PSB 2002 Tutorial

PSB 2002

Advanced Statistical Analysis of Microarray Studies

Description

This tutorial will focus on analysis of RNA expression data derived from microarray images. Both oligonucleotide and spotted arrays will be discussed. The course is organized into ten sections:

1. Sources of variability in microarray studies (25 minutes);
2. Combining information on oligonucleotide arrays (20 minutes);
3. An ontology of inferential methods (10 minutes);
4. Identifying genes by testing hypotheses (40 minutes);
5. Mining data through clustering - options, benefits, limitations (10 minutes);
6. Dimension reduction through linear components models (20 minutes);
7. Latent class models (20 minutes);
8. Issues in experimental design (20 minutes);
9. Quantitation adjustments in microarray data analysis (20 minutes); and
10. Forthcoming analytic developments for microarray studies (15 minutes).

The instructor regularly conducts continuing education in this area for statisticians. This tutorial is geared towards quantitative analysts from varied backgrounds who are already substantially familiar with one or more microarray technologies. Attendees will become familiar with data analytic issues associated with microarray studies, and gain a basic understanding of application, benefits and limitations of available methods to analyze data from microarray studies, from a statistical perspective.

Biographical Sketch

Dr. Lazaridis is faculty in the Interdisciplinary Oncology Program of the University of South Florida (USF). He co-directs the activities of the Biostatistics Core of the H. Lee Moffitt Cancer Center and Research Institute at USF, where he also serves on the Scientific Review Committee. Dr. Lazaridis directs the USF Pilot Center for Mathematical-modeling of Image Data Across the Sciences (MIDAS), which was established in 1999. The mission of the MIDAS center is to create interdisciplinary collaborations among imaging, quantitative and biological scientists, in order to develop new analytic models of image-related data and to train researchers of various disciplines in modeling techniques. He joined the cancer center from Indiana University School of Medicine (IUSM), where he was affiliated with the Department of Medicine, the Regenstrief Institute for Health Care and the Herman B. Wells Center for Pediatric Research. He is a graduate of the statistics program of the University of Chicago.

Dr. Lazaridis has successfully competed for funding as Principal Investigator in studies of aging, cancer and statistical methods through the NIH, the American Cancer Society, and the Department of Defense. In addition he has developed substantial and productive collaborations with other medical researchers. He directs the research computing section of the Molecular Oncology Program Project (PO1-CA-82533), which seeks to develop more effective therapies for human cancer based on a better mechanistic understanding of tumor cell survival and drug resistance. Dr. Lazaridis participates in the Biomarker Development Laboratory at Moffitt (U01-CA-84973), where he develops technologies for biomarker testing in lung cancer. As co-PI he directs all biometrics research for the studies Molecular Fingerprint of STAT3 Regulated Genes for Early Detection of Human Cancer (DAMD-17-98-1-8659) and similarly as co-investigator for Decoding Fingerprints Portending Colon Cancer Metastasis (UO1-CA85052). Both of these studies employ spotted cDNA and oligonucleotide microarrays.

In past work, Dr. Lazaridis has collaborated with investigators in diabetes research (Diabetes Care, Journal of General Internal Medicine; Archives of Family Medicine), pharmacology (American Journal of the Medical Sciences), cancer research (Blood; Bone Marrow Transplantation) and aging (Journals of Gerontology: Medical Sciences; Facts and Research in Gerontology). In addition to his collaborative, analytic and scientific expertise, he has published articles on Bayesian models and applications (Statistics in Medicine; Communications in Statistics: Stochastic Models; Computer Methods and Programs in Biomedicine), and regarding innovations in instructional methods and theory (Academic Medicine; Journal of Statistics Education). Recent articles include "Linking Image Quantitation and Data Analysis in Modern Biological Experimentation", "Statistical Contributions to Molecular Biology", and "Introduction to Microarray Experimentation and Analysis."

His short course, "Statistical Methods in Molecular Biology: RNA and Protein Analyses", first sponsored by the American Statistical Association at the 2000 Joint Statistical Meetings, received rave reviews from the attendees, with 52% grading the course an overall A, and 34% an overall B. 64% of the statisticians attending the JSM short course rated his knowledge concerning the topics of instruction at the A level, and an additional 28% at a B level. Dr. Lazaridis regularly offers continuing education to molecular biologists, statisticians and other quantitative analysts in areas of microarray, proteomics, flow cytometry and other molecular biology technologies and their associated analytic methods.