untitled


Genetic Epidemiology 34 : 146–150 (2010)

Extent and Distribution of Linkage Disequilibrium in the
Old Order Amish

Cristopher V. Van Hout,1� Albert M. Levin,1� Evadnie Rampersaud,2 Haiqing Shen,2

Jef frey R. O’Connell,2 Braxton D. Mitchell,2 Alan R. Shuldiner,2,3 and Julie A. Douglas1y

1Department of Human Genetics, University of Michigan School of Medicine, Ann Arbor, Michigan
2Department of Medicine, University of Maryland School of Medicine, Baltimore, Maryland

3Geriatric Research and Education Clinical Center, Veterans Administration Medical Center, Baltimore, Maryland

Knowledge of the extent and distribution of linkage disequilibrium (LD) is critical to the design and interpretation of gene
mapping studies. Because the demographic history of each population varies and is often not accurately known, it is
necessary to empirically evaluate LD on a population-specific basis. Here we present the first genome-wide survey of LD in
the Old Order Amish (OOA) of Lancaster County Pennsylvania, a closed population derived from a modest number of
founders. Specifically, we present a comparison of LD between OOA individuals and US Utah participants in the
International HapMap project (abbreviated CEU) using a high-density single nucleotide polymorphism (SNP) map. Overall,
the allele (and haplotype) frequency distributions and LD profiles were remarkably similar between these two populations.
For example, the median absolute allele frequency dif ference for autosomal SNPs was 0.05, with an inter-quartile range of
0.02–0.09, and for autosomal SNPs 10–20 kb apart with common alleles (minor allele frequencyZ0.05), the LD measure r2

was at least 0.8 for 15 and 14% of SNP pairs in the OOA and CEU, respectively. Moreover, tag SNPs selected from the
HapMap CEU sample captured a substantial portion of the common variation in the OOA (�88%) at r2Z0.8. These results
suggest that the OOA and CEU may share similar LD profiles for other common but untyped SNPs. Thus, in the context of
the common variant-common disease hypothesis, genetic variants discovered in gene mapping studies in the OOA may
generalize to other populations. Genet. Epidemiol. 34 : 146–150, 2010. r 2009 Wiley-Liss, Inc.

Key words: single nucleotide polymorphism; population genetics; human genetics; founder population; linkage
disequilibrium; haplotypes

Contract grant sponsor: NIH; Contract grant numbers: U01 HL72515; R01 CA122844.
�The first two authors contributed equally to this work.
yCorrespondence to: Julie A. Douglas, Department of Human Genetics, University of Michigan, 1241 E. Catherine St., 5912 Buhl Building,
SPC 5618 Ann Arbor, MI 48109-5618. E-mail: jddoug@umich.edu
Received 27 January 2009; Revised 14 April 2009; Accepted 10 June 2009
Published online 20 August 2009 in Wiley InterScience (www.interscience.wiley.com).
DOI: 10.1002/gepi.20444

INTRODUCTION

Many genetic studies of complex traits and diseases are
being conducted in population isolates, including the Old
Order Amish (OOA) of Lancaster County Pennsylvania
[Ginns et al., 1998; Hsueh et al., 2000; Mitchell et al., 2001,
2008; Streeten et al., 2006; Post et al., 2007; Douglas et al.,
2008; Wang et al., 2009]. Whether results from these studies
will generalize to other populations is dependent (in part)
on the similarity of allele frequencies and patterns of
linkage disequilibrium (LD) between populations. To
inform future genetic studies of the OOA and facilitate
comparisons of findings with other populations, we
conducted the first genome-wide survey of LD in the
OOA and compared our findings to the International
HapMap project [Frazer et al., 2007].

Most of the present-day OOA of Lancaster County are
the descendants of approximately 200 individuals [Cross,
1976] from central western Europe who immigrated to the
United States in the early eighteenth century [McKusick
et al., 1964]. Although recent data indicate that the

dif ferences in LD between isolated and cosmopolitan
populations for common alleles are modest [Bonnen et al.,
2006; Service et al., 2006], the uncertain but unique
demographic history of the OOA necessitates empirical
evaluation of LD.

