 |
|||||||||
|
NMBIS Speakers
Atul Butte, M.D., Ph.D. Keynote Address: Exploring Genomic Medicine Using Integrative Biology Abstract: The past 10 years have led to a variety of measurements tools in molecular biology that are nearly-comprehensive in nature. For example, microarrays are just one of at least 30 large-scale measurement or experimental modalities available to investigators in molecular biology. Instead of focusing on the cell, or the genotype, or on any single measurement modality, using integrative biology allows us to think holistically and horizontally. A disease like diabetes can lead to myocardial infarction, nephropathy, and neuropathy; to study diabetes in genomic medicine would require reasoning from a disease to all its various complications to the genome and back. To enable such research, we have been studying the process of intersecting genome-scale data sets in molecular biology, such as those from genome scans, microarrays, proteomics, RNAi, and many others. I will show how we have built computational tools that reason over these types of data to help enable discoveries in genomic medicine, with specific applications for obesity and diabetes mellitus. Though standards are increasingly being required and used for genome-scale data, representing the experimental context using a structured vocabulary has not yet happened, yet is a crucial set towards automated integrative biology. I will show how the largest unified biomedical vocabulary can now be used to represent microarray sample annotations and show examples of visualization, searching, and analysis using this coding that could not have been done before. I will end with a consideration of ways we can use genome-scale data to provide new ways to classify disease, and show how this broad recasting of disease nosology allows identification of new therapeutic opportunities, and of the specificity, or lack thereof, of disease biomarkers. Bio: Atul Butte, M.D., Ph.D. is an Assistant Professor in Medicine (Medical Informatics) and Pediatrics at the Stanford University School of Medicine, and a board-certified pediatric endocrinologist. Dr. Butte received his undergraduate degree in Computer Science from Brown University in 1991, and worked in several stints as a software engineer at Apple Computer (on the System 7 team) and Microsoft Corporation (on the Excel team). He graduated from the Brown University School of Medicine in 1995, during which he worked as a research fellow at NIDDK through the Howard Hughes/NIH Research Scholars Program. He completed his residency in Pediatrics and Fellowship in Pediatric Endocrinology in 2001, both at Children's Hospital, Boston. Dr. Butte received a Ph.D. in Health Sciences and Technology from the Medical Engineering / Medical Physics Program in the Division of Health Sciences and Technology, at Harvard Medical School and Massachusetts Institute of Technology. Dr. Butte's laboratory focuses on solving problems relevant to genomic medicine by developing new biomedical-informatics methodologies in integrative biology. Dr. Butte has authored more than 25 publications in bioinformatics, medical informatics, and molecular diabetes and has delivered more than 30 presentations world-wide on bioinformatics, including nine at the National Institutes of Health or NIH-sponsored meetings. Dr. Butte's recent awards include the 2006 PhRMA Foundation Research Starter Grant, the 2003 Emory University School of Medicine Pathology Residents' Choice Award, the 2002 and 2003 American Association for Clinical Chemistry Outstanding Speaker Award, the 2002 Endocrine Society Travel Award based on presentation merit, the 2001 American Association for Cancer Research Scholar-In-Training Award and the 2001 Lawson Wilkins Pediatric Endocrine Society Clinical Scholar Award. Along with Isaac Kohane and Alvin Kho, Dr. Butte has co-authored one of the first books on microarray analysis titled "Microarrays for an Integrative Genomics" published by MIT Press.
Jason Gans, Ph.D. Talk: Development of a multi-pathogen and backgrounds microarray for environmental monitoring. N. Doggett, R. Leach, R. Cary, M. Doyal, P. Pardington, C. Stubben, J. Song, M. Wolinsky, K. Creek and J. Gans. Abstract: We are developing a microarray for detection and identification of a broad range of pathogens and backgrounds in the environment. Our approach involves the polymerase chain reaction (PCR) to selectively amplify at predetermined sites and the use of primer extension microarrays to type phylogenetic single nucleotide polymorphisms (SNPs) within amplified targets. A key feature of this approach is the selection of a minimal set of PCR oligonucleotide primers, which are selected through by comparative genomic and computational approaches to amplify only pathogens and the use of a collection of conserved gene targets for the selection of PCR primers to amplify broadly across pathogens, their near neighbors and other background species. We have demonstrated the utility of this approach with the design and implementation of a multi-pathogen microarray which can detect and differentiate among 41 different species (total of 61 strains) of bacterial pathogens following amplification with only 15 pathogen specific PCR primers pairs. This multi-pathogen microarray contains 625 probes which type phylogenetically selected SNPs within amplimers to provide robust and redundant discrimination among 41 different species (61 different strains) of bacterial pathogens. The selection of phylogenetic SNPs for typing is made both at leaf and branch nodes on the tree thereby permitting not only the identification of previously sequenced bacterial species but also enabling detection of unsequenced species at their appropriate branch nodes. The design of the background/near neighbor component of the array permits detection and identification of over 570 taxa using 48 PCR primers, or over 450 taxa using only 20 PCR primers. This new array could prove useful as a first pass for detection of pathogen and neighbor species for environmental monitoring applications.
