one of the largest volume synthetic chemicals in the world.
Despite their importance, little is known about the molecular structure of dissolved silicates and even less about the transformation of these molecules into amorphous gels although gelling is important in almost every utilization. FTIR and Raman spectroscopy can quickly test the molecular structures of both solids and liquids thus uniquely suited for studying this solidification process in situ. Vibrations of Si-O bonds have been subjects of exhaustive theoretical studies. IR and Raman spectra have been routinely used for distinguishing 3, 4, 5, etc. member rings in zeolites and Q0, Q1,…..Q4 connectivities of [SiO4] tetrahedra in glasses for decades. However, these structural assignments to characteristic vibrational bands could not be employed to the dissolved or gelled silicates without contradictions presumably because of interactions with the water molecules present.
To be able to deduct the molecular structure of aqueous silicates from their vibrational spectra and explain some unpredicted experimental observations we turned to computational chemistry. Here we show in same examples that carefully designed models can resemble quite well the real FTIR and Raman spectra. To establish appropriate model conditions for more complex siloxane bonds in aqueous environment, we calculated first spectra of some of the very few materials which contain almost exclusively monomer [SiO4] units verified by independent XRD, Si29 NMR, and molecular weight measurements. We prove that water can indeed affect the characteristic Si-O vibrations. Examples will be shown for vibrational spectra from both semi-empirical and DFT calculations using VAMP, DMOL3, and CASTEP modules from the Accelrys’ Materials Studio program package with different basis sets. All calculations were performed on workstations with 1, 16, and 32 processors.
Kelly L Anderson
Proctor & Gamble
On the Possibilities for Materials Component Collection in Consumer Packaged Goods R&D
With dramatic reductions in the cost for high performance computing (cost/GFLOP < $0.15 ) and the recent work to expose Materials Studio functionality to Pipeline Pilot, one can start to imagine the potential for screening molecules or material systems using grid technology and pipeline pilot. Having Mesocite’s coarse-graining and dissipative particle dynamics methods exposed to workflow management software leads to potential to coarse- and fine- graining of molecules or systems on-the-fly or to at least further refine the parameterization of highly accurate and transferable coarse-grained parameters. The grid capabilities in Pipeline Pilot allow large collections of condensed phase simulations to be setup and managed from a number of possible interfaces. However, there is still much science to do. Atomistic forcefields are still missing a large number of parameters or are sometimes inaccurate, and we are often asked to simulate the intractable. What can we do to make the intractable feasible and deployable?
Stewart Clark
The Castep Group
Recent developments in Castep: New functionality and speed enhancements
The Castep program within the Materials Studio package is a first principles quantum mechanical code for performing electronic structure calculations. Within the density functional formalism it can
be used to simulate a wide range of materials including crystalline
solids, surfaces, molecules, liquids and amorphous materials; the
properties of any material that can be thought of as an assembly of
nuclei and electrons can be calculated with the only limitation being
the finite speed and memory of the computers being used. Castep uses
this to simulate many properties of such materials. In this talk I
will give a summary of recent developments of Castep's functionality
focusing on both computational speed and new functionality. This
includes on the calculation of experimental spectroscopic results such
as infra-red and Raman spectroscopy (phonons) using first-principles DFT
methods.
Alex Goldberg, Nick Reynolds
Accelrys
Density Functional Theory Study on Al Nanostructures
Al clusters have been a focus of several studies because they exhibit unique energetic and structural properties. Al13 cluster exists in several isomeric configurations. It has been shown both theoretically and experimentally that the icosahedral structure is the most stable isomer. It has been suggested as a good candidate for hydrogen absorption with potential application in hydrogen storage. It has been theoretically demonstrated that a single hydrogen atom adsorbs on the surface of Al13 clusters without crossing a potential barrier. Three stable minima have been reported with hydrogen atom adsorbed at atop, bridge and hollow positions. We also report that the most stable structure of
Al13H2 cluster has one hydrogen atom on the vertex with the other on the edge of the icosahedron. This finding suggests that potentially many different combinations of adsorbed hydrogen atom positions can occur to cover all available sites. Using the DFT calculations we have found a stable
Al13H12 cluster with hydrogen atoms bonded to all vertex Al atoms.
It appeared that Al12 cluster with icosahedral symmetry, an atomic “cage” obtained by removing the central Al ion from the most stable Al13 cluster can absorb even more hydrogen atoms. The Al12H20 cluster is of particular interest for hydrogen storage application as it has a relatively high hydrogen storage capacity. Comparing structures and properties of bare and hydrogenated Al12 cages with the corresponding Al13 clusters we found that in contrast to Al13H12 cluster the Al12H12 expands upon hydrogen adsorption on its surface. This expansion can be attributed to the missing central Al atom. Due to the cluster expansion there is more surface area available for hydrogen atoms resulting in a weakly bound but nevertheless rather stable, Al12H20 cluster.