SUBJECTS AND METHODS

OOA study subjects were recruited and genotyped
(n 5 861) in the course of the Heredity and Phenotype
Intervention (HAPI) Heart study [Mitchell et al., 2008],
which was designed to identify gene-environment inter-
actions influencing cardiovascular traits. Because many
closely related individuals were deliberately ascertained,
we used a simulated annealing algorithm [Douglas and
Sandefur, 2008] to select a set of minimally related
individuals (30 men and 30 women). The median (range)
pair-wise kinship coef ficient was 0.03 (0.01–0.04) for the
set of 60 vs. 0.03 (0.01–0.3) for the entire sample of 861. For
comparison with the OOA, we also utilized 30 men and 30
women (or 60 unrelated parents) from a US Utah

r 2009 Wiley-Liss, Inc.



population with northern and western European ancestry
(abbreviated CEU) in the International HapMap project
[Frazer et al., 2007].

GENOTYPING AND QC METHODS

DNA was extracted from whole blood by standard
methods as described previously [Mitchell et al., 2008].
The Af fymetrix GeneChips Human Mapping 500K Array
Set was used for the comparison of LD patterns in both the
OOA and CEU samples. Genotype calls were made using
a Bayesian Robust Linear Model with Mahalanobis
(BRLMM) distance classifier [Af fymetrix, 2006]. Genotype
data for the CEU sample and corresponding annotation for
the platform, including chromosome and genomic posi-
tions for all single nucleotide polymorphisms (SNPs) on
the array, were obtained from the Af fymetrix website
(www.af fymetrix.com).

Individuals with 45% missing genotypes, and/or for
men 41% heterozygous genotypes on the X chromosome,
were excluded. A subset of autosomal SNPs (2,068), which
were selected to have high information content (minor
allele frequency (MAF) Z0.3), low pair-wise LD (maxi-
mum r2 of 0.44), and coverage across all autosomes
(average intermarker spacing of 1.3 cm) in the OOA, were
used to infer relationships using the maximum likelihood
method implemented in Relpair [Epstein et al., 2000].
We excluded individuals who had an inferred relation-
ship that dif fered from the pedigree relationship with
a likelihood ratio greater than 106. Based on these
combined criteria, a total of 24 individuals (out of 861)
were excluded from further analysis.

SNPs were required to satisfy the following quality
control criteria in both samples: (1) r5% uncalled
genotypes; (2) r5 and r1 Mendelian inconsistencies in
OOA and CEU samples, respectively, using pedigree
diagnostics as implemented in PedCheck [O’Connell
and Weeks, 1998]; and (3) Hardy Weinberg Equilibrium
P-valueZ10�6 by Fisher’s exact test [Wigginton
et al., 2005] as implemented in Haploview [Barrett et al.,
2005]. To assess genotyping accuracy, we used duplicate
genotype data for 61 of the 861 OOA subjects for whom
data from the Af fymetrix Genome-Wide Human SNP
Array 6.0 (overlap of 482,235 SNPs with Af fymetrix
GeneChips Human Mapping 500K Array Set) were also
available. Only SNPs with o2 duplicate inconsistencies
were retained for analysis. Of the 500,447 genotypes that
mapped to a single location in the human genome, 82,404
failed at least one QC measure in at least one sample.
Those SNPs were removed, leaving a total of 409,071
autosomal (Table I) and 8,972 X chromosome (Table AI in
the Appendix) SNPs. For the SNPs that passed our quality
control criteria, the genotype consistency rate among 61
duplicate pairs was 99.4%.

STATISTICAL ANALYSES

Fisher’s exact test was used to compare allele frequency
distributions between the OOA and CEU. For common
SNPs (MAFZ0.05) on the same chromosome and within
10 Mb of each other, we used the expectation-maximiza-
tion (EM) algorithm to obtain maximum likelihood
estimates of two-SNP haplotype frequencies and mea-
sured pair-wise LD by the r2 and D0 statistics [Lewontin,
1964]. Based on common SNPs, we also identified

haplotype blocks in the CEU using an extension of the
four-gamete rule [Wang et al., 2002] and estimated
haplotype frequencies in both the CEU and OOA using
the EM algorithm with a partition-ligation method
[Qin et al., 2002] for blocks with 410 SNPs as implemen-
ted in Haploview [Barrett et al., 2005]. For each sample,
we then calculated and compared the ef fective number
of haplotypes in each block, i.e., (Spi

