7:00-8:00 Conference Registration and Morning Coffee

8:00-8:15 Welcoming Remarks from Conference Director
Julia Boguslavsky, Cambridge Healthtech Institute

8:15-8:20 Introduction from Executive
Monette Greenway, Ph.D., Vice President, General Manager, Laboratory Automation and Cellular Imaging, Thermo Fisher Scientific

8:20-9:20 Panel Discussion with End-Users and Vendors

9:20-10:20 Coffee Break with Exhibit and Poster Viewing

HCA for Toxicology

10:20-10:45 in vitro and Clinical Pathology Application of HCA Assays for Prediction, Mechanistic Understanding, and Biomarkers of Toxicity
Peter J. O’Brien, D.V.M., Ph.D., Veterinary Clinical Pathologist, Pathology, University College Dublin, Ireland
The high effectiveness of HCA in testing for drug-induced, sublethal cytotoxicity is demonstrated. Application is illustrated in different cell-based models, including circulating blood cells for clinical pathology and various in vitro human cell lines for predictive toxicology. Such diverse applications of HCA provide some translational safety biomarkers with potential to track toxicity through drug discovery and development.  HCA is also demonstrated to be useful in elucidating subcellular mechanisms of pathogenesis of drug toxicities. Critical features of cell models and effective diagnostic parameters are identified for HCA cytotoxicity testing.  

10:45-11:00  Sponsored Presentation                    
A Portfolio of CSB™ Toxicity Panels for Discovery Tox Applications
Deb Hope, Ph.D., MBA, Director, Product Management,
Discovery Toxicology, Cellumen, Inc.
The Cellular Systems Biology (CSB™) approach to in vitro toxicology combines mechanistic indicators of toxicity, cell-based models of relevant target organs and proprietary classification algorithms to predict in vivo toxicity.  Human and rodent CSB panels have been developed in hepatic, cardiac and neuronal systems and provide a reliable and objective method to predict the safety risk of compounds in development. 

11:00-11:25 To be Announced

11:25-11:50  Using the in vitro High-Content Imaging Methods for Target and Non-Target Organ Toxicity Assessment
Madhu S. Mondal, Ph.D., Head, Early Toxicology, Novartis Institutes for BioMedical Research, Inc.
Abstract unavailable at the time of printing.

HCA Data Management

10:20-10:45    Data Management in the HCAWorld
Keith Grochow, IT Lead Analyst, Scientific Information Sciences, Roche
Management of large, and ever growing, amounts of HCA data is a challenge facing many organizations. This discussion will explore how Roche is dealing with the challenges of growing amounts of data in this business area. Other topics will include: things to consider to provide a good support model for the HCA group; storage, archiving and retention of data; partnering with the local HCA group.

10:45-11:10    Developing Enterprise-Level IT Solution to Support High-Content Screening Experiments
Michael Lenard, Informatics Technical Lead, Informatics & Technology Automation, Bristol-Myers Squibb Co.
The increasing role and growing use of high-content screening in modern drug discovery requires new approach to HCS informatics to fully benefit from HCS technology and, at the same time, increase productivity of HCS users. The presentation focuses on challenges in development of HCS Road – the enterprise-level IT solution implemented in BMS. Among topics addressed are: how standardizing on single HCS data model and choosing correct data storage strategy allows efficient and cost effective data management, how to efficiently access and work with HCS data across enterprise network, and how to collect and integrate data from different HCS imaging platforms (including approach to automation). Other topics covered are: annotation of HCS experiments, how linking of the annotation and high-content data benefits users by allowing flexible visualization of all experimental data (e.g. plate annotation, high-content data, images, etc.), and examples of some data analysis and presentation techniques.

11:10-11:35    The Open Microscopy Environment:  Open Image
Informatics and Metadata Management for Biological Microscopy and High-Content Screening
Jason Swedlow, Ph.D., Wellcome Trust Centre for Gene Regulation and Expression, Wellcome Trust Biocentre, University of Dundee (UK)
Many advances have been made in instrumentation for light microscopy and image analysis algorithms, but computational cell biology remains out of reach for many investigators. Proprietary file formats and closed, non-standardized programming interfaces make integration of new imaging methods and algorithms difficult at best.  We have developed an open-source software framework to address these limitations called the Open Microscopy Environment (OME).  OME has three components—an open data model for biological imaging, standardized file formats and libraries for data file conversion and applications for image data management and analysis.  The OME Data Model is a foundation for the whole OME project, and has recently been updated to more fully support fluorescence filter sets, unique identifiers, and HC imaging using multi-well plates. Bio-formats is a pluggable Java library for reading data from proprietary file formats and the OME-TIFF file format is an open specification for an exchangeable image data file.  These resources enable access to data by different processing and visualization applications, and data sharing between collaborators.  The OMERO application platform combines an image metadata database, a binary image data repository and an image management application to provide a multi-user image management, visualization, and analysis facility.

