Executive Sponsor:
Corporate Sponsors:
Sponsoring Publications:
Arnold Genome Research IOS Press Taylor and Francis Scientific Computing World CHI's Outlook for DNA Microarrays CHI's Life Sciences Informatics: From Data to Drugs
Sponsoring Society: International Informatics Society Web Partners: Lab-on-a-Chip.com Biochipnet

Miniaturized platform technologies are no longer new concepts. Chips and microarrays are becoming standard tools for the high-throughput analysis of gene expression and continue to grow in information quality and to discover new applications. The power of microarray analysis lies in the ability to compare large sets of genes in different tissues or conditions to identify pathways and regulatory networks. Building upon the successful foundation of Cambridge Healthtech Institute's Fifth Annual Lab-on-a-Chip and Microarrays, the critical topic of Informatics and Microarray Data Analysis completes the picture. The meetings are a must-attend event for biological researchers, drug developers, diagnosticians, engineers, statisticians, bioinformaticists, database developers, and computer software designers. Each discipline (topic) contributes to a successful team effort and like the microarray must be measured as a whole.

Preconference Short Course Statistical Idiosyncrasies of Microarray Data and Pitfalls to Avoid Mr. Thomas J. Downey, President, Partek Incorporated

Keynote Presentation
Dr. Uwe Dengler, Novaris Pharma AG

Data Quality/Experimental Design
Dr. Peter Adorjan, Epigenomics AG
Dr. Andrew B. Goryachev, GeneData AG
Dr. Sandrine Imbeaud, CNRS
Dr. Alexandre Bureau, Genome Therapeutics Corporation

Tools for Utilizing Data
Mr. Misha Kapushesky, European Bioinformatics Institute
Dr. Christopher M.L.S. Bouton, LION Bioscience Research, Inc.
Dr. Nathalie Cotte, INSERM
Dr. Joaquín Dopazo, CNIO
Dr. Petra Scheffer, SPSS Science Software GmbH
Dr. Mark Demesmaeker, European Manager, Spotfire AB

Statistical Evaluation
Dr. Robert Nadon, McGill University
Dr. Torsten C. Kroll, University of Jena
Dr. Matthew A. Roberts, Nestlé Research Center
Dr. Chiara Romualdi, University of Padova
Dr. Paul Vauterin, Applied Maths BVBA
Dr. Michel Bellis, CNRS

Data Interpretation and Analysis
Dr. Doron Lancet, The Weizmann Institute of Science
Dr. Michael D. Kramer, Lynx Therapeutics GmbH
Dr. Tarif Awad, Manager, Affymetrix Genomics Collaborations
Dr. Matthias Fellenberg, Biomax Informatics AG
Dr. Andre Nantel, National Research Council of Canada

Scientific Advisors
Dr. Douglas E. Bassett, Jr., Rosetta Biosoftware
Dr. Alvis Brazma, European Bioinformatics Institute
Dr. Stephen Sharp, Iobion Informatics

 

MONDAY, FEBRUARY 10

13.00 Short Course Registration 14.00-17.00 Preconference Short Course: Statistical Idiosyncrasies of Microarray Data Mr. Thomas J. Downey, President, Partek Incorporated Because of the large number of genes and the relatively small number of samples available in the typical microarray study, application of many statistical methods needs some special attention. Having many more genes than biological samples (curse of dimensionality) will have a significant impact on the way we use statistics to find which genes are differentially expressed and also on the problem of estimating the predictive capabilities of diagnostic classifiers based on the expression of a small set of genes. 17.00.19:00 Early Conference Registration and Poster and Exhibit Set-up

 

TUESDAY, FEBRUARY 11

7.30 Conference Registration, Poster and Exhibit Viewing, and Light Continental Breakfast

8.30 Chair's Opening Remarks
Dr. Stephen Sharp, Director of Marketing, Iobion Informatics

8.45 Keynote Presentation: Interpretation of Microarray Data for Functional Genomics
Dr. Uwe Dengler, Research Program Team Head, Life Science Informatics, Functional Genomics, Novartis Pharma AG
The basis for large scale microarray gene expression analysis in big research organizations are standardized mechanisms to capture and annotate data. Rapid evolution of gene expression analysis requires a flexible infrastructure that can be connected with a variety of tools and allows for an integration of gene expression data with other types of genomic data. Besides the application of expression profiling within a specific experimental context, growng repositories of microarray data open new avenues of research applying database-wide searches to understand the function and interaction of genes and their products.