2)�1, where pi is
the frequency of the ith haplotype in the block. As a
measure of redundancy, we identified the number of SNPs
(or proxies) that were in strong LD with each SNP at
various thresholds of r2 in each sample. To evaluate the
extent to which SNPs selected to tag variation in the CEU
capture common variation in the OOA, we selected
common tag SNPs in the CEU using the greedy algorithm
[Carlson et al., 2004] implemented in Haploview [Barrett
et al., 2005] such that every unselected SNP had an r2Z0.8
with one or more selected SNPs. We then calculated r2

between the tag SNPs and the remaining ‘‘non-tagged’’ but
typed SNPs in the OOA. Unless specified otherwise, all
analyses were carried out using a combination of in-house
R, Perl, and C programs.

RESULTS

For the 418,043 SNPs that passed QC, mean hetero-
zygosity was 0.26 and 0.27 for the autosomes in the OOA
and CEU, respectively, and 0.23 and 0.24 for the X
chromosome. The slightly lower heterozygosity in the
OOA reflects the larger number of monomorphic
SNPs in the OOA relative to the CEU, e.g., 68,869 vs.
57,669 for the autosomes (Table I). Among all SNPs that

TABLE I. Summary of autosomal SNPs

OOA CEU Overlap

Total genotyped 489,922 489,922 489,922
41 duplicate inconsistencya 51,459 NA NA
45% missing datab 50,085 16,896 8,973
Mendelian inconsistenciesb,c 3,188 1,168 202
Po10�6 for HWE testd 379 217 116

Passed QC filtere 415,440 472,851 409,071
Passed QC in both OOA and CEU

Monomorphicd 68,869 57,669 52,467
Polymorphicd

MAFZ0.05 297,605 310,704 287,476
MAFZ0.10 256,614 267,149 240,375
MAFZ0.20 182,941 189,133 161,062

OOA, Old Order Amish; CEU, US Utah residents from HapMap;
MAF, minor allele frequency; SNPs that failed a QC measure in
either sample were excluded from further analysis, and SNPs with
MAFZ0.05 passing QC in both samples (n 5 287,476) were used
for LD analysis.
aBased on the 61 OOA individuals who were also genotyped on
the Af fymetrix 6.0 array; SNPs with more than one duplicated
genotype discrepancy were excluded.
bBased on 837 OOA and 90 CEU individuals (30 trios).
cSNPs with 45 and 41 Mendelian inconsistencies in OOA and
CEU, respectively.
dBased on 60 unrelated individuals (30 men and 30 women) from
each sample.
eSNPs may fail QC in more than one way, so rows do not sum to
the subtotal passing QC.

147Linkage Disequilibrium in Old Order Amish

Genet. Epidemiol.



were polymorphic in at least one sample, the median
absolute allele frequency dif ference was 0.05 for the
autosomes and 0.07 for the X chromosome. At
P-valueo10�6, OOA and CEU allele frequencies were
significantly dif ferent for 799 autosomal and 137 X
chromosome SNPs.

The percentage of SNP pairs within 10 Mb of each other
and between which strong LD was observed was
remarkably similar between the OOA and CEU for
the autosomes (Table II) and the X chromosome
(Table AII in the Appendix). For example, for autosomal
SNPs at an inter-marker distance of o10 kb, no evidence
of recombination (D0 5 1) was observed for 79 and
75% of SNP pairs, perfect LD (r2 5 1) was observed for
20 and 19% of SNP pairs, and useful LD (r2Z0.8) was
observed for 30 and 29% of SNP pairs in the OOA and
CEU, respectively. Based on the CEU sample, we identi-
fied 58,097 autosomal haplotype blocks, with a median of
three SNPs per block and an inter-quartile range of [3, 4].
Among all autosomal blocks, the median ef fective number
of haplotypes (ne) was 2.43 and 2.47 in the OOA and CEU,
respectively, and the median of the dif ferences in ne (CEU
minus OOA) per block was 0.04, with an inter-quartile
range of �0.2 to 0.3, suggesting modestly greater haplo-
type diversity in the CEU. Results based on haplotype
blocks defined in the OOA did not qualitatively dif fer
from those based on blocks defined in the CEU (data not
shown).