11:35-12:00    Insight into HCS Data Stream: Analytical Strategy and Applications
Ansuman Bagchi, Ph.D., Director, Applied Computer Science and Mathematics, Merck & Co., Inc.
Abstract unavailable at the time of printing.

Flow Cytometry

10:20-10:45    High-Content/High-Throughput Screening of Multiplexed Targets by Flow Cytometry
Larry A. Sklar, Ph.D., Regents Professor, Pathology; Distinguished Professor, Pharmacy; Director, Basic Research, UNM Cancer Research and Treatment Center; Director, New Mexico Molecular Libraries Screening Center
On behalf of the NIH Roadmap Molecular Libraries Initiative and in collaboration with many investigative teams, we have completed screens of six families of multiplex targets: 1) FPR family; 2) ABC transporters;  3) Bcl-2;  4) RGS; 5) small GTPases; and 6) MMP substrate families.  These screens reflect millions of data points. Biological content is increased with mixtures of targets that are color-coded as suspension arrays. Chemical content is increased with combinatorial mixtures provided by TPIMS of thousands of compounds per well arrayed in combinatorial fashion. The GTPase family screens, for example, reveal both small molecule activators and inhibitors with intracellular activities.  We are currently implementing molecular multiplex targets that include DNA-protein and RNA-protein interactions, kinase domain protein-protein interactions, and interactions of low molecular GTPases with their regulatory proteins. In cell-based high-content analysis, we are exploring yeast multiplex model systems for TOR pathway analysis, cell cycle regulation, transporters, protein-DNA repair interactions, and host-pathogen yeast two hybrid protein interactions.

10:45-11:10    Phospho Flow Cytometry: Multiparameter Kinase Assays at the Single Cell Level
Peter O. Krutzik, Ph.D., Senior Scientist, Baxter Laboratory in Genetic, Pharmacology, Department of Microbiology and Immunology, Stanford University
Flow cytometry is a powerful tool for analyzing 10 or more parameters at the single cell level. Recently, the use of phospho-specific antibodies has allowed us to measure intracellular signaling events in addition to classical surface markers, enabling us to analyze kinase signaling cascades in complex primary cell populations such as peripheral blood. Using phospho flow, we performed a small molecule drug screen with a library of natural products in primary cells. The screen yielded both pathway- and novel cell type-specific inhibitors of cytokoine-induced Jak-Stat signaling. The method was used both in vitro and in vivo to confirm drug activity. To improve sample throughput, we employed Fluorescent Cell Barcoding (FCB), a multiplexing method that enables combination of samples prior to antibody staining. For data analysis, we have developed software tools that provide rapid exploratory overviews, including heatmaps, of high dimensional datasets.

11:10-11:25    Sponsored Presentation(Opportunity Available.)
Contact Katelin Fitzgerald, Manager, Business Development, at 781-972-5458 or kfitzgerald@healthtech.com.

11:25-11:50    Multiparametric Flow Cytometry Analysis of Immune Responses for Vaccine Development
Mario Roederer, Ph.D., Senior Investigator, ImmunoTechnology Section, Vaccine Research Center, NIAID, NIH
Current efforts to develop vaccines designed to elicit effective T cell responses, likely required for HIV, TB, and malaria, are hampered by the lack of knowledge as to what constitutes the appropriate response.  T cells are very heterogeneous, comprising dozens of phenotypically distinct subsets, each of which can express dozens of functions in different combinations.  Using highly multiparametric flow cytometry, we can quantify a large number of functions and phenotypes simultaneously on individual antigen-specific T cells.  This technology is being applied in mouse and non-human primate models where responses can be correlated to protection from challenge.  By measuring responses engendered in humans by different vaccine regimens, we hope to optimize vaccines to provide the greatest level of desirable T cell responses.  Over the past ten years, the challenges for applying this technology have evolved from hardware to reagents and applications, to what remains to be the greatest hurdle still extant:  the enormously complex bioinformatics and statistics required by the large volume of highly complex data.

