|
Activity Number:
|
350
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #303156 |
|
Title:
|
Model-Based Clustering of Array CGH Profiles: A Recursive-Partitioning Algorithm for Wavelet Decompositions
|
|
Author(s):
|
David Engler*+ and Brent Coull and Rebecca Betensky and E. Andres Houseman
|
|
Companies:
|
Brigham Young University and Harvard University and Harvard School of Public Health and University of Massachusetts Lowell
|
|
Address:
|
223 TMCB, Provo, UT, 84602,
|
|
Keywords:
|
array CGH ; wavelet decomposition ; hierarchical clustering
|
|
Abstract:
|
Array-based comparative genomic hybridization (aCGH) allows for the assessment of DNA sequence copy number changes in tumor tissue. A frequent objective of interest is the identification of clinically-relevant subgroups that share common aCGH profiles. Current clustering methods often fail to make efficient use of the underlying structure of the aCGH data or, alternatively, can suffer from substantial information loss. Moreover, identification of the appropriate number of clusters is not straightforward under current approaches. We present a novel, computationally-efficient approach for the clustering of aCGH profiles. Clustering is conducted through a model-based recursive partitioning algorithm applied to the wavelet decompositions of aCGH profiles. Identification of relevant clusters is obtained through inference on the resulting nested classes.
|