Jian-Jie Liang
Accelrys
Fitting vdW Parameters in Forcefield Development by a Combined GA/GFA Method Implemented through a Pipeline Pilot Workflow
Nature presents us with various ways that weak forces such as van der Waals (vdW) interactions have dramatic influence on macroscopic properties. Examples include organic-inorganic composites as in abalone shells or human tooth enamel that impart mechanical strengths orders of magnitudes higher than the constituent materials alone. Harnessing such weak interactions in nano-engineering has been challenging, and atomistic modeling is perceived to play important role towards distinguishing the underlying physical principles of such processes.
Central to the atomistic modeling of the vdW interactions is the parameterization of the interaction terms for each atomic species in a target system. Such parameterization has still been largely manual and subjective. Development of high-quality parameters had been lagging behind the comparably more rapid development in new materials design.
Preliminary work has shown that statistical model-building through the Neural Network (NN) method as available in Materials Studio can substantially automate the process. Our most recent work further demonstrated that the process can be made substantially more efficient and less subjective through a combined Genetic Algorithm (GA) and Genetic Function Approximation (GFA) method. Individual vdW parameters are sampled independently against thermodynamic observables (such as Cohesive Energy Density, CED). A GFA is fitted to reproduce the observables as functions of the vdW parameters. The GFA models are used, along with the thermodynamic observables, to determine the optimum combination of these VDW parameters via an evolutionary (or genetic) algorithm (GA). Currently, the GFA model building and the GA optimization processes have been implemented into a Pipeline Pilot protocol.
Chemistry Track
Jacquelyn Klug-McLeod
Pfizer
Development of a Mechanism-Based Screening Library
To date most screening within the pharmaceutical industry has relied on chemical diversity or localized subsets of compounds with known pharmacology. There has been limited effort to generate screening libraries using the vast amount of pharmacological data as the source, to pick the mode of action and then ensure the compound is selective. Pfizer’s Mechanism-Based Screening Library currently covers 722 targets representing 3155 compounds. The library allows the delineation of which mechanisms may be playing a role in an observed phenotypic functional endpoint, assisting in the discovery of new mechanisms of interest and ultimately new innovative drug discovery projects. This talk will describe Pfizer’s efforts to automatically triage internal and external biological data using Accelrys’ Pipeline Pilot to identify the most potent/selective compounds known for each mechanism and select those available as solids for screening.
Jim Metz
Abbott
Using Protein Ligand Interaction Networks to Navigate Kinome Polypharmacology
Polypharmacology of kinase inhibitors has recently been recognized as an important and clinically useful aspect of the design of drugs for this family of proteins. We are using Pipeline Pilot to create tools to elucidate and facilitate the rapid analysis of kinome pharmacological networks based on ligand profiling data from the Abbott kinome panel. We are also developing a deeper understanding of various similarity metrics as well as the interplay between information content, assay quality, experimental noise, and the robustness of ligand-target and target-target networks. I will present various aspects of this research including the creation of a novel network similarity metric and visualization tools for pharmacological networks.
Dana Honeycutt
Accelrys
Mad about MAD and More
This talk will cover some of the many enhancements made to Pipeline Pilot's statistical capabilities for the 8.0 release. These include new components and capabilities that allow you to treat both native and R models equivalently, simplifying comparisons when trying to determine the best type of model for a given problem. I will focus on new measures of a statistical model's applicability domain (MAD) and how we can use these measures to assess (and sometimes improve) the expected accuracy when making predictions. I will discuss new options for computing distances (dissimilarities) between samples for clustering, similarity searching, and applicability domain calculations. These options give you greater flexibility in determining how distances are computed, including the option to allow (and correct for) missing data.
Rob Brown & Neil Eccles
Accelrys
Recent Advances and Future Plans in Chemistry and Cheminformatics
This talk will review the recent major developments in the Chemistry, ADMET, Plate Analytics and Cheminformatics Collections in Pipeline Pilot 7.5 and 8.0 including
- Pipeline Pilot Chemistry Cartridge and Components
- Functionality and useability enhancements to the Database Search client
- New Bioisoster Enumeration components
- Improved reaction handling
- Improved structure normalization and clean-up
- New re-learnable Tox models based on the TOPKAT data set
We will also discuss the future development plans for these collections.