11:50-12:15    Use of the ImageStream Imaging Flow Cytometer Platform to Investigate the Role of PI3K in Whole Blood Phagocytosis of GFP-labeled E.coli
Gary Elliott, Ph.D., Senior Scientist, HTS Molecular Pharmacology, Amgen
Phagocytosis of pathogens is an extremely complex process involving numerous receptors and intracellular signaling pathways. Because the innate immune system acts as the first line of defense against invading microorganisms, inhibited or aberrant phagocytosis may lead to an increase in opportunistic infections. However, understanding the impact of inhibiting individual molecules is further complicated in heterogenous cell populations such as whole blood by the presence of more than one type of professional phagocyte. PI3 kinase has been demonstrated to play a role in Fc receptor mediated phagocytosis of immunoglobulin or complement opsonized pathogens, but it is unclear what role inhibition of specific PI3 kinase isoforms would play in ingestion or killing of bacteria in whole blood. The ImageStream is a flow cytometer platform capable of obtaining images of cells in flow simultaneously in six parameters with multiple lasers. The ImageStream has previously demonstrated utility in applications as diverse as nuclear translocation, receptor internalization and immunophenotyping. Here, we examined the potential for developing highly multiplexed assays on the ImageStream that may combine one or more of these assays with measurement of phagocytosis. We utilized several small molecule inhibitors of various PI3 kinase family members to characterize the role these signaling molecules may play in various leukocyte subsets during phagocytosis of green fluorescent protein-expressing bacteria.

Luncheon Showcase: High-Content Screening

12:15-12:45 Speed: Recent Developments for the MDS Complete Imaging Solution                                            
Michael Sjaastad, Ph.D., Marketing Director, Cellular Imaging, Molecular Devices (Now Part of MDS Analytical Technologies)
The Complete Imaging Solution from MDS Analytical Technologies has evolved to provide more speed. The IsoCyte high-speed, whole well imaging system is now part of the Complete Solution.  IsoCyte can scan a 1536 well plate in 3-5 minutes at cellular resolution.  We have also developed a high-speed parallel processing for MetaXpress application modules analysis. The solution decreases image analysis times 10-50 fold depending on the number of processors deployed.

12:45-1:00  Advances in HCA Image Analysis Using IN Cell Investigator Software
Daniel J. Collins, Applications Specialist, GE Healthcare Advanced Systems
The combination of data from fluorescent probes and sensors coupled with detailed morphological analyses provides in-depth insight into the effects of drugs, siRNA and other perturbations of cellular functions. GE Healthcare’s IN Cell Analyzer 1000 is a versatile platform used across the drug discovery process, from target identification and validation, high throughput screening and lead optimization, to the evaluation of compound toxicology. New image analysis methods for object segmentation, label-free object identification utilizing transmitted light-derived images, and image stitching are a few of the recent additions to the IN Cell Investigator software. In addition, the instrument software has been modified to include a sufficient cell count feature during image acquisition. Potential applications of these methods will also be discussed.

1:00-1:45     Sponsored Presentations (Opportunities Available.)
Contact Katelin Fitzgerald, Manager, Business Development, at 781-972-5458 or kfitzgerald@healthtech.com.

Luncheon Showcase: High-Content Data Analysis

12:15-12:30    Extracting ‘Omic Knowledge from HCS Data
Thermo Scientific

12:30-12:45    Streamlining Toxicology Workflow from Assay Development to Decision
Abhay Kini, PhD. Product Marketing Manager GE Healthcare IN CELL Software

12:45-1:15      Open Systems and Software Interoperability in HCA: A Working Example
J. Scott McDonald, Vice President, Software Research & Development, Bioimaging Systems, BD Biosciences
With an increase in both the number of assay types being performed and the number of software tools available for image and data analysis, it has become important for the HCA community to have the ability to select the software tools of choice to process and analyze images and data.  The various levels of software interoperability will be discussed from close integration through open integration and open data access.  Specifically, the integration of BD IPLabÔ image analysis and Pipeline PilotÔ data analysis into the BD PathwayÔ software and an open data interchange mechanism will be presented.