 

DATA QUALITY/EXPERIMENTAL DESIGN

9.30 Statistical Process Control for Large-Scale Microarray Experiments
Dr. Peter Adorjan, Vice President, Information Sciences, Epigenomics AG
We introduce a novel method for maintaining and controlling data quality in large-scale microarray studies. When experiments are standardized enough, then process-dependent alterations are relatively rare events. Therefore, instead of reducing these effects by repetitions one should rather detect problematic slides or slide batches and repeat only those. This can only be achieved by controlling process stability. We demonstrate the power of our approach on three large sets of DNA methylation microarray data.

10.00 High-Throughput Quality Assessment and Processing of Microarray Data
Dr. Andrew B. Goryachev, Senior Scientist and Product Manager, GeneData AG
A variety of array-based gene expression profiling technologies are successfully applied in biomedical research as well as in the development of drugs and diagnostics. As mRNA laboratories are scaling up their facilities, manual assessment of chip images becomes cumbersome and is prone to subjective criteria. To standardize data quality control we have developed Expressionist Refiner, a system for the highly automated assessment and optimization of data quality. The system monitors a variety of quality parameters and offers means to minimize noise, remove systematic measurement errors, and reveal potential array defects. We show examples of high-throughput studies using Affymetrix GeneChips and two-color microarrays and demonstrate the increase in value of array experiments due to enhanced data reliability and experimental sensitivity.

10.30 Poster and Exhibit Viewing, Refreshment Break

11.15 "The 39 Steps" in Expression Profiling with Microarrays
Dr. Sandrine Imbeaud, Microarray Team Leader, Human Functional Genomics Program, CNRS
High-density cDNA arrays are powerful and robust tools enabling the production of large gene expression data sets, but a number of questions still need to be addressed. What is the validity and quantitative accuracy of the observed changes? Which genes should be prioritized for further study? How does one determine whether a given change is a cause rather than a consequence of a specific biological condition? Based on the results of case studies in cancer expression profiling, the presentation will give an overview on the organization of a microarray workflow in the laboratory, focusing on standardization, quality control assessment, annotation, and authentication procedures to manage and share expression data.

11.45 Modeling Variance as a Function of Expression Level in Microarray Experiments
Dr. Alexandre Bureau, Biostatistician, Statistical Genetics, Genome Therapeutics Corporation
Microarray experiments to compare gene expression levels between different conditions often include too few replicates to reliably estimate gene-specific null distributions of summary statistics. As a result, many methods proposed to analyze such experiments rely on the assumption that summary statistics are identically distributed for all genes. That assumption is contradicted by the observation that the variability in expression measurements at different degrees of replication depends on the level of expression. When at least two replicates are available, we propose to standardize measures of expression differences by an estimate of standard deviation that is a function of the expression level. We investigated parametric and nonparametric functions. Empirical assessment and simulations show that local polynomial regression of squared residuals provides a better fit than parametric variance functions. The methods apply to both oligonucleotide and spotted cDNA microarray data sets and adapt to various experimental designs and various degrees of replication. Data sets from gene expression profiling experiments in microorganisms are used to illustrate the methods.

12.15 Panel Discussion

12.45 Lunch (on your own)

 

TOOLS FOR UTILIZING DATA

14.00 Chair's Remarks
Dr. Douglas E. Bassett, Jr., Vice President and General Manager, Rosetta Biosoftware

14.05 Managing and Analyzing Microarray Data Online
Mr. Misha Kapushesky, Scientific Application Programmer, EMBL Outstation, European Bioinformatics Institute
Expression Profiler (EP, http://ep.ebi.ac.uk/) is a set of tools for the analysis and interpretation of gene expression and other functional genomics data. These tools perform expression data clustering, visualization and analysis, integration of expression data with protein interaction data and functional annotations, like GeneOntology, and the analysis of promoter sequences for predicting transcription factor binding sites. Several clustering analysis method implementations and tools for sequence pattern discovery provide a rich data-mining environment for various types of biological data. A new highly flexible interface has recently been developed for these web-based tools, which allows further simultaneous integration of expression data analysis with other tools and databases, for instance SOTA (self-organizing tree array), and the public microarray database ArrayExpress.

