A novel power-based approach to Gaussian kernel selection in the kernel-based association test

Xiang Zhan, Debashis Ghosh

Research output: Contribution to journalArticle

Abstract

Kernel-based association test (KAT) is a widely used tool in genetics association analysis. The performance of such a test depends on the choice of kernel. In this paper, we study the statistical power of a KAT using a Gaussian kernel. We explicitly develop a notion of analytical power function in this family of tests. We propose a novel approach to select the kernel so as to maximize the analytical power function of the test at a given test level (an upper bound on the probability of making a type I error). We assess some theoretical properties of our optimal estimator, and compare its performance with some similar existing alternatives using simulation studies. Neuroimaging data from an Alzheimer's disease study is also used to illustrate the proposed kernel selection methodology.

Original languageEnglish (US)
Pages (from-to)180-191
Number of pages12
JournalStatistical Methodology
Volume33
DOIs
StatePublished - Dec 1 2016

Fingerprint

Gaussian Kernel
kernel
Power Function
Genetic Association
Neuroimaging
Statistical Power
Alzheimer's Disease
Type I error
Maximise
Simulation Study
Upper bound
Estimator
Methodology
Alternatives

All Science Journal Classification (ASJC) codes

  • Statistics and Probability

Cite this

@article{c731d35e5e6d4f378d3c1497031078b8,
title = "A novel power-based approach to Gaussian kernel selection in the kernel-based association test",
abstract = "Kernel-based association test (KAT) is a widely used tool in genetics association analysis. The performance of such a test depends on the choice of kernel. In this paper, we study the statistical power of a KAT using a Gaussian kernel. We explicitly develop a notion of analytical power function in this family of tests. We propose a novel approach to select the kernel so as to maximize the analytical power function of the test at a given test level (an upper bound on the probability of making a type I error). We assess some theoretical properties of our optimal estimator, and compare its performance with some similar existing alternatives using simulation studies. Neuroimaging data from an Alzheimer's disease study is also used to illustrate the proposed kernel selection methodology.",
author = "Xiang Zhan and Debashis Ghosh",
year = "2016",
month = "12",
day = "1",
doi = "10.1016/j.stamet.2016.09.003",
language = "English (US)",
volume = "33",
pages = "180--191",
journal = "Statistical Methodology",
issn = "1572-3127",
publisher = "Elsevier",

}

A novel power-based approach to Gaussian kernel selection in the kernel-based association test. / Zhan, Xiang; Ghosh, Debashis.

In: Statistical Methodology, Vol. 33, 01.12.2016, p. 180-191.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A novel power-based approach to Gaussian kernel selection in the kernel-based association test

AU - Zhan, Xiang

AU - Ghosh, Debashis

PY - 2016/12/1

Y1 - 2016/12/1

N2 - Kernel-based association test (KAT) is a widely used tool in genetics association analysis. The performance of such a test depends on the choice of kernel. In this paper, we study the statistical power of a KAT using a Gaussian kernel. We explicitly develop a notion of analytical power function in this family of tests. We propose a novel approach to select the kernel so as to maximize the analytical power function of the test at a given test level (an upper bound on the probability of making a type I error). We assess some theoretical properties of our optimal estimator, and compare its performance with some similar existing alternatives using simulation studies. Neuroimaging data from an Alzheimer's disease study is also used to illustrate the proposed kernel selection methodology.

AB - Kernel-based association test (KAT) is a widely used tool in genetics association analysis. The performance of such a test depends on the choice of kernel. In this paper, we study the statistical power of a KAT using a Gaussian kernel. We explicitly develop a notion of analytical power function in this family of tests. We propose a novel approach to select the kernel so as to maximize the analytical power function of the test at a given test level (an upper bound on the probability of making a type I error). We assess some theoretical properties of our optimal estimator, and compare its performance with some similar existing alternatives using simulation studies. Neuroimaging data from an Alzheimer's disease study is also used to illustrate the proposed kernel selection methodology.

UR - http://www.scopus.com/inward/record.url?scp=84992129619&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84992129619&partnerID=8YFLogxK

U2 - 10.1016/j.stamet.2016.09.003

DO - 10.1016/j.stamet.2016.09.003

M3 - Article

VL - 33

SP - 180

EP - 191

JO - Statistical Methodology

JF - Statistical Methodology

SN - 1572-3127

ER -