Research Reports

NEW TIME: Fridays from 3:15 - 4:20 in Foege Auditorium (S 060)

Our program will include reports on research progress and proposed work by graduate students.  Talks are planned for 25 minutes each.  If you are unable to speak at the appointed time, please arrange to change dates with someone else.  Please email Brian Giebel (bgiebel [at] u.washington.edu) with any changes.

Follow this link for a listing of past Research Reports topics.

SPRING 2008

May 9 -

Brig Mecham
"Identifying and controlling for sources of heterogeneity in genome- wide association studies"

Abstract:

Microarray technology has been used to measure variation in gene expression for a variety of diseases, including many types of cancer. As this technology becomes cheaper and users more comfortable with analysis of array data there has been a change in the emphasis on small studies consisting of two to six arrays, to larger studies consisting of hundreds to even thousands of samples in a single experiment. Intuition tells us that larger studies are more likely to suffer from spuriously correlated data as the sheer volume of samples requires arrays to be grouped with respect to date of processing, manufactoring lot (or array batch), and even labelling reagent. These factors are examples of Expression Heterogeneity: a general class of factors that confound associations between gene
expression variation and the biological quantity of interest. Controlling for EH is the motivation for all commonly used normalization procedures, including the lowess-dye bias method, quantile normalization, and VSN. In this talk I'll explore expression variation associated with tumor status in three panels of prostate cancer studies. The effects of EH on each study are investigated, and a novel method is described that can account for bias introduced by certain types of EH. The method is applied to these data and shown to considerably increase the consistency across these studies when compared to the results obtained from other common analysis techniques. We also apply this method to other studies that extend its relevance to multiple generations of Affymetrix platform as well as two-color assays.
This work is of direct interest to people involved in the analysis or interpretation of genomic studies using microarray technology. It is probably also of interest to people using the new Solexa data as similar models should apply here as well.

May 16 - reserved

May 23 - reserved

May 30 - First Year Rotation Talks:
Renee George, Sayer Herin, Ray Malfavon-Borja, Sarah Ng

June 6 - First Year Rotation Talks:
Alex Nord, Rupali Patwardhan, Dan Skelly, Cailyn Spurrell
slight time change: 3:00 - 4:00

June 13 - time reserved for K. Dhillon dissertation defense.