1:15-1:30          Systematic Processing, Analysis and Management of High-Content Screening Data
Stephan Heyse, Ph.D., Head, Lead Discovery Informatics; Manager, Scientific Development and Business, Genedata Screener, Genedata
Data complexity and volume in a high-content screen of a compound library or a genome-wide siRNA screen exceed the capacity of in-depth scientific review. Typically, this leads to an extreme simplification of analysis (e.g. using only a single outcome like cell number as the result.) In our case study, we show a systematic approach to extract the most information-rich features from cell images, aggregate these to activity profiles which qualify and quantify the observed biological processes, separate effects from artifacts by in-depth QC, and compile results from a complete screen. We show how this approach increases the robustness of results and the value derived from high-content screening experiments.

1:30-1:45        Sponsored Presentation

Pipeline Pilot and Scitegic Enterprise Server: a New Flexible, Modular Approach to HCS Workflow Development and Automation
Kurt M. Scudder, Ph.D., Scientist, Accelrys, Inc.
Many practitioners of HCS find the need to go beyond the capabilities of the software provided with their HCS instrumentation and develop custom workflows for some aspect of image analysis, data analysis, data management, or integration. Standard image analysis programs typically do not have all workflow functionality, nor do they support custom user interface and database connectivity. The use of a workflow building-block system comprised of customizable interoperable modules can provide all of the core functions for HCA data workflows. Accelrys’ Pipeline Pilot and Scitegic Enterprise Server products implement just such a system. It can be used to build anything from small workflows which can connect to an existing HCS data management infrastructure, up to entire HCS data management systems. The Pipeline Pilot toolbox contains elements for image analysis, plate data analysis and results management, graphical user interface development, database and file reading and writing, as well as a complete suite of chemically intelligent components for joining HCS results with medicinal chemistry data, sequence analysis, and gene expression data. Finally, developed methods can be rolled out to users via the web browser available on every computer. Several examples of HCS workflow automation using Pipeline Pilot will be shown.

HCA for Compound Screening

2:15-2:40     High-Content Screen to Look for Inhibitors of Virus Entry
Marjo Simonen, Ph.D., Lab Head, High-Content Screening, Lead Finding Platform, Novartis Institutes for BioMedical Research
Human cytomegalovirus (CMV) belongs to a subfamily of Herpesviridae. Approx-imately 60-90% of adults worldwide are infected and have detectable antibody levels to CMV. CMV infection can be life threatening for immunocompromised patients including organ transplant recipients and patients with AIDS. CMV is also the most significant viral cause of birth defects in industrialized countries leading to mental retardation and deafness of neonates.  We have developed an imaging assay to screen for inhibitors of virus entry to normal human fibroblasts and human retinal pigmented epithelial cells. The assay exploits a genetically modified virus expressing a GFP fusion protein as part of its virion tegument. Upon successful entry, the GFP tagged tegument protein translocates to the nucleus permitting monitoring of virus infection. As a measure of active compounds the number of infected cells where the virus reaches the nuclear region was determined based on a threshold setting (% infected cells). The readout for cell number enabled the classification of the compounds to toxic and non-toxic ones in presence of the virus. In addition, the readout for granularity was used to categorize the compounds based on their mode of action. The compounds that induced granular appearance of the virus at cell surface allowed docking of the virus but no further internalization. Compounds without granular pattern might inhibit the docking onto the host cells. A good correlation of tool compound activity, mode of action and toxicity was found between the two cell lines.  The screen demonstrates the ability of HCS to characterize compounds in an early phase of discovery by using only a single assay approach.

2:40-3:05     Our Experience with High-Content Screening Using Wound Healing Assay
Miroslav Cik, Ph.D., Head, Assay Development, Department of AD & HTS, Johnson & Johnson
Wound healing assay is being used to study the cell migration process and to estimate the migration rates of different cell types. Typically, a small area in MW 96 or 384 is disrupted and displaced by scratching a line through a confluent monolayer. The monolayer recovery is followed by automated microscopy and automated image analysis. Our experience with screening of a large library collection of small molecules and siRNAs will be presented.