Of common autosomal SNPs, 72 and 64% had at least
one proxy at r2Z0.8 and 55 and 44% had at least one
perfect proxy (r2 5 1) in the OOA and CEU, respectively,
indicating that fewer independent SNPs are required to
represent variation in the OOA relative to the CEU. At
r2Z0.8, 170,979 of 310,704 common SNPs in the CEU were
selected as tag SNPs and captured �88% of the ‘‘non-
tagged’’ SNPs in OOA, suggesting that SNPs selected to
tag common variation in the CEU capture much of the
same variation in the OOA. SNPs not captured by the CEU
tag SNPs tended to be of lower MAF (data not shown).
Results for the X chromosome were qualitatively similar.

DISCUSSION

In general, we found a high degree of similarity in allele
frequencies and LD patterns in the OOA and CEU
samples. Allele frequencies were not significantly dif ferent
between the OOA and CEU for 499% of SNPs. Based on
common SNPs, which comprised 74 and 66% of autosomal
SNPs in the OOA and CEU, respectively, the distribution
and extent of LD were remarkably similar between these
two samples. These data are consistent with previous
theoretical predictions [Kruglyak, 1999; Pritchard and
Przeworski, 2001] and recent empirical data [Bonnen
et al., 2006; Service et al., 2006; Navarro et al., 2009;
Thompson et al., 2009], all of which point to modest
dif ferences in LD between isolated and cosmopolitan
populations for common alleles. The situation for rare
alleles, however, is likely to be dif ferent as has been
demonstrated in applications of LD mapping for mono-
genic diseases and traits.

Demographic and historical information indicate that the
OOA were founded relatively recently (�10–15 generations
ago) by a modest number of individuals (several hundred)
and then expanded rapidly to a current census population
size exceeding 30,000 [Lancaster County Amish, 2002].
Though the precise demographic details are unknown, it is
apparent that the number of founders and rate of growth
were sufficient and that the subsequent isolation of the
OOA was too short for genetic drift and/or recombination
to have meaningfully altered the common allele or
haplotype frequency spectrum. Our recent study of
variation on the Y chromosome supports these observa-
tions in that much of the diversity observed in non-isolated
populations of similar ancestry is present in the OOA
[Pollin et al., 2008]. It appears that inbreeding due to the
finite population size of the OOA was also insufficient to
meaningfully alter the allele frequency distribution or
extent of LD. Based on the 60 OOA individuals included in
our analyses, the average inbreeding coef ficient F [Wright,
1922] was 0.026 (range of 0.0003–0.046), which is too weak
to generate substantial differences in LD relative to a non-
isolated population [Hill and Robertson, 1968].

Owing to similar allele frequencies and LD patterns
in the OOA and CEU, CEU-derived tag SNPs performed
well in capturing common variation in the OOA,
consistent with previous studies in other samples of
European ancestry, including those from isolated popula-
tions [Willer et al., 2006; Service et al., 2007]. These results
suggest that the OOA and CEU samples may also share
similar LD profiles for other common but untyped SNPs.
Thus, findings from gene mapping studies in the OOA
may generalize to other populations in the context of the
common variant-common disease hypothesis.

ACKNOWLEDGMENTS

We gratefully acknowledge the Amish Research Clinic
Staf f, our Amish liaisons, and the Amish community,
whose extraordinary support and cooperation made
this study possible. We also thank Drs. Alejandro
Schaf fer and Richa Agarwala at the NIH/NCBI for
providing the pedigree information and the Center for
Inherited Disease Research (CIDR), NIH for providing
duplicate genotypes from the Af fymetrix Genome-Wide
Human SNP Array 6.0.