14.35 The Role of the Local Research Microarray Database System in the Data Analysis Process
Dr. Stephen Sharp, Director of Marketing, Iobion Informatics
Summary unavailable at time of printing.

15.05 Analyzing Large-Scale Gene Expression Data
within a Biological Context Using DRAGON and DRAGON View
Dr. Christopher M.L.S. Bouton, Bioinformatician, LION Bioscience Research, Inc.
Large-scale gene expression data can be annotated with various types of biological information in a simultaneous, comprehensive fashion using the DRAGON database (Bouton and Pevsner, 2000). Preannotation of expression data sets allows for novel forms of analysis that incorporate biologically relevant information such as encoded protein domains, participation in cellular pathways, and chromosomal location. DRAGON View (Bouton and Pevsner, 2002) provides a set of web-accessible tools that aid in the analysis of biologically annotated expression data sets. Using example data I will demonstrate the use of both DRAGON and DRAGON View and show how they can be used to conduct novel forms of data analysis within a biological context.

15.35 Poster and Exhibit Viewing, Refreshment Break

16.00 GenMAPP and MAPPFinder, New Tools for Viewing and Analyzing Microarray Data on Biological Pathways
Dr. Nathalie Cotte, Postdoctoral Fellow, INSERM, Collège de France
GenMAPP (www.GenMAPP.org), a stand-alone free application, displays gene expression data on biological pathways by color-coding the genes based on data and criteria provided by the user. Its graphical capacities allow the modification of downloadable pathways and the custom design of new ones. To quickly identify the gene expression changes occurring in a data set across all areas of biology, MAPPFinder relates the microarray data set to the Gene Ontology hierarchy and displays the results as a ranked list in a searchable browser. Practical examples and results from using GenMAPP are highlighted in this presentation.

16.30 Sample Classification and Selection of Relevant Genes in DNA Microarray Data Using Discriminatory Capacity and Biological Information
Dr. Joaquín Dopazo, Bioinformatics Unit, Centro Nacional de Investigaciones Oncológicas (CNIO)
A combined approach is described that produces an efficient classification of different classes (e.g., cancer types, tissues, etc.) defined among a series of experimental conditions in microarray gene expression experiments and, at the same time, permits the selection of the genes that account for the differences among classes. The approach implies three steps. First, the dimensionality of the data set of gene expression profiles is reduced to a smaller number of nonredundant clusters of co-expressing genes. Second, the average values of these clusters are used for training a perceptron, a widely used type of neural network. The relative importance of the clusters of genes in the definition of the classes can be inferred from the magnitudes of the interconnection weights in the trained perceptron. And, finally, the proportion of different GO (Gene Ontology) terms found in the clusters connected with strongest weights is compared to the basal distribution of terms in the rest of the genes by means of a Fisher exact test. Only clusters that, in addition to strong weights, display a distribution of GO terms that significantly differs from the background are selected. In that way both discriminatory capacity and biological relevance are used for the selection of genes of interest.

17.00 Data Mining of Microarray Gene Expression Data
Dr. Petra Scheffer, Senior Marketing Manager, SPSS Science Software GmbH
The analysis of microarray gene expression data is moving toward more robust statistics and mathematical analytics that go far beyond the fold analysis and basic clustering. Data mining has proven to be an indispensable tool in microarray data analysis by uncovering patterns and relationships in gene expression data and turning them into predictive, deployable information. Data mining of microarray gene expression data presents a number of challenges given the small number of samples and the large number of genes. Reliable analysis has to address this challenge with proper approaches for normalizing, feature reduction, selecting the best set of genes, controlling false positives, and validating the results. This talk will give an introduction into data mining in this area and will show approaches as to how to address the challenges in microarray data analysis with two publicly available data sets.