3:05-3:30     Evaluation of High-Content Analysis Platforms to Support GPCR Screening
Andrea Weston, Ph.D., Research Investigator, Lead Discovery, Bristol-Myers Squibb & Co.
G protein-coupled receptors (GPCRs) represent one of the most popular target classes for therapeutic intervention. Accordingly, there are several cell-based assay platforms available to support small-molecule screens for modulators of GPCR signaling.  With improved efficiency of high-content screening, we are better poised than ever to identify compounds that interfere with GPCR function in a specific and selective way.  Focusing on a GPCR target of interest, we have evaluated the use of high-content assays (HCA) such as the Transfluor β-arrestin internalization and receptor internalization using the Opera and Cellomics’ imaging platforms, respectively.  Using a test-set of compounds, these technologies were evaluated alongside an Aequorin assay for calcium mobilization, Invitrogen’s Tango assay, which indirectly measures receptor coupling to β-arrestin, and a receptor binding assay.  This comprehensive analysis has provided a clearer understanding of the utility of HCA in GPCR screening.

HCA Data Analysis

2:15-2:40       Practical Handling and Analysis of High-Content Screening Data
Karol Kozak, Ph.D., Head of Computation Analysis, HCA/HTS Informatics, LMC-RISC, Institute for Biochemistry 
Cell-based High-Content Screening (HCS) using automated microscopy is an upcoming methodology for the investigation of cellular processes and their alteration by multiple chemical or genetic perturbations. The analysis of the large amount of data generated in HCS experiments represents a significant challenge and is currently a bottleneck in many screening projects.  This workshop session reviews the different ways to analyze large sets of HCS data, including the questions that can be asked and the challenges in interpreting the measurements.  The main data mining approaches used in HCS, such as image descriptors computations and classification algorithms, will be outlined.

2:40-3:05       Application of Novel Multiparametric High-Content Toxicity and Phenotypic Assays: Integration with Advances in Automation, Image Analysis, Multivariate Statistics and in silico Modeling
Neil Carragher, Ph.D., Associate Principal Scientist, Advanced Science and Technology Lab, AstraZeneca Research & Development, Charnwood
We have developed a suite of novel high-content toxicity and phenotypic assays with customized image analysis algorithms that enable accurate cell segmentation and the extraction of multiple features across a panel of physiologically relevant cell types. The image-based assays and analysis methods presented describe the ability to assess compound effect upon several aspects of hepatotoxicity and cell phenotype including, cell-cycle, cell and nuclear morphology, actin and microtubule cytoskeleton and 3-dimensional cell invasion. By integrating these assays with the latest advances in automation, image informatics, multivariate statistics and in silico modelling we can apply these approaches to large scale compound profiling activities to provide detailed information on biological mode-of-action, supporting compound  differentiation and target selectivity studies.

3:05-3:30       Evaluation of Data Analysis Tools for the Multivariate Classification of Compound Responses Detected by High-Content Screening
Anne Kuemmel, Ph.D., Post Doc Researcher, Novartis Institutes for BioMedical Research
High-content screening providing automated sub-cellular imaging and image analysis generates multidimensional readouts that report on cellular phenotypes. These phenotypic profiles constitute a detailed, cell-by-cell data basis about the cellular effects in response to tested compounds. Multivariate statistics provide a range of data reduction and classification tools to not only identify hits but also to classify the compound’s effect and to consider different responses in subpopulations. It is however not clear which are the most suitable data analysis methods to apply to the analysis of high-content screening results to exploit the information provided by HCS to full capacity. Different approaches for data normalization, multivariate characterization and classification exploiting the phenotypic, cell-by-cell data are discussed.

Novel Probes and Biosensors

2:15-2:40        A Novel p53-HDM2 Protein-Protein Interaction Biosensor HCS: Assay Development, Screening and Hit Characterization
Paul A. Johnston, Ph.D., University of Pittsburgh Drug Discovery Institute, Research Associate Professor, Department of Pharmacology and Chemical Biology, School of Medicine
The p53 tumor suppressor is the central node of the p53 network/pathway that integrates the many diverse signals that regulate cell cycle arrest, apoptosis, senescence, DNA repair, metastasis, and angiogenesis. TP53 is the most frequently mutated gene (~ 50%) in human tumors, and when p53 is not mutated, aberrations occur (~ 50%) in other components of the p53 signaling pathway. p53 activation requires both stabilization and accumulation of the protein, a complex process tightly regulated through its interactions with MDM2 and MDM4 protein partners. A novel p53-HDM2 protein-protein interaction biosensor assay developed by Cellumen, Inc. was utilized to screen more than 200,000 compounds for disruptors of the interaction between p53 and HDM2, and the performance of the HCS and characterization of the chemical probes identified will be described.

