|
Activity Number:
|
245
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #309867 |
|
Title:
|
Likelihood-Based Procedures for Disease Gene Localization with General Pedigree Data
|
|
Author(s):
|
Shuyan Wan*+ and Shili Lin
|
|
Companies:
|
Merck & Co., Inc. and The Ohio State University
|
|
Address:
|
126 E Lincoln Ave, Rahway, NJ, 07065,
|
|
Keywords:
|
disease gene localization ; confidence interval method ; model averaging ; importance sampling ; linkage analysis
|
|
Abstract:
|
We proposed a confidence interval method for disease gene localization by testing every position on each chromosome of interest for its possibility of being a disease locus and including those not rejected into the interval. Two generalized likelihood ratio tests with or without model averaging (GLRT/MA and GLRT) were proposed to perform the test. Null distribution based on GLRT was estimated by importance sampling method. We also proposed asymptotic approaches based on both GLRT and GLRT/MA as alternatives that are much more efficient computationally but depends on the reliability of the limiting distributions. Besides its efficiency, the asymptotic procedure based on GLRT/MA also takes model uncertainty into consideration. Performance of various methods was compared by ROC-like curves based on both simulated data and a real data set from Genetic Analysis Workshop.
|