17.30 Facing the Challenge of Complexity, Diversity and Quality of Data Generated by Modern High Throughput Technologies
Dr. Mark Demesmaeker, European Manager, Spotfire AP

18.00 Networking Reception (sponsored by Cambridge Healthtech Institute)

19.00 Close of Day One

 

WEDNESDAY, FEBRUARY 12

7.30 Poster and Exhibit Viewing and Light Continental Breakfast

Statistical Evaluation

8.00 Chair's Remarks
Dr. Matthew A. Roberts, Team Leader, Nutritional Genomics for Humans and Companion Animals, and Manager of the NRC Core Facility for Functional Genomics, c248 Molecular Analytics, Nestlé Research Center

8.05 Processing and Statistical Analysis of Microarray Data
Dr. Robert Nadon, Department of Human Genetics and Montreal Genome Centre, McGill University
Processing issues such as bias correction, random error estimation, false positive and false negative control, and outlier detection have figured prominently in recent research on microarray data analysis. An overview of this work will be presented, with emphasis on its relevance to practical aspects of determining differential expression.

8:35 Comparison of Normalization Methods for Gene Expression Array Data
Dr. Torsten C. Kroll, Research Scientist, Molecular Biology, University of Jena
Normalization of data is one crucial step in successful gene expression array data analysis. Various methods are available to normalize the data to an inherent standard (i.e., total signal) to eliminate variations occurring in the measurement process. We developed criteria and visualization tools to choose the appropriate method for a given data set based on ranking (RIP-ranked intensity plots) leading to an optimization in subsequent data analysis. Furthermore, we use two strategies to enhance common global normalization procedures, which allow some adjustment of the normalization avoiding background and saturation effects.

9.05 The Limit Fold Change Model: A Practical Approach for Selecting Differentially Expressed Genes from Microarray Data
Dr. Matthew A. Roberts
The gene selection model presented herein is based on the observation that (1) variance of gene expression is a function of absolute expression, (2) one can model this relationship in order to set an appropriate lower fold change limit of significance, and (3) this relationship defines a function that can be used to select differentially expressed genes. The model first evaluates fold change (FC) across the entire range of absolute expression levels for any number of experimental conditions. Genes are systematically binned and those genes within the to X% of highest FCs for each bin are evaluated both with and without the use of replicates. A function is fitted through the to X% of each bin, thereby defining a limit fold change. All genes selected by the 5% FC model lie above measurement variability using a within standard deviation (SDwithin) confidence level of 99.9%. Real time-PCR (RT-PCR) analysis demonstrated 85.7% concordance with microarray data selected by the limit function.

9.35 Pattern Recognition in Gene Expression Profiling Using cDNA Microarray: A Comparative Study of Different Statistical Methods Applied to Cancer Classification
Dr. Chiara Romualdi, Research Fellow at CRIBI Biotechnology Center, Biology Department, University of Padova
Over the last two years a series of computational tools have been developed for the analysis of different aspects of gene expression profiling. After this great effort on the computational aspects of the microarray data analysis, there is a need of comparative studies that test simultaneously different statistical methodologies. Our work tried to compare the efficacy of various supervised statistical techniques of classifying correctly different tumor types. A simulation approach was initially used to control the huge source of variation among and between patients and to evaluate the capability of different algorithms of a correct classification with different experimental variables. Also, the influence of various dimension reduction techniques has been studied. Furthermore, simulation results have been tested applying the selected classification algorithms to two experimental microarray data sets of human cancers.

10.05 Poster and Exhibit Viewing, Refreshment Break

10.45 A New Bootstrap Implementation for Measuring the Significance of Clusters Obtained from Large Data Sets
Dr. Paul Vauterin, Director of Software Development, Applied Maths BVBA
An important issue in the analysis of gene expression data is the identification of significant clusters or groups inside huge data matrices. As most known grouping techniques tend to produce groups or clusters, regardless of the underlying structure, bootstrap methods have been devised to address the significance of obtained clusters. A new bootstrap technique is presented, which can be used to obtain reliable significance values for virtually every method that splits the data into a number of groups, with only minimal assumptions about the underlying data structure. The method provides a powerful statistical tool to filter out nontrivial clusters from insignificant clusters and has proven to be suitable for large data sets, such as microarray studies comprising several thousands of genes and/or samples.