2:40-3:05        Applications of Small Molecule Activity-Based Probes for Non-Invasive Optical Imaging
Matthew Bogyo, Ph.D., Assistant Professor, Department of Pathology, Stanford University School of Medicine
Abstract unavailable at the time of printing.

3:05-3:30        Chemical Reporters for Probing and Visualizing Lipoproteins and Lipids in Cells
Rami Hannoush, Ph.D., Scientist, Protein Engineering, Genentech, Inc.
Novel chemical reporters were developed as probes for post-translational modification of proteins. Fatty acylation of cellular proteins is critical for mediating protein-protein and protein-lipid interactions. In cancer, protein fatty acylation is important for regulating signaling pathways like apoptosis and in tumorogenesis. We report here the design and synthesis of novel non-radioactive reporters and their  application for the rapid detection and fluorescence-based visualization of protein fatty acylation in cells. These new reporters are useful for studying the behavior of cellular lipoproteins and lipids, have potential in investigating protein-lipid interactions at the cellular level, and are amenable for use in high-content assays aimed at investigating the behavior of cellular lipids. 

HCA for Compound Screening
(continued)

4:45-5:10     Phenotypic Screening: A Complementary Drug Discovery Paradigm
Jonathan A. Lee, Ph.D., Senior Research Advisor, Quantitative Biology, Eli Lilly and Company
Lilly Discovery is exploring the use of cell based phenotypic and pathway assays to chemically interrogate complex biological systems composed of multiple or unknown biochemical components/pathways.  Such Phenotypic Drug Discovery (PDD) approaches complement classical target-directed drug discovery strategies.  Potential advantages of PDD approaches include direct identification of cell active compounds, systematic inclusion of both known and unknown molecular components/mechanisms and interrogation of molecular targets in a relevant, cellular context. Concerns with cell based approaches include design of appropriate compound libraries, the statistical robustness and throughput of assays, and whether cellular assays can provide compound structure-activity relationships. This presentation is a status report on the use of PDD approaches at Eli Lilly and Company.  Our experience to date indicates how multiplexed cellular imaging and the resulting compound activity fingerprints can be combined with advanced data analysis tools and computational hit expansion algorithms to identify compounds with known and novel mechanisms. 

5:10-5:35     Glucocorticoid Nuclear Hormone Receptor Translocation HCS Assay: Screening the NIH Compound Library to Identify Inhibitors of Dynein-Mediated Cargo Transport on Microtubules
Paul A. Johnston, Ph.D., Research Associate Professor, Department of Pharmacology and Chemical Biology, School of Medicine, University of Pittsburgh Drug Discovery Institute
Cytoplasmic dynein is the molecular motor that carries cargo to the minus ends of microtubules (MTs) (e.g., from the cytoplasm to the nucleus), and provides the mechanical force for many other important functions, including nuclear envelope breakdown and sister chromatid exchange during mitosis. Dynein motor complexes (dynein+dynactin+adaptor proteins) interact with transport cargos such as the glucocorticoid nuclear hormone receptor (GNHR) and utilize the hydrolysis of ATP to power their movement along MTs towards the nucleus. The Mouse mammary adenocarcinoma cell line (3617.4) stably expressing a rat GNHR green fluorescent fusion protein (GNHR-EGFP) under the control of a tetracycline (tet)-regulated promoter has been utilized as a model system to investigate hormone dependent GNHR-EGFP translocation. The development of a GNHR-EGFP translocation HCS assay using this cell line will be presented together with the results from screening the NIHs 212,000 compound library, and characterization of the Dynein inhibitor probes identified. 