TABLE II. Percentage of autosomal SNP pairsa showing
no evidence of recombination (D0 5 1), perfect LD
(r2 5 1), or where useful LD is observed (r2Z0.8)

D0 5 1 r2 5 1 r2Z0.8

Inter-SNP distance (kb) OOA CEU OOA CEU OOA CEU

r10 79 75 20 19 30 29
10–20 60 53 9 7 15 14
20–50 43 34 4 3 9 7
50–100 28 20 1 1 3 2
100–200 20 11 0 0 1 1
200–500 14 7 0 0 0 0
500–1,000 12 6 0 0 0 0
1,000–2,000 11 5 0 0 0 0
2,000–5,000 10 5 0 0 0 0
5,000–10,000 8 5 0 0 0 0

OOA, Old Order Amish (n 5 60); CEU, US Utah residents from
HapMap (n 5 60).
aRestricted to SNPs with minor allele frequency Z0.05 in both
samples (n 5 287,476).

148 Van Hout et al.

Genet. Epidemiol.



REFERENCES
Af fymetrix. 2006. BRLMM: an Improved Genotype Calling Method

for the GeneChip Human Mapping 500K Array Set. [http://

www. af fymetrix. com / support / technical/whitepapers/brlmm_

whitepaper.pdf].

Barrett JC, Fry B, Maller J, Daly MJ. 2005. Haploview: analysis and

visualization of LD and haplotype maps. Bioinformatics 21:263–265.

Bonnen PE, Pe’er I, Plenge RM, Salit J, Lowe JK, Shapero MH,

Lifton RP, Breslow JL, Daly MJ, Reich DE, Jones KW, Stof fel M,

Altshuler D, Friedman JM. 2006. Evaluating potential for whole-

genome studies in Kosrae, an isolated population in Micronesia.

Nat Genet 38:214–217.

Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA.

2004. Selecting a maximally informative set of single-nucleotide

polymorphisms for association analyses using linkage disequili-

brium. Am J Hum Genet 74:106–120.

Cross HE. 1976. Population studies and the Old Order Amish. Nature

262:17–20.

Douglas JA, Sandefur CI. 2008. PedMine—a simulated annealing

algorithm to identify maximally unrelated individuals in popula-

tion isolates. Bioinformatics 24:1106–1108.
Douglas JA, Roy-Gagnon MH, Zhou C, Mitchell BD, Shuldiner AR,

Chan HP, Helvie MA. 2008. Mammographic breast density—

evidence for genetic correlations with established breast cancer

risk factors. Cancer Epidemiol Biomarkers Prev 17:3509–3516.

Epstein MP, Duren WL, Boehnke M. 2000. Improved inference of

relationship for pairs of individuals. Am J Hum Genet 67:1219–1231.

