JSM 2004 - Toronto

Abstract #301800

This is the preliminary program for the 2004 Joint Statistical Meetings in Toronto, Canada. Currently included in this program is the "technical" program, schedule of invited, topic contributed, regular contributed and poster sessions; Continuing Education courses (August 7-10, 2004); and Committee and Business Meetings. This on-line program will be updated frequently to reflect the most current revisions.

To View the Program:
You may choose to view all activities of the program or just parts of it at any one time. All activities are arranged by date and time.

The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.

Back to main JSM 2004 Program page
Activity Number: 200
Type: Contributed
Date/Time: Tuesday, August 10, 2004 : 9:00 AM to 10:50 AM
Sponsor: Biometrics Section
Abstract - #301800
Title: Thick-tailed and Heteroscedastic Linear Models for the Analysis of cDNA Microarray Data
Author(s): Juan P. Steibel*+ and Guilherme J.M. Rosa and Robert J. Tempelman
Companies: Michigan State University and Michigan State University and Michigan State University
Address: 1261 Anthony Hall, East Lansing, MI, 48824,
Keywords: microarray ; heteroscedasticity ; thick tails ; MCMC
Abstract:

A number of model specifications were compared for the analysis of a cDNA microarray. The dataset consisted of a dye swap experiment involving two treatments and two arrays. A total of 76 genes were spotted four times and two genes were spotted 16 times. A Bayesian approach was used to fit the following models: (1) linear model with homoscedastic residuals across genes, (2) linear model with independent heteroscedastic residuals across genes, (3) linear model with exchangeable heteroscedastic residuals across genes, (4) a linear model with homoscedastic student-t distributed residuals, (5) linear model with student-t exchangeable residuals. This approach allowed not only the estimation of treatment effects for each gene, but also the posterior distribution for the ranking of genes according to these effects. Results were compared on the basis of predicted treatment effects, size of the credibility intervals and rankings of genes. All models yielded similar point estimates for treatment effects, but with varying width of the credibility intervals. The most discordant results were obtained between the independent heteroscedastic and homoscedastic residual specifications.

  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2004 program

JSM 2004 For information, contact jsm@amstat.org or phone (888) 231-3473. If you have questions about the Continuing Education program, please contact the Education Department.
Revised March 2004