Pierre Ducrot
Discngine
Use of data graph to navigate through complex data in Pipeline Pilot based scientific applications
Nowadays, extracting, navigating and visualizing data content has become a major task for chemoinformaticians supporting scientists. Most data points are related to each other via common assays, targets, fragments, scaffolds, etc. Hence, extracting context dependent and accurate information that would help scientist to identify new ideas or validate hypothesis without overwhelming them with data may be a difficult talks.
Similarly to hyperlinks which have been a revolution in the way people read documents, data connections have proven very efficient in multi objective data analysis. In this presentation we will demonstrate how data graphs can be used in conjunction with Pipeline Pilot reporting collection to extract and display relevant information from complex data structures in the context of structure-biological data relationship analysis.
Biology Track
Nancy Latimer
Acclerys
Big Bytes for Pipeline Pilot
With the dramatic advances in sequencing technology, we are approaching a fundamental shift in how we do science. It may soon become financially feasible to completely sequence the genomes or targeted chromosomal regions for all patients in a clinical trial! Successful users of deep sequencing platforms often have in-house bioinformatics expertise and locally customizable software. This is typically not included with the platform vendor’s hardware. The informatics challenge is making sense of all the data, especially as not all organizations have in-house expertise or the budget to create customized tools. However, many workflows are routine and predictable and, as such, are amenable to automated analyses. Using a data-pipelining approach we demonstrate several common workflows that start after base calling and continue through variation analysis. The examples are implemented using Pipeline Pilot’s Next Generation Sequencing Collection (NGS) in beta.
Eric Scott
Roche
Tracking H1N1 Influenza Virus Sequence Mutations Around the World
The 2009 H1N1 Influenza pandemic has been a major concern for governments and health organizations worldwide. The first swine-origin triple re-assorted influenza cases were reported in humans in March 2009, and it was declared a pandemic in June. As of September, the World Health Organization (WHO) has reported more than 3200 deaths associated with Pandemic H1N1. There are several web-based systems monitoring and tracking spread of the disease and incidence of deaths over time around the world. Concurrently, there has also been a significant increase in the number of viral sequences submitted to the Influenza Virus Resource at NCBI. Tracking viral sequences and polymorphisms is especially important for monitoring possible development of resistance to anti-viral drugs that have been effective against the virus. We developed a web-based influenza sequence tracker that is updated daily and provides an automated method for monitoring mutations in viral sequences. Sequence analysis results and detailed statistics are reported separately for the pandemic and seasonal H1N1 and other influenza strains. This has been especially useful to virologists and the influenza team.
Donavan Cheng
Roche
Automated in silico drug target discovery via a Pipeline Pilot platform for meta-analysis of gene expression data
Bioinformatics analyses of genomic scale data are emerging as key pillars of target discovery and biomarker identification efforts in Pharma Research. In mining microarray data, one shortcoming has been that individual datasets are often underpowered, with sample sizes too small to allow adequate differentiation of coordinated gene expression from noise. To answer this issue, efforts at Roche have been directed at integrating microarray data from multiple sources within a meta-analysis framework, focusing on gene signature identification for disease indications e.g. in oncology and inflammation fields. Meta-analysis methods offer the advantage of more accurate and sensitive detection of gene expression signatures due to the increased statistical power from combining multiple datasets.
Pipeline Pilot presents an opportunity for rapid platform prototyping and standardized module development. Herein, we describe a Pipeline Pilot workflow that implements our meta-analysis framework for integrating microarray data. First, microarray datasets matching a disease indication of interest are extracted from public repositories, e.g. GEO, ArrayExpress, and analyzed using standard methods (RMA, ANOVA etc.) to identify differentially expressed genes. Two meta-analysis methods are used to combine data: a nonparametric rank-based method and an inverse-variance weighted effect size method. Both methods assign high scores to genes that derive broad support across multiple datasets with large median fold changes that are consistent in their direction of induction (up vs. down) across studies. Biological interpretation of significant gene expression markers in terms of enriched pathway and functional annotations is further enabled by querying against Gene Ontology, GSEA, BioCarta and KEGG databases. Using breast cancer as an example, we show significant overlap between genes identified via our meta-analysis platform and known breast cancer targets, thereby validating our approach. Importantly, implementation of this workflow in Pipeline Pilot facilitates quick dissemination and reusability of our platform, engendering impact across multiple target discovery efforts.
Stephen Schurer
University of Miami
Kinome‐wide predictive modeling from diverse high‐quality data sets
Kinases are now one of the most intensely pursued classes of drug targets with approximately 30 distinct kinase targets and over 200 small molecule inhibitors in clinical development. The Kinase Knowledge Base (KKB, http://kinasedata.com/) covers a large body of Kinase SAR data that has been generated in numerous research programs and spans the entire human kinome. Pipeline Pilot makes it easy to aggregate and standardize data sets and to generate hundreds of kinase predictors employing different machine learning techniques. Because the data sets are curated from several thousand peerreviewed journal and patent publications from numerous laboratories comprising various assay technologies, it is interesting to investigate to what extend such data sets can be used to generate predictive models, which modeling technique(s) are best suited, and to characterize data sets that lead to good predictors.
