Day 1
Day 2
Day 3
Short Courses
Hotel & Travel
PDF Download
Press Pass 
Request Brochure 



Agilent Technologies

Biorad labs logo

Biotrove Logo

Fluidigm logo

Bio-IT World

The Scientist

Bio Array News

drug discovery news

Genome Technology



Genome Web

SelectScience logo


Integrative Data Analysis - Day 2

Conference Proceeding CD Now Available
  • Speaker Presentations
  • Poster Abstracts
  • and More!


(Formerly Microarray Data Analysis)


7:30 am Breakfast Technology Workshop (Sponsorship Available)


8:30 Chairperson’s Opening Remarks

8:40 Keynote Presentation
Massively Parallel High Throughput DNA Sequencing: Automation for Microbial Community, Gene Expression and de Novo Deciphering of New Genomes
Bruce A. Roe, Ph.D., George Lynn Cross Research Professor of Chemistry and Biochemistry, Advanced Center for Genome Technology, Stephenson Research and Technology Center, University of Oklahoma
This presentation discusses the modifications Dr. Roe’s laboratory has introduced to streamline the up-front sample preparation protocols, to improve sample read length and to automate the library and sample preparation steps prior to DNA sequencing.   The results obtained from applying these techniques to study microbial communities, eukaryotic gene expression and the de novo sequencing of new genomes will be discussed and compared to those obtained using more classical approaches.
9:25 Choice of Normalization Algorithm Affects Consistency Between Microarray and RT-PCR in Clinical Samples
Balazs Gyorffy,  M.D.,  Ph.D.,  Semmelweis University,  Budapest
One of the main prerequisites of building reliable, clinically applicable gene expression based classifiers is the accurate, or at least consistent, measurement of gene expression levels. However, more than 10 years after the inception of microarray technology, it is still unclear how robustly we can quantify gene expression in clinical samples. While several reassuring studies were published showing high levels of platform consistency and accuracy of microarray-based gene expression profiling, all these studies were based on highly artificial RNA samples, which do not reflect the true heterogeneity of RNA extracted in a real clinical setting. In order to create a benchmark data set, we have microarray profiled 36 colon cancer samples using the Affymetrix HG-U133 plus 2.0 gene chip, along with the independent quantification of close to one hundred genes by TaqMan arrays. We have determined which of the widely used microarray normalization algorithms (RMA, GCRMA, dCHIP, MAS5) produces the most consistent gene expression estimates with QRT-PCR measurements.

9:55 Leveraging Comparative Genomics to Infer-Functional Consequences of Gene Expression Changes: Accelerating Microarray Discovery via Phenomics Annotation
Kristopher Irizarry, Ph.D., Assistant Professor, College of Veterinary Medicine, Western University of Health Sciences
Methods of enhancing the functional annotation of genes can provide tremendous insight into mechanisms underlying clinically relevant phenotypes. The addition of novel types of functional annotation can increase the value of microarray experiments by increasing the biological signal in the data. Using phenotypic information to cluster and mine gene expression data offers an additional level of annotation with direct implications for interpreting gene expression changes which can accelerate discovery in the post-genomics era.

10:20 Morning Coffee, Poster and Exhibit Viewing


11:00 Molecular Classification of Gliomas: Let the Data Inform Treatment 
Jean-Claude Zenklusen, M.S., Ph.D., Staff Scientist, Neuro-Oncology Branch, National Cancer Institute
Although there has been quite a number of reports using genomic technology to help in the classification of tumors from a molecular perspective, all of these attempts have been plagued by the use of preconceived variables (either histopathological classes, survival, time to recurrence, etc.) to obtain the initial groupings, instead of allowing the large amount of data generated to guide the class determination.  Our approach to allow the expression profiles to determine the classes coupled with validation of the targets found makes for a more discrete, rationally selected geneset that helps focus the research for possible therapeutic targets.

11:30 Removing the Hurdles From the Path Towards Personalized Cancer Therapy
Zoltan Szallasi, M.D., Children’s Hospital Informatics Program, Boston and Technical University of Denmark, Lyngby
Genome-scale analysis by microarrays or array CGH are expected to yield information that would enable clinical oncologists to select the most efficient therapy for a given cancer patient. However, the success of such an endeavor is highly dependent on several factors, including the general noise structure and noise level of high-throughput measurements, the strength of association between biomarkers and gene modules and clinical outcome and the rather unfavorable ratio between the number of alternative hypotheses (e.g. quantified genes) and available clinical samples. We will address two important issues deeply rooted in the high-throughput nature of genome scale profiling and highly relevant for the meaningful analysis of clinical microarray data: systematic bias in clinical microarray data and extracting robust, convergent and clinically useful information from multiple breast cancer data sets. We will also provide evidence from a clinical cohort that an appropriately selected, biologically motivated robust gene expression signature can determine which of two widely used chemotherapeutic agents will be more effective for a given ovarian cancer patient.

12:00 pm Close of Morning Session

12:15 Luncheon Technology Workshop
 Sponsored by Agilent Technologies 


2:00 Chairperson’s Opening Remarks

2:05 Gene Expression Analysis Employing All Reported Protein-Protein Interactions Reveals Abnormal IFN-Beta Expression in Therapy-Nãive Relapsing-Remitting Multiple Sclerosis
Hugh Salamon, Ph.D., Founder, AbaSci, LLC 
An approach to defining gene sets defined by protein interactions is introduced. These gene sets in turn are utilized in an analysis of gene expression profiles of mononuclear cells of both therapy-nãive and interferon-beta-treated multiple sclerosis patients. A statistical model provides strength of effect and rejection statistics that highlight abnormal expression in therapy-nãive patients. The general approach to mechanistic dissection of expression changes in disease and treatment can be applied more generally, with an example from
in vitro cancer chemotherapeutic resistance briefly presented.