Steve Haase, Ph.D. Talk: Identifying Cdk-Independent Cell Cycle Transcription Networks Using Microarrays Abstract: It is well established that transcriptional control is fundamental to the proper regulation of cell cycle events in most cells. But in early embryos, periodic cell cycle events can be triggered even in the absence of active transcription. The biochemical oscillator controlling these cyclic events is centered on the activity of cyclin-dependent kinases (Cdks). Cdks are also thought to act as the central cell cycle oscillator in somatic cells and yeast. Collectively, cyclin/Cdk oscillations and checkpoint mechanisms can account for much of the observed regulation of the cell cycle, including regulation of periodic transcription. However, we have previously shown in budding yeast that a small subset of genes normally expressed in G1 continue to be periodically transcribed, on schedule, in B-cyclin mutants that arrest at the G1/S border. These findings suggested that the mechanisms regulating periodic activation of transcription might function independent of B-cyclin/Cdks and cell cycle progression. We have analyzed genome-wide transcription over time in synchronized populations of B-cyclin mutant cells using microarrays. Our preliminary analysis revealed that genes normally transcribed in G1, S-phase, and mitosis, continue to be activated on schedule even though B-cyclin mutant cells cannot progress into S-phase or mitosis. These results indicate that a substantial portion of the cell cycle-regulated transcription network remains intact in cells lacking B-cyclin genes. Although cyclin/Cdks can regulate transcription, our findings support a model in which the cell cycle-regulated transcription network can function, at least in part, independent of B-cyclins and cell cycle progression. We are currently constructing transcription network models based on microarray data to determine if periodic activity could be an emergent property of the transcription network. Bio: Steve Haase earned a B.S. in Biochemistry from Colorado State University, and a Ph.D. from the Department of Genetics at Stanford University. Dr. Haase pursued his postdoctoral work at The Scripps Research Institute in La Jolla, California. He was awarded a Leukemia Society of America Fellowship in 1994, and received a FASEB Young Investigator Award in 1996. Dr. Haase is currently an Assistant Professor in the DCMB Group, Department of Biology at Duke University, and is a member of the Duke Comprehensive Cancer Center, the University Program in Genetics and Genomics, and the Computational Biology and Bioinformatics Program. He is also an active member of the Biological Networks Group at Duke. The Haase Lab is investigating the molecular mechanisms that regulate cell division using budding yeast as a model system.
Brendan Mumey, Ph.D. Talk: Discovering Genes and Classes in Microarray Data Using Island Counts. Joint work with Louise Showe and Michael Showe of the Wistar Institute, Philadelphia, PA. Abstract: We consider the problem of unsupervised class discovery from DNA microarray data and present a new algorithm to simultaneously discover tissue classes and identify a set of genes that well-characterize these classes. We employ a combinatorial optimization approach where the object is to simultaneously identify an interesting set of genes and a partition of the array samples that optimizes a certain scoring function based on a novel statistic called island counts. While we show that the underlying optimization problem is NP-complete, in many instances we can solve problems of interest to optimality using a branch-and-bound algorithm. We have tested the algorithm on a 30 sample Cutaneous T-cell Lymphoma data set; it was able to almost perfectly discriminate short-term survivors from long-term survivors and normal controls. Another useful feature of the method is that can handle missing expression data. Bio: Brendan Mumey is an associate professor of Computer Science at Montana State University, Bozeman. His research interests are in algorithm design and computational biology. He earned a B.Sc. in Mathematics at the University of Alberta, a M.Sc. in Computer Science at the University of British Columbia and a Ph.D. in Computer Science at the University of Washington. He has published over 20 papers and has been awarded one patent.