Frazer KA, Ballinger DG, Cox DR, Hinds DA, Stuve LL, Gibbs RA,

Belmont JW, Boudreau A, Hardenbol P, Leal SM, Pasternak S,

Wheeler DA, Willis TD, Yu F, Yang H, Zeng C, Gao Y, Hu H, Hu W,

Li C, Lin W, Liu S, Pan H, Tang X, Wang J, Wang W, Yu J, Zhang B,

Zhang Q, Zhao H, Zhao H, Zhou J, Gabriel SB, Barry R,

Blumenstiel B, Camargo A, Defelice M, Faggart M, Goyette M,

Gupta S, Moore J, Nguyen H, Onofrio RC, Parkin M, Roy J, Stahl E,

Winchester E, Ziaugra L, Altshuler D, Shen Y, Yao Z, Huang W,

Chu X, He Y, Jin L, Liu Y, Shen Y, Sun W, Wang H, Wang Y,

Wang Y, Xiong X, Xu L, Waye MM, Tsui SK, Xue H, Wong JT,

Galver LM, Fan JB, Gunderson K, Murray SS, Oliphant AR,

Chee MS, Montpetit A, Chagnon F, Ferretti V, Leboeuf M,

Olivier JF, Phillips MS, Roumy S, Sallée C, Verner A, Hudson TJ,

Kwok PY, Cai D, Koboldt DC, Miller RD, Pawlikowska L,

Taillon-Miller P, Xiao M, Tsui LC, Mak W, Song YQ, Tam PK,

Nakamura Y, Kawaguchi T, Kitamoto T, Morizono T,

Nagashima A, Ohnishi Y, Sekine A, Tanaka T, Tsunoda T,

Deloukas P, Bird CP, Delgado M, Dermitzakis ET, Gwilliam R,

Hunt S, Morrison J, Powell D, Stranger BE, Whittaker P,

Bentley DR, Daly MJ, de Bakker PI, Barrett J, Chretien YR,

Maller J, McCarroll S, Patterson N, Pe’er I, Price A, Purcell S,

Richter DJ, Sabeti P, Saxena R, Schaf fner SF, Sham PC, Varilly P,

Altshuler D, Stein LD, Krishnan L, Smith AV, Tello-Ruiz MK,

Thorisson GA, Chakravarti A, Chen PE, Cutler DJ, Kashuk CS,

Lin S, Abecasis GR, Guan W, Li Y, Munro HM, Qin ZS, Thomas DJ,

McVean G, Auton A, Bottolo L, Cardin N, Eyheramendy S,

Freeman C, Marchini J, Myers S, Spencer C, Stephens M,

Donnelly P, Cardon LR, Clarke G, Evans DM, Morris AP,

Weir BS, Tsunoda T, Mullikin JC, Sherry ST, Feolo M, Skol A,

Zhang H, Zeng C, Zhao H, Matsuda I, Fukushima Y, Macer DR,

Suda E, Rotimi CN, Adebamowo CA, Ajayi I, Aniagwu T,

Marshall PA, Nkwodimmah C, Royal CD, Leppert MF, Dixon M,

Peif fer A, Qiu R, Kent A, Kato K, Niikawa N, Adewole IF,

Knoppers BM, Foster MW, Clayton EW, Watkin J, Gibbs RA,

Belmont JW, Muzny D, Nazareth L, Sodergren E, Weinstock GM,

Wheeler DA, Yakub I, Gabriel SB, Onofrio RC, Richter DJ,

Ziaugra L, Birren BW, Daly MJ, Altshuler D, Wilson RK,

Fulton LL, Rogers J, Burton J, Carter NP, Clee CM, Grif fiths M,

Jones MC, McLay K, Plumb RW, Ross MT, Sims SK, Willey DL,

Chen Z, Han H, Kang L, Godbout M, Wallenburg JC,

L’Archevêque P, Bellemare G, Saeki K, Wang H, An D, Fu H,

Li Q, Wang Z, Wang R, Holden AL, Brooks LD, McEwen JE,

Guyer MS, Wang VO, Peterson JL, Shi M, Spiegel J, Sung LM,

Zacharia LF, Collins FS, Kennedy K, Jamieson R, Stewart J. 2007.

A second generation human haplotype map of over 3.1 million

SNPs. Nature 449:851–861.

Ginns EI, St Jean P, Philibert RA, Galdzicka M, Damschroder-Williams P,

Thiel B, Long RT, Ingraham LJ, Dalwaldi H, Murray MA, Ehlert M,

Paul S, Remortel BG, Patel AP, Anderson MC, Shaio C, Lau E,

Dymarskaia I, Martin BM, Stubblefield B, Falls KM, Carulli JP,

Keith TP, Fann CS, Lacy LG, Allen CR, Hostetter AM, Elston RC,

Schork NJ, Egeland JA, Paul SM. 1998. A genome-wide search for

chromosomal loci linked to mental health wellness in relatives at

high risk for bipolar affective disorder among the Old Order Amish.

Proc Natl Acad Sci USA 95:15531–15536.

Hill WG, Robertson A. 1968. Linkage disequilibrium in finite

populations. Theor Appl Genet 38:226–231.

Hsueh WC, Mitchell BD, Aburomia R, Pollin T, Sakul H, Gelder Ehm M,

Michelsen BK, Wagner MJ, St Jean PL, Knowler WC, Burns DK,

Bell CJ, Shuldiner AR. 2000. Diabetes in the Old Order Amish:

characterization and heritability analysis of the Amish Family

Diabetes Study. Diabetes Care 23:595–601.

Kruglyak L. 1999. Prospects for whole-genome linkage disequilibrium

mapping of common disease genes. Nat Genet 22:139–144.

