|
Activity Number:
|
341
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #300224 |
|
Title:
|
Bayesian Models for Repeatedly Repeated Data
|
|
Author(s):
|
Gary L. Rosner*+ and Peter Mueller and Fernando A. Quintana
|
|
Companies:
|
The University of Texas M.D. Anderson Cancer Center and The University of Texas M.D. Anderson Cancer Center and Pontifícia Universidad Católica de Chile
|
|
Address:
|
1515 Holcombe Boulevard, Houston, TX, 77030,
|
|
Keywords:
|
Bayesian inference ; Nonparametric models ; Loss of heterozygosity ; Repeated measures ; Binary sequences ; Hierarchical models
|
|
Abstract:
|
We discuss inference for data with repeated measurements at many levels. Inference typically concerns characteristics of the repeated measurements within repeating cycles or dependence across cycles or both. We illustrate modeling and inference with an example relating to alterations of patients' DNA represented by sequences of indicators of loss of heterozygosity. The data involve three nested levels of repetition for each patient: chromosomes, regions within chromosomes, and single nucleotide polymorphisms nested within regions. Our Bayesian semiparametric hierarchical model for these multi-level repeated binary data includes a mixture-of-Markov-chains model, defined with respect to the Markov transition probabilities and a nonparametric prior for the random mixing measure.
|