5:35-6:00     Application of Quantitative High-Throughput Screening to High-Content Analysis of Cell-Based Assays
Sunita J. Shukla, M.P.H., Ph.D., Research Fellow, National Institutes of Health Chemical Genomics Center
Traditional high-content assays, such as cell cycle analysis, have used microscope-based possessing that prohibits high-throughput screening of large chemical libraries. We have pioneered the application of laser-scanning cytometry using the Acumen Explorer eX3 to various cell-based assays to screen libraries of greater than 200,000 compounds at multiple concentrations. However, the Acumen provides low-resolution images, thus we have sought to develop a pipeline that couples the throughput of the Acumen to systems with high-resolution imaging capabilities. Thus, the Kalypsys automated robotic system was used for large-scale screening of cell-based assays on the Acumen. As a proof-of-principle study, we describe the effects of small molecules on various cell cycle phases by scanning 1536 well plates on the Acumen and then selecting active wells for higher resolution imaging. For cell cycle analysis, we profiled several thousand compounds, including those that are part of the Environmental Protection Agency and National Toxicology Program compound libraries, as seven point concentration response curves (CRCs). This presentation will address the utilization of an automated informatics pipeline for the wells that showed high quality CRCs from the Acumen in order to image them on the IN Cell Analyzer 1000, an automated wide-field fluorescent microscope imager. The use of an automated integrated imaging system and our informatics pipeline should enable the screening of greater than 100,000 compounds for cell-based imaging assays involving multiple readers.

HCA Data Analysis
(continued)

4:45-5:10       Statistical Analysis of High-Content Data: Harnessing the Full Data Matrix
Amir A. Handzel, Ph.D., Senior Biometrician, Biometrics Research, Merck & Co., Inc.
High-Content Analysis of medium-to-high throughput assays presents both challenges and opportunities in terms of data analysis methodologies.  Current analysis typically relies on well level summaries that do not always capture all useful information at the cell population level.  In the other dimension of the data matrix, extracted cellular image features are often treated individually, that is in a univariate way.  We take advantage of several multivariate statistical methods that take into account cell-level measurements for multiple features simultaneously.  They can be applied for diverse purposes, including small molecule hit detection and statistical selection of relevant image features. 

5:10-5:35       Obtaining Better Mechanistic Understanding of Patterns in High-Content Analysis
Robert F. Murphy, Professor, Departments of Biological Sciences and Biomedical Engineering, Carnegie Mellon University
Many proteins or other biological macromolecules are localized to more than one subcellular structure. The fraction of a protein in different compartments can be measured by colocalization with organelle-specific markers, requiring availability of fluorescent probes for each compartment and acquisition of images for each in conjunction with the macromolecule of interest. Alternatively, tailored high-content analysis algorithms allow finding particular regions in images and quantifying the amount of fluorescence they contain. Unfortunately, this requires extensive hand-tuning of algorithms and is often cell type dependent.  I will describe a machine learning approach for estimating the amount of fluorescent signal in different subcellular compartments without hand tuning, requiring only the acquisition of separate training images of markers for each compartment. The method has been validated using high-throughput microscopy.  It will enable automated and unbiased determination of the distributions of protein across every cellular compartment, and will significantly improve imaging-based high-throughput assays and facilitate large-scale, proteome-wide localization efforts.  I will also describe new approaches for automatically capturing and modeling nuclear and cell shape, which has many high-content analysis applications including interpretation of results from drug and RNAi analysis.

5:35-6:00       Weak Signals and Strong Artifacts: HCS in Antiviral Drug Discovery
Emmanuel Gustin, Ph.D., Principal Scientist, Tibotec BVBA, Jonhson & Johnson
The adoption of high-content screening for the discovery of small-molecule antiviral drugs promises to provide better access to targets other than well-behaved soluble enzymes, increased predictivity of assay data towards the in vivo efficacy, and multiplexing of assay end points. Assay development challenges are inherent in the nature of these targets, which may be a small number of events of short duration during the viral life cycle. This calls for the use of HCS instruments with the best available resolution and sensitivity. In addition, small molecule libraries contain numerous compounds that will induce artifacts such as fluorescence, crystallization and toxicity. The multi-parametric nature of HCS also creates an excellent opportunity to effectively deal with these issues, if a strong image analysis and data analysis effort is integrated in all stages of the process, from assay development and equipment choice to final reporting. Given tools with high performance and flexibility, methods can be established to improve the specificity of signals, to avoid artifacts, to exploit data redundancy, to detect problems, and to measure known off-target effects. A toolbox can be inventoried, that encompasses the broad set of skills that should be available in the HCS team, as well as the flexible tools and techniques it should have available.