Maggie Werner-Washburne, Ph.D. Talk: Microarray analysis for biological processes in yeast Abstract: We have used microarrays to study various processes related to stationary phase (SP) and quiescence in yeast. We view microarrays as a systems engineering process and abide by the rules of "design, control, or randomize." During this work, we have contributed, with colleagues from Sandia National Labs and elsewhere, to the development of the hyperspectral imaging scanner, an automated sampler, and various programs for microarray analysis. These tools have been used to analyze both time-course and single time-point data. The three major datasets we have produced will be discussed. Time-course data of exit from SP led to the identification of 127 transcripts abundant in cells from SP cultures. From this we identified 32 novel SP mutants that affect viability during different times during entry into SP and more than 1000 transcripts whose abundance increased within 5 minutes after glucose addition. Further characterization of stress responses in cells from SP cultures identified more than 2000 transcripts held in a protease-labile, phenol-insoluble form in these cells and released in a stress-specific manner. Recently, our ability to isolate quiescent and non-quiescent cells from SP cultures allowed us to complete a single time-point analysis, including almost 400 arrays of mutant and parental cells. Transcript abundance is consistent with physiological differences determined for these cells. We are in the process of carrying out additional experiments based on computational analysis of this data. Future work and the implications of these studies will be presented. Bio: Maggie Werner-Washburne has worked in the area of stress response and stationary phase in yeast for most of her career. Her contributions include the discovery that HSP70 proteins are chaperones, establishing the research area of yeast stationary phase, and, most recently, using microarrays to discover the sequestration of thousands of mRNAs in cells from SP cultures and the identification and characterization of quiescent and non-quiescent/apoptotic cells from yeast SP cultures, including finding a true G0 state in yeast. Her approaches have included biochemistry, molecular biology and genetics, and, in the past 8 years, genomics and computational biology. She collaborates with a broad range of researchers at UNM, Dalhousie University, the University of Graz, Sandia National Laboratories, and elsewhere. The work in her laboratory has never been limited by ideas and excellent post docs and graduate students are always encouraged to apply. NMBIS Workshop Speakers
George Davidson, M.S. Workshop: Knowledge Mining With VxInsight Abstract: Large collections of microarray experiments offer the opportunity to identify genes with similar expression patterns or, alternately, to cluster the arrays into groups by overall similarities. The first kind of clustering is useful when seeking to identify possible gene functions for unannotated genes. In this case, one hopes to find known genes with expression patterns very similar to that of the unannoted gene, which may suggest a like function. The second kind of clustering can be useful when trying to understand which genes may be important with respect to a particular biological state. This approach could be used to try to understand which genes are important for cancer patient outcomes. This workshop will examine both types of clusterings using VxInsight, a visual data mining tool. In going beyond clustering, one may wish to identify the genes that have different expression patterns between groups, say for example cancer patients with good or bad outcomes. We will examine how to use analysis of variance to rank genes with respect to different clusters, or patient outcomes. We will see how to use bootstrapping to see if the gene lists are robust to slight perturbations in the data, and will, then, briefly examine the really difficult next step in analysis: using the identified genes, the available databases and the published literature to interpret the results. Bio: George Davidson has been involved in data analysis and knowledge visualization for most of his career at Sandia National Laboratories, which began with the developing data collection and analysis environments for the huge fossil energy experiments in the 1970's. More recently, he managed the visualization group that developed virtual reality systems and applied them to the analysis of large-scale supercomputer computations. One of the tools developed by that group is the visual data mining tool VxInsight, which was originally used to look at the structure of science as revealed by citations graphs assembled from published papers. VxInsight, also, proved useful in analyzing large collections of microarray experiments, and has been the basis of joint work with researchers from across the country and here in New Mexico. His current interests include (1) exploring the similarities between bioinformatics and the informatics needs of the intelligence community, and (2) developing knowledge management tools for team working with biological data.
Gavin Pickett, Ph.D. Workshop: New Genomic Tools and Their Impact on Experimental Design Abstract: Advances in technology and new insights into mechanisms of genomic expression are fostering a new age in experimental design. Now researchers have many more microarrary choices when it incomes to exploring genomics. When the research path leads into the realm of Genomics, the first question of experimental design is now: "What are the appropriate tools to investigate my particular area of research?" There are now more kinds of microarrays than we could have imagined 5 years ago. This makes experimental design more difficult but much more interesting scientifically. This talk will focus on some of the new microarray technologies that are available including alternative-splicing, tiling, and SNP analysis. Bio: Gavin Pickett is the Technical/Scientific Director of the Keck-UNM Genomics Resource (KUGR) facility at the University of New Mexico Health Science. He grew up here in New Mexico and received his B.S., M.S. and Ph.D. in Molecular Biology from UNM in 1992. After leaving New Mexico for Post-Doctoral training at Duke, he returned in 1995 to help start Phase-1 Molecular Toxicology in Santa Fe. It was here he started designing and printing microarrays for use in gene-expression studies focused on phase-1 toxicology studies with Astra-Zeneca, Glaxo-Wellcome and the FDA. He has been at UNM in his current position for 5 years.
Faye D. Schilkey Workshop: Online Resources for Analyzing Microarray Data Abstract: Once data has been generated from a microarray experiment, what online analysis resources are available to the researcher? How do these compare to client-side tools? What data-formats are required by these resources? This workshop focuses on introducing and presenting key-features of various online analysis resources through interactive lab work. Online resources such as NCBI's Gene Expression Omnibus and EBI's ArrayExpress will be covered and compared to client side tools such as TIGR's MeV. MIAME compliance and formats such as MAGE-ML will be covered. Participants may suggest resources to cover by emailing me at fds@ncgr.org by 3/17/06. Bio: Ms. Schilkey received her B.S. degree in computer engineering in 1986 from Oakland University in Rochester, Michigan. Working at NCGR since 1996, she has overseen and worked on several database and software projects including GSDB, ISYS, XGI, PathDB, TAIR, GEYSIR and INBRE. Her focus is in Agile project management (Scrum), object oriented programming and unit testing to yield on-target, extendable software, in an iterative and timely fashion. Before coming to NCGR, she worked for 10 years in the areas of real-time embedded software systems for military applications (General Dynamics), and network systems (Los Alamos Technical Associates). |
||||||||
