Mixed-effects logistic approach for association following linkage scan for complex disorders

H. Xu, Sanjay Shete

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

An association study to identify possible causal single nucleotide polymorphisms following linkage scanning is a popular approach for the genetic dissection of complex disorders. However, in association studies cases and controls are assumed to be independent, i.e., genetically unrelated. Choosing a single affected individual per family is statistically inefficient and leads to a loss of power. On the other hand, because of the relatedness of family members, using affected family members and unrelated normal controls directly leads to false-positive results in association studies. In this paper we propose a new approach using mixed-model logistic regression, in which associations are performed using family members and unrelated controls. Thus, the important genetic information can be obtained from family members while retaining high statistical power. To examine the properties of this new approach we developed an efficient algorithm, to simulate environmental risk factors and the genotypes at both the disease locus and a marker locus with and without linkage disequilibrium (LD) in families. Extensive simulation studies showed that our approach can effectively control the type-I error probability. Our approach is better than family-based designs such as TDT, because it allows the use of unrelated cases and controls and uses all of the affected members for whom DNA samples are possibly already available. Our approach also allows the inclusion of covariates such as age and smoking status. Power analysis showed that our method has higher statistical power than recent likelihood ratio-based methods when environmental factors contribute to disease susceptibility, which is true for most complex human disorders. Our method can be further extended to accommodate more complex pedigree structures.

Original languageEnglish (US)
Pages (from-to)230-237
Number of pages8
JournalAnnals of Human Genetics
Volume71
Issue number2
DOIs
StatePublished - Mar 1 2007
Externally publishedYes

Fingerprint

Disease Susceptibility
Linkage Disequilibrium
Pedigree
Single Nucleotide Polymorphism
Case-Control Studies
Dissection
Logistic Models
Smoking
Genotype
DNA
Power (Psychology)

Keywords

  • Association study
  • Mixed model
  • Pedigree

ASJC Scopus subject areas

  • Genetics
  • Genetics(clinical)

Cite this

Mixed-effects logistic approach for association following linkage scan for complex disorders. / Xu, H.; Shete, Sanjay.

In: Annals of Human Genetics, Vol. 71, No. 2, 01.03.2007, p. 230-237.

Research output: Contribution to journalArticle

@article{9f445f5d46a44b4f98675b5910883c60,
title = "Mixed-effects logistic approach for association following linkage scan for complex disorders",
abstract = "An association study to identify possible causal single nucleotide polymorphisms following linkage scanning is a popular approach for the genetic dissection of complex disorders. However, in association studies cases and controls are assumed to be independent, i.e., genetically unrelated. Choosing a single affected individual per family is statistically inefficient and leads to a loss of power. On the other hand, because of the relatedness of family members, using affected family members and unrelated normal controls directly leads to false-positive results in association studies. In this paper we propose a new approach using mixed-model logistic regression, in which associations are performed using family members and unrelated controls. Thus, the important genetic information can be obtained from family members while retaining high statistical power. To examine the properties of this new approach we developed an efficient algorithm, to simulate environmental risk factors and the genotypes at both the disease locus and a marker locus with and without linkage disequilibrium (LD) in families. Extensive simulation studies showed that our approach can effectively control the type-I error probability. Our approach is better than family-based designs such as TDT, because it allows the use of unrelated cases and controls and uses all of the affected members for whom DNA samples are possibly already available. Our approach also allows the inclusion of covariates such as age and smoking status. Power analysis showed that our method has higher statistical power than recent likelihood ratio-based methods when environmental factors contribute to disease susceptibility, which is true for most complex human disorders. Our method can be further extended to accommodate more complex pedigree structures.",
keywords = "Association study, Mixed model, Pedigree",
author = "H. Xu and Sanjay Shete",
year = "2007",
month = "3",
day = "1",
doi = "10.1111/j.1469-1809.2006.00321.x",
language = "English (US)",
volume = "71",
pages = "230--237",
journal = "Annals of Human Genetics",
issn = "0003-4800",
publisher = "Wiley-Blackwell",
number = "2",

}

TY - JOUR

T1 - Mixed-effects logistic approach for association following linkage scan for complex disorders

AU - Xu, H.

AU - Shete, Sanjay

PY - 2007/3/1

Y1 - 2007/3/1

N2 - An association study to identify possible causal single nucleotide polymorphisms following linkage scanning is a popular approach for the genetic dissection of complex disorders. However, in association studies cases and controls are assumed to be independent, i.e., genetically unrelated. Choosing a single affected individual per family is statistically inefficient and leads to a loss of power. On the other hand, because of the relatedness of family members, using affected family members and unrelated normal controls directly leads to false-positive results in association studies. In this paper we propose a new approach using mixed-model logistic regression, in which associations are performed using family members and unrelated controls. Thus, the important genetic information can be obtained from family members while retaining high statistical power. To examine the properties of this new approach we developed an efficient algorithm, to simulate environmental risk factors and the genotypes at both the disease locus and a marker locus with and without linkage disequilibrium (LD) in families. Extensive simulation studies showed that our approach can effectively control the type-I error probability. Our approach is better than family-based designs such as TDT, because it allows the use of unrelated cases and controls and uses all of the affected members for whom DNA samples are possibly already available. Our approach also allows the inclusion of covariates such as age and smoking status. Power analysis showed that our method has higher statistical power than recent likelihood ratio-based methods when environmental factors contribute to disease susceptibility, which is true for most complex human disorders. Our method can be further extended to accommodate more complex pedigree structures.

AB - An association study to identify possible causal single nucleotide polymorphisms following linkage scanning is a popular approach for the genetic dissection of complex disorders. However, in association studies cases and controls are assumed to be independent, i.e., genetically unrelated. Choosing a single affected individual per family is statistically inefficient and leads to a loss of power. On the other hand, because of the relatedness of family members, using affected family members and unrelated normal controls directly leads to false-positive results in association studies. In this paper we propose a new approach using mixed-model logistic regression, in which associations are performed using family members and unrelated controls. Thus, the important genetic information can be obtained from family members while retaining high statistical power. To examine the properties of this new approach we developed an efficient algorithm, to simulate environmental risk factors and the genotypes at both the disease locus and a marker locus with and without linkage disequilibrium (LD) in families. Extensive simulation studies showed that our approach can effectively control the type-I error probability. Our approach is better than family-based designs such as TDT, because it allows the use of unrelated cases and controls and uses all of the affected members for whom DNA samples are possibly already available. Our approach also allows the inclusion of covariates such as age and smoking status. Power analysis showed that our method has higher statistical power than recent likelihood ratio-based methods when environmental factors contribute to disease susceptibility, which is true for most complex human disorders. Our method can be further extended to accommodate more complex pedigree structures.

KW - Association study

KW - Mixed model

KW - Pedigree

UR - http://www.scopus.com/inward/record.url?scp=34047261588&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34047261588&partnerID=8YFLogxK

U2 - 10.1111/j.1469-1809.2006.00321.x

DO - 10.1111/j.1469-1809.2006.00321.x

M3 - Article

C2 - 17032287

AN - SCOPUS:34047261588

VL - 71

SP - 230

EP - 237

JO - Annals of Human Genetics

JF - Annals of Human Genetics

SN - 0003-4800

IS - 2

ER -