11.15 The Rank Shift Model: How to Use Replicates to Construct Robust Statistics
Dr. Michel Bellis, Principal Investigator, CNRS
We have developped a complete informatics solution, Arrayon, aimed at managing, displaying and analyzing transcriptome results. We paid special attention to developing rigorous methods in order to detect genes undergoing statistically significant variation and to provide the end user with three meaningful selection parameters (p-value, false discovery rate, and sensitivity). Case examples involving different situations and exploring how to adopt the best analysis strategy according to the structure of the experimental scheme will be discussed.

11.45 Panel Discussion

12.15 Lunch (on your own)

 

DATA INTERPRETATION AND ANALYSIS

13.30 Chair's Remarks

13.35 Molecular Recognition in Biological Repertoires and in Transcription Control Networks
Dr. Doron Lancet, Head, Crown Human Genome Center, Department of Molecular Genetics, The Weizmann Institute of Science
Molecular recognition is a most crucial aspect of any biological system. While at many nodes in a cellular network there are highly specialized recognition elements such as enzymes and receptors, other pathways may be governed by chance encounters between members of a random repertoire and a molecular target, a phenomenon exemplified by the immune and olfactory pathways. In such systems, the affinity between the target and members of the repertoire is distributed with a probability function describing the propensity of obtaining a particular affinity value. We have previously proposed a phenomenological Receptor Affinity Distribution (RAD) formalism, which describes this probability function based on simple statistical considerations. The RAD model is found to provide the best description for binding data for over eight orders of magnitude on the affinity scale and to account for a relationship between repertoire size and the maximal obtainable affinity within different repertoires.

14.05 In-depth Expression Profiling for Systems Biology Experiments on the Human Immune System
Dr. Michael D. Kramer, Managing Director, Lynx Therapeutics GmbH
In order to further our understanding of immune cell differentiation we used LYNX s Massively Parallel Signature Sequencing (MPSSÒ) to establish the world s largest expression database on immune cells with currently over 60 million analyzed transcripts. These datasets lay the basis for our systems biology approach for the identification of novel biomarkers and target candidates. We will present the technological background of database generation and we will also discuss selected examples of our mining experiments, which will further demonstrate the value of MPSSÒ for Systems biology research.

14.35 Analyzing Gene Expression Data: Knowing When to Stop
Dr. Tarif Awad, Manager, Data Analysis Team, Affymetrix Genomics Collaborations
One of the primary challenges in microarray data analysis is "knowing when to stop" analyzing raw data and when to begin focusing on biological interpretation. With the proliferation of tools and approaches, there is a tendency to dwell on the numbers before getting back to the biology. Reaching a useful, biologically meaningful end-point in microarray data analysis requires tools for annotation mining, pathway analysis, and data integration from a variety of sources. Solutions to this challenge are being developed; some will be illustrated in this presentation using a data set from a blood cell development study, together with freely available tools such as NetAffx, Gene Ontologies, and GenMAPP.

15.05 Poster and Exhibit Viewing, Refreshment Break

15.30 Workflow of an Integrated Analysis of Expression Data
Dr. Matthias Fellenberg, Manager, Bioinformatics Projects, Biomax Informatics AG
Gene expression profiles and protein-protein interactions are the most prominent examples for large data sets now produced with high-throughput approaches. Numerical and statistical evaluations of these data sets are well developed but do not necessarily reveal the biologically significant signals that are hidden in the data. We have developed an integrative approach to the analysis of high-throughput data sets beyond statistics. This approach aims at revealing biological causalities by integrating biological information with the large data sets. The employed biological annotations can be functional categories using the Biomax FunCat™ functional catalog, textbook metabolic pathways, protein-protein interactions, or systematic annotations to categories defined by the user. The Biomax Expression Analysis Suite is capable of visualizing the distribution of genes belonging to a specific functional category over the gene clusters. Groups of co-expressed genes with similar biological roles can be automatically identified. Co-expressed metabolic pathways are visually highlighted, marking those genes of a pathway that are not present in the respective gene clusters. Hypothetical metabolic networks can be dynamically calculated from a group of co-expressed genes. The methods employed for the integrative approach are part of a framework that enables the flexible definition of analysis workflows.

