Premium
Microsatellites versus single‐nucleotide polymorphisms in confidence interval estimation of disease loci
Author(s) -
Papachristou Charalampos,
Lin Shili
Publication year - 2006
Publication title -
genetic epidemiology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.301
H-Index - 98
eISSN - 1098-2272
pISSN - 0741-0395
DOI - 10.1002/gepi.20122
Subject(s) - linkage disequilibrium , single nucleotide polymorphism , loss of heterozygosity , tag snp , locus (genetics) , microsatellite , snp , genetics , inference , biology , computational biology , multiple comparisons problem , allele , computer science , statistics , genotype , gene , artificial intelligence , mathematics
Abstract With cost‐effective high‐throughput Single Nucleotide Polymorphism (SNP) arrays now becoming widely available, it is highly anticipated that SNPs will soon become the choice of markers in whole genome screens. This optimism raises a great deal of interest in assessing whether dense SNP maps offer at least as much information as their microsatellite (MS) counterparts. Factors considered to date include information content, strength of linkage signals, and effect of linkage disequilibrium. In the current report, we focus on investigating the relative merits of SNPs vs. MS markers for disease gene localization. For our comparisons, we consider three novel confidence interval estimation procedures based on confidence set inference (CSI) using affected sib‐pair data. Two of these procedures are multipoint in nature, enabling them to capitalize on dense SNPs with limited heterozygosity. The other procedure makes use of markers one at a time (two‐point), but is much more computationally efficient. In addition to marker type, we also assess the effects of a number of other factors, including map density and marker heterozygosity, on disease gene localization through an extensive simulation study. Our results clearly show that confidence intervals derived based on the CSI multipoint procedures can place the trait locus in much shorter chromosomal segments using densely saturated SNP maps as opposed to using sparse MS maps. Finally, it is interesting (although not surprising) to note that, should one wish to perform a quick preliminary genome screening, then the two‐point CSI procedure would be a preferred, computationally cost‐effective choice. Genet. Epidemiol. 30:3–17, 2006. © 2005 Wiley‐Liss, Inc.