Lancaster County Amish. 2002. Church Directory of the Lancaster

County Amish. Gordonville, PA: The Diary.

Lewontin RC. 1964. The interaction of selection and linkage. II.

Optimum models. Genetics 50:757–782.

McKusick VA, Hostetler JA, Egeland JA. 1964. Genetic studies of the

Amish, background and potentialities. Bull Johns Hopkins Hosp

115:203–222.

Mitchell BD, Hsueh WC, King TM, Pollin TI, Sorkin J, Agarwala R,

Schaf fer AA, Shuldiner AR. 2001. Heritability of life span in the

Old Order Amish. Am J Med Genet 102:346–352.
Mitchell BD, McArdle PF, Shen H, Rampersaud E, Pollin TI, Bielak LF,

Jaquish C, Douglas JA, Roy-Gagnon MH, Sack P, Naglieri R,

Hines S, Horenstein RB, Chang YP, Post W, Ryan KA, Brereton NH,

Pakyz RE, Sorkin J, Damcott CM, O’Connell JR, Mangano C,

Corretti M, Vogel R, Herzog W, Weir MR, Peyser PA, Shuldiner AR.

2008. The genetic response to short-term interventions affecting

cardiovascular function: rationale and design of the Heredity and

Phenotype Intervention (HAPI) Heart Study. Am Heart J 155:823–828.
Navarro P, Vitart V, Hayward C, Tenesa A, Zgaga L, Juricic D, Polasek O,

Hastie ND, Rudan I, Campbell H, Wright AF, Haley CS, Knott SA.

2009. Genetic comparison of a Croatian isolate and CEPH European

Founders. Genet Epidemiol, DOI: 10.1002/gepi.20443.

O’Connell JR, Weeks DE. 1998. PedCheck: a program for identification

of genotype incompatibilities in linkage analysis. Am J Hum Genet

63:259–266.

Pollin TI, McBride DJ, Agarwala R, Schaf fer AA, Shuldiner AR,

Mitchell BD, O’Connell JR. 2008. Investigations of the Y chromo-

some, male founder structure and YSTR mutation rates in the Old

Order Amish. Hum Hered 65:91–104.

Post W, Bielak LF, Ryan KA, Cheng YC, Shen H, Rumberger JA,

Sheedy 2nd PF, Shuldiner AR, Peyser PA, Mitchell BD. 2007.

Determinants of coronary artery and aortic calcification in the Old

Order Amish. Circulation 115:717–724.

Pritchard JK, Przeworski M. 2001. Linkage disequilibrium in humans:

models and data. Am J Hum Genet 69:1–14.

Qin ZS, Niu T, Liu JS. 2002. Partition-ligation-expectation-maximiza-

tion algorithm for haplotype inference with single-nucleotide

polymorphisms. Am J Hum Genet 71:1242–1247.

Service S, DeYoung J, Karayiorgou M, Roos JL, Pretorious H,

Bedoya G, Ospina J, Ruiz-Linares A, Macedo A, Palha JA,

Heutink P, Aulchenko Y, Oostra B, van Duijn C, Jarvelin MR,

Varilo T, Peddle L, Rahman P, Piras G, Monne M, Murray S,

Galver L, Peltonen L, Sabatti C, Collins A, Freimer N. 2006.

Magnitude and distribution of linkage disequilibrium in

149Linkage Disequilibrium in Old Order Amish

Genet. Epidemiol.



population isolates and implications for genome-wide association

studies. Nat Genet 38:556–560.
Service S, Sabatti C, Freimer N. 2007. Tag SNPs chosen from HapMap

perform well in several population isolates. Genet Epidemiol

31:189–194.

Streeten EA, McBride DJ, Pollin TI, Ryan K, Shapiro J, Ott S,

Mitchell BD, Shuldiner AR, O’Connell JR. 2006. Quantitative trait

loci for BMD identified by autosome-wide linkage scan to

chromosomes 7q and 21q in men from the Amish Family

Osteoporosis Study. J Bone Miner Res 21:1433–1442.