Arvind ThirumalaiswamySekhar
Dow Chemical Company
Automated Allergenicity Assessment of proteins using Pipeline Pilot
Gene products are assessed for potential allergenicity prior to commercialization. The industry agreed in-silico assessment involves the use of multiple bioinformatic software applications that collectively help predict potential allergenicity of a product. Each application has its own data input and output formats and the complexity of the applications necessitate a dedicated resource to run the applications on the sequence data and complete the assessment. The applications themselves are not user friendly and pose a bottleneck to the gene discovery pipeline.
We have successfully automated the allergenicity assessment workflow using Pipeline Pilot as the platform and making remote calls to the bioinformatic applications from the workflow thus abstracting the underlying complexities of the application themselves. The workflow also enables Dow AgroSciences to simultaneously assess multiple protein products in a single run. As a result we have removed the bottleneck from our gene discovery pipeline and reduced the turn-around-time for assessments from days to hours.
MA Mancini, FJ Ashcroft
Baylor College of Medicine
A systems biology approach : Multi-parametric high content analysis of estrogen receptor functions reveals unique biological response fingerprints
Workshops
IT Best Practices
The workshop will focus on best practices related to PP administration and development. We encourage the audience to discuss best practices in their organization.
Topics to include: Component Design, Protocol Design, Application Packaging, Production Deployment, Server Upgrades, Migration and Disaster Recovery, Protocol Regression System, Server Sizing and OS Selection, Server Configuration, Development, Integration, SharePoint Integration, Deploying Web Port Beyond the Corporate Firewall, Database Integration & Security Configuration.
Target Audience: PP administrators, PP developers
Biology Workshop
The first half of the workshop will be an introductory exploration of the Pipeline Pilot Biology Collections: Perform advanced analyses on your ‘omics data using the robust collections of algorithms and best-practice protocols offered within the Mass Spectrometry for Proteomics, Gene Expression, Sequence Analysis, and Next Generation Sequencing Component Collections.
The second half of the workshop will be a hands on exploration of new tools for automated Image Informatics. Emphasis on deign mode an interactive learning to create and automate workflows to segment, label and annotate and measure images and objects
Target Audience: Biologist Directors of Research and IT
Best Practices in Protocol Development
This workshop will be split into two parts (across the coffee break). Part 1 will focus on basic good practices for protocol development. Following the break, more advanced topics will be discussed. Topics may include:
- Understanding data pipelining and designing protocols to make best use of it
- Basic protocol building with Design Mode (new feature in Pipeline Pilot 8.0 – excellent opportunity to show it off)
- Merging, joining, and caching: Optimizing computationally/memory-intensive operations
- Refactoring protocols: Reviewing completed protocols to see where efficiencies can be obtained
- Help with Reporting protocols: Understanding containers and other general concepts
- Interactive reporting: Good practices when linking protocols
Target audience: The first part of the workshop is intended primarily for relatively new users of Pipeline Pilot, and for those who are interested in learning more about new Pro client features in Pipeline Pilot 8.0. The second part of the workshop will feature more advanced topics, that would particularly suit intermediate-level protocol developers.
Discovery Studio
The goal of this workshop will be to have a live walk through of how Discovery Studio can be used to allow you to gather as much information out of your favorite structure. We will explore the various elements that are found on a typical PDB file and the steps used in the cleanup process that one has to go through prior to making any structure based calculation. We will learn how to calculate the interaction energy between a protein-ligand complex, as well as other various analysis tools that can help us gain a better understanding of the various relationships within a molecular system.
Target Audience: New Discovery Studio users seeking to do structure based calculations and Current DS users who would like to reaffirm their skills and explore best practices in structure analysis.
Materials Studio & Materials Studio Collection Workshop
The workshop will develop strategies for investigating automating calculations in Materials Studio. We will use the calculation of solubility parameters as an example, specifically the polymer cohesive energy density.
In the first part of the workshop, workflows based on Materials Studio will be developed, including the use of Materials Scripting. In the second part, the implementation will be extended to use the new Materials Studio Collection in Pipeline Pilot, highlighting further automation, deployment and reporting capabilities.
The strategies and techniques that participants learn will be applicable to broad classes of materials including organic molecules, polymers, catalysts, and semiconductors.
Target Audience:
Current users of Accelrys Materials Studio
*Please note that titles and abstracts are subject to change.
If you have any questions regarding the event please email Byron Konold.