16.00 Use of Microarray Analysis Software as a Novel Tool to Visualize Comparative Genomics Data
Dr. Andre Nantel, Research Officer, Biotechnology Research Institute, National Research Council of Canada
During the annotation of newly sequenced genomes, gene functions are generally predicted based on sequence similarities with other genes. Currently there are no simple bioinformatics tools available for the simultaneous visualization of comparative genomic results for thousands of genes over multiple genomes. We demonstrate that software used in the visualization of transcriptional profiling experiments can also be used in comparative genomics and applied this tool to a comparison of the budding yeast proteome to eight prokaryotic and eukaryotic genomes. We will present examples of the use of this concept from our ongoing studies on cancer and the pathogenesis of Candida albicans.

16.30 Panel Discussion

17.00 Close of Conference

Corporate Sponsor Biographies

Gene Logic Inc. is a leading genomics-based biocontent and bioinformatics company focused on developing information products related to gene activity in human disease and toxicity to optimize rapid, reliable and cost-effective pharmaceutical and biotechnology research and development. Through its expertise in biosamples, high-throughput genomics technologies and software development, the Company has developed a series of gene expression information solutions based on its core product, the GeneExpress® Suite.

Iobion Informatics presents GeneTraffic software for two-color microarray data management and analysis. GeneTraffic software allows you to access data and projects on a desktop PC, from any location within your Network, manage data, perform computational analyses and query your data. With GeneTraffic software you can qualify and validate your microarray data prior to biological analysis.

Iobion Informatics is a Delaware LLC, headquartered in La Jolla, CA with offices in Toronto, Canada, and Austin, Texas.

Hotel Information

Swissôtel Zurich Am Marktplatz Oerlikon CH-8050 Zurich, Switzerland T: 41-1-317-3111 F: 41-1-312-3425 Room Rate: 270 CHF S/D Cut-off Date: December 20, 2002 Please call the hotel directly to make your room reservation. Identify yourself as a Cambridge Healthtech Institute conference attendee to receive the reduced room rate. Reservations made after the cut-off date or after the group room block has been filled (whichever comes first) will be accepted on a space-and-rate-availability basis. Rooms are limited, so please book early.

Call for Sponsors and Exhibitors

This exciting event will cover many aspects of microarray data analysis as well as latest advances and innovation in miniaturized platform technologies. Many sponsorship opportunities are available for your company to maximize its exposure and influence—including overall event and conference specific sponsorships as well as technology workshops, networking receptions, delegate bags, and badge lanyards. Over the course of the week, over 400 delegates will have access to the exhibit hall. The early rate for exhibit spaces is October 25, 2002; exhibit registrations received by that date will save your company up to $675! For more information on available sponsorship packages and exhibit space, please contact Angela Parsons at 781-972-5467 aparsons@healthtech.com or Deborah Brooks at 781-972-5412 dbrooks@healthtech.com.

Call for Posters

Cambridge Healthtech Institute encourages attendees to gain further exposure by presenting their work in the poster sessions. Please fill out the registration form, with the poster title and primary author. To ensure inclusion in the conference CD, a one-page summary must be submitted and registration must be paid in full by January 3, 2003. Click here for poster instructions

Special ski packages are offered for conference attendees at Davos and Interlaken. Booking Cut-off Date: December 20, 2002 For additional information please contact: Chips Lindenmeyr, Lindenmeyr Travel, Ltd. - "Skiing as it ought to be!" T: 212-725-2807 or 800-248-2807, F: 212-779-2239, www.lindenmeyrtravel.com

 

CHI Home   |  Conferences   |  Exhibits  |  Sponsorship  |  Request Info  |  CD Orders  |  Privacy Policy

---

Phone: 617-630-1300, Fax:  617-630-1325
Email: chi@healthtech.com