Thompson EE, Sun Y, Nicolae D, Ober C. 2009. Shades of gray: a

comparison of linkage disequilibrium between Hutterites and

Europeans. Genet Epidemiol, DOI: 10.1002/gepi.20442.

Wang N, Akey JM, Zhang K, Chakraborty R, Jin L. 2002. Distribution

of recombination crossovers and the origin of haplotype blocks: the

interplay of population history, recombination, and mutation. Am J

Hum Genet 71:1227–1234.
Wang Y, O’Connell JR, McArdle PF, Wade JB, Dorf f SE, Shah SJ, Shi X,

Pan L, Rampersaud E, Shen H, Kim JD, Subramanya AR,

Steinle NI, Parsa A, Ober CC, Welling PA, Chakravarti A,

Weder AB, Cooper RS, Mitchell BD, Shuldiner AR, Chang YP.

2009. Whole-genome association study identifies STK39 as a

hypertension susceptibility gene. Proc Natl Acad Sci USA 106:6.

Wigginton JE, Cutler DJ, Abecasis GR. 2005. A note on exact tests of

Hardy-Weinberg equilibrium. Am J Hum Genet 76:887–893.

Willer CJ, Scott LJ, Bonnycastle LL, Jackson AU, Chines P, Pruim R,

Bark CW, Tsai YY, Pugh EW, Doheny KF, Kinnunen L, Mohlke KL,

Valle TT, Bergman RN, Tuomilehto J, Collins FS, Boehnke M. 2006.

Tag SNP selection for Finnish individuals based on the CEPH Utah

HapMap database. Genet Epidemiol 30:180–190.

Wright S. 1922. Coef ficients of inbreeding and relationship. Am Nat

56:330–338.

APPENDIX

Summary and percentage of X chromosomes are given in
Tables AI and AII.

TABLE AI. Summary of X chromosome SNPs

OOA CEU Overlap

Total genotyped 10,525 10,525 10,525
41 duplicate inconsistencya 1,061 NA NA
45% missing datab 547 461 261
Mendelian inconsistenciesb,c 44 246 10
Po10-6 for HWE testd 0 0 0

Passed QC filtere 9,139 10,064 8,972
Passed QC in both OOA and CEU

Monomorphicd 2,272 1,905 1,805
Polymorphicd

MAFZ0.05 5,763 6,106 5,516
MAFZ0.10 4,971 5,376 4,449
MAFZ0.20 3,571 3,925 2,929

OOA, Old Order Amish; CEU, US Utah residents from HapMap; MAF, minor allele frequency. SNPs that failed a QC measure in either
sample were excluded from further analysis, and SNPs with MAFZ0.05 passing QC in both samples (n 5 5,516) were used for LD analysis.
aBased on the 61 OOA individuals who were also genotyped on the Affymetrix 5.0 array; SNPs with more than one duplicated genotype
discrepancy were excluded.
bBased on 837 OOA and 90 CEU individuals (30 trios).
cSNPs with 45 and 41 Mendelian inconsistencies in OOA and CEU, respectively.
dBased on 60 unrelated individuals (30 men and 30 women) from each sample.
eSNPs may fail QC in more than one way, so rows do not sum to the subtotal passing QC.

TABLE AII. Percentage of X chromosome SNP pairsa showing no evidence of recombination (D0 5 1), perfect LD (r2 5 1),
or where useful LD is observed (r2Z0.8)

D0 5 1 r2 5 1 r2Z0.8

Inter-SNP distance (kb) OOA CEU OOA CEU OOA CEU

r10 88 85 39 35 51 49
10–20 72 64 23 19 34 31
20–50 60 48 12 9 21 18
50–100 44 31 6 3 11 10
100–200 31 19 3 1 6 4
200–500 22 11 1 0 2 1
500–1,000 18 7 0 0 0 0
1,000–2,000 17 7 0 0 0 0
2,000–5,000 15 7 0 0 0 0
5,000–10,000 13 7 0 0 0 0

OOA, Old Order Amish (n 5 60); CEU, US Utah residents from HapMap (n 5 60).
aRestricted to SNPs with minor allele frequency Z0.05 in both samples (n 5 5,516).

150 Van Hout et al.

Genet. Epidemiol.