|
Activity Number:
|
110
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #304680 |
|
Title:
|
Power Computation in Testing Hardy-Weinberg Equilibrium
|
|
Author(s):
|
Marepalli B. Rao*+ and Subramaniam Venkatesan and Subramanyam Kasala
|
|
Companies:
|
University of Cincinnati and University of Cincinnati and The University of North Carolina at Wilmington
|
|
Address:
|
Center for Genome Information, Cincinnati, OH, 45267,
|
|
Keywords:
|
Hardy-Weinberg Equilibrium ; Convex Set ; Extreme Point ; Power Function
|
|
Abstract:
|
Testing whether or not a population is in Hardy-Weinberg Equilibrium is a key step in Epidemiological and Genetical Studies. There are a number of tests available. Power function is used to assess the performance of any test and also for comparing the performance of any two tests. However, the power function is a function of two arguments and thus comparing the power functions of two tests is fraught with difficulties. We introduce a novel method which facilitates power computations and comparisons as easy as an apple pie. The key idea is the extreme point methodology inducted from Functional Analysis. A natural convex set is identified and its extreme points determined. The power function of a test thus becomes a convex combination of a fundamental set of probabilities associated with the test. Comparing two power functions is tantamount to comparing their fundamental sets.
|