2:35 A Search of Public Gene Expression Microarray Data Identifies Novel Agents That Selectively Eradicate Leukemic Stem Cells
Duane Hassane, Ph.D., Post-doctoral Research Associate, James P. Wilmot Cancer Center, University of Rochester School of Medicine and Dentistry
Cancer stem cells initiate and perpetuate disease in a variety of malignancies, including acute myelogenous leukemia. These cells are resistant to traditional chemotherapeutics and we have thus found a novel agent, parthenolide, that selectively eradicates the leukemic stem cell. However, parthenolide demonstrates limited bioavailability and does not represent an ideal therapeutic. Using similarity measures and machine learning approaches, we demonstrate the application of a “search agent” to identify chemically diverse parthenolide-like compounds  from the collective multi-center microarray gene expression data that is found in the Gene Expression Omnibus.

3:05 Technology Spotlight Sponsored by Agilent Technologies 
Analysis and Visualization of Heterogeneous Data in GeneSpring GX
Pam Tangvoranuntakul, Ph.D., GeneSpring Product Manager, Agilent Technologies, Inc.
As multi-assay studies become more prevalent in Genomic research, the need for software applications that support integrative data analysis becomes critical to the discovery of linkages and concordance between different data types such as chromosomal copy number, miRNA, and gene expression.  This session will focus on tools in GeneSpring GX that will allow researchers to make these discoveries.  Methods of integrative data analysis in GeneSpring GX will be demonstrated by comparing miRNA with gene expression changes.

3:20 Refreshment Break/Poster and Exhibit Viewing (last chance for viewing)

4:00 Genomic Diversity, Population Structure, and Signatures of Recent Selection in a Genome-Wide Study of 3,796 Humans
Adam Auton, Ph.D., Post-Doc, Department of Biological Statistics and Computational Biology, Cornell University  
We have analyzed a survey of 3,796 individuals from four distinct continental populations. Individuals were successfully genotyped at approximately 457,000 Single Nucleotide Polymorphisms (SNPs). We identified both global and sub-continental patterns of population structure, including very fine-scale detail within Europe. Furthermore, we observed significant differences in patterns of homozygosity between both individuals and populations-suggestive of differences in recent common ancestry and of selective pressures.

4:30 Capturing Exons with Synthetic Oligonucleotide Baits for Next-Generation Sequencing 
Carsten Russ, Ph.D., Research Scientist, Genome Sequencing and Analysis Program, Broad Institute of MIT and Harvard
New technology has substantially reduced the cost of DNA sequencing, but a significant cost efficiency can still be gained by sequencing subsets of genomes, such as protein-coding exons. We have developed an efficient, cost-effective, highly scalable and highly-multiplexed method to capture targets for sequencing from genomic DNA. We have applied this method to study variation in cancer samples and in association with human genetic disease. We will describe a key new method that greatly expands the capabilities of new sequencing technology. The method has general application in studies of human genetics and cancer, examples of which we will describe.

5:00 Discovery of Useful Epigenetic Biomarkers using MethylPlex Libraries and Genome-Wide Profiling Tools: Cross-Platform Comparison
Vladimir Makarov, Ph.D., Chief Scientific Officer, Rubicon Genomics
Promoter hyper- and hypo-methylation is a common epigenetic event in many human cancers and it can be used as a biomarker for cancer detection and prognosis, or for patient stratification to predict and/or monitor the efficacy of cancer therapy. MethylPlexTM is a novel enzymatic approach to study DNA methylation on a genome-wide scale. Analysis of MethylPlexTM libraries by qPCR proved to be very sensitive and allows a detection of ~2-10 methylated DNA copies per sample and analysis of multiple promoter regions from a single DNA sample. As a next step we developed a simple and robust protocol for obtaining MethylPlex-PlusTM DNA libraries with a greatly reduced complexity and ~ 100-1,000-fold enrichment for DNA sequences that are differentially methylated in cancer cells and analyzed these libraries using three analytical platforms: next-generation sequencing (Solexa), CpG microarray profiling (Agilent), and qPCR. We will discuss application of the genome-wide DNA methylation analysis for discovery of novel biomarkers by comparing these analytical platforms and 3 different DNA sources – frozen tissue, FFPE tissue, and serum (plasma).

5:30 ChIP-Seq Based Transcription Regulation Mapping of Human Liver and Data Validation by ChIP-DSL & ChIP-qPCR
Jeffrey Falk, Ph.D., Director of Technology Applications, Molecular Biology, Aviva Systems Biology
Studies will be described that illustrate the integrated power of using multiple ChIP (Chromatin Immunoprecipitation) – based technologies for understanding tissue specific gene transcription regulation. In these studies, next-generation sequencing-based ChIP-Seq was used to analyze various histone markers in normal human liver. These results were combined with ChIP-on-ChIp screening using an automated system to analyze over 200 specific transcription factors. Specific gene loci demonstrating factor binding in these studies were subsequently validated by ChIP-qPCR studies that combined nuclei receptor and co-activator interactions. These studies clearly demonstrate how the parallel development of ChIP - based tools (ChIP-Seq, ChIP-DSL, and ChIP-qPCR ) can be used to effectively fingerprint changes in transcriptional networks associated with various human diseases.

6:00 Close of Day

By Series:
By Region:

& Course Catalog

CHI Catalog 

Short Course DVDs