|
Activity Number:
|
436
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Learning and Data Mining
|
| Abstract - #305676 |
|
Title:
|
A Comparison of Logic Regression and CART for Identifying Important SNP Interactions
|
|
Author(s):
|
Wonsuk Yoo*+ and Michele Cote and Ann Schwartz
|
|
Companies:
|
Wayne State School of Medicine and Karmanos Cancer Institute/Wayne State University and Karmanos Cancer Institute
|
|
Address:
|
4201 St Antoine St., UHC-4H, Detroit, MI, 48201,
|
|
Keywords:
|
lung cancer ; multigenic study ; logic regression ; CART
|
|
Abstract:
|
Lung cancer is the most leading cause of death in women in the United States, and still remains most dangerous threat to women life regardless of development in early detection and medical techniques increases for the chance of full recovery. Lung cancer is a multigenic disease which is more likely associated with the combined effects of multiple genes, not a single gene effect. In this study, we provide a comprehensive comparison of two classification methods between logic regression (Schwender et al, 2007) and CART (BreWang et al, 2007) to identify the associations of a comprehensive panel of genetic polymorphisms as well as demographic and environmental characteristics with lung cancer risk in white ever smoking womes. An ad-hoc analysis are applied to compare with the result of multiplicative interactions identified by two classification tree methods.
|
