Content uploaded by Lindsay Farrer
Author content
All content in this area was uploaded by Lindsay Farrer on Jun 09, 2014
Content may be subject to copyright.
Nature GeNetics VOLUME 45 | NUMBER 11 | NOVEMBER 2013 1375
Macular degeneration is a common cause of blindness in the
elderly. To identify rare coding variants associated with a large
increase in risk of age-related macular degeneration (AMD),
we sequenced 2,335 cases and 789 controls in 10 candidate
loci (57 genes). To increase power, we augmented our control
set with ancestry-matched exome-sequenced controls. An
analysis of coding variation in 2,268 AMD cases and 2,268
ancestry-matched controls identified 2 large-effect rare
variants: previously described p.Arg1210Cys encoded
in the CFH gene (case frequency (fcase) = 0.51%; control
frequency (fcontrol) = 0.02%; odds ratio (OR) = 23.11) and
newly identified p.Lys155Gln encoded in the C3 gene
(fcase = 1.06%; fcontrol = 0.39%; OR = 2.68). The variants
suggest decreased inhibition of C3 by complement factor H,
resulting in increased activation of the alternative
complement pathway, as a key component of disease biology.
Genetic and environmental factors contribute to AMD1,2, a major
cause of vision loss in elderly individuals3. Pioneering discovery
of association of AMD with complement factor H (encoded by
CFH4–6) was quickly followed by the identification of additional
susceptibility loci that now include ARMS2-HTRA1 (refs. 7,8)
and complement genes C3, C2-CFB and CFI9–12. Genome-wide
association studies (GWAS) of AMD cases and controls have now
identified common susceptibility variants at ~20 different loci13,14
and have begun to uncover specific cellular pathways involved in
AMD biology.
Whereas common variants tag an associated genomic region, rare
coding variants can provide more specific clues about the under-
lying disease mechanism15. For example, rare variant p.Arg1210Cys
encoded in the CFH gene was recently associated with a large increase
in AMD risk using targeted sequencing of rare CFH risk haplotypes16.
The resulting altered protein has decreased binding to C3b, C3d,
heparin and endothelial cells17–19. A reduction in the ability of CFH
to inactivate C3, leading to increased cell killing activity of the com-
plement pathway, could contribute to AMD, representing a much
more specific and testable hypothesis about disease mechanism than
provided by common CFH variants whose mechanistic consequences
are unclear.
To systematically identify rare, large-effect variants, we carried out
targeted sequencing of eight AMD risk loci identified in GWAS20 (near
CFH, ARMS2, C3, C2-CFB, CFI, CETP, LIPC and TIMP3-SYN3) and
two candidate regions (LPL and ABCA1) (Supplementary Table 1).
We resequenced these regions in 3,124 individuals (2,335 cases and
789 controls) recruited in ophthalmology clinics at the University
of Michigan and the University of Pennsylvania and in Age-Related
Eye Disease Study (AREDS) participants20,21. We enriched genomic
targets using a set of 150-bp probes designed by Agilent Technologies
and generated sequence data on Illumina Genome Analyzer and
HiSeq instruments. The 10 loci comprised 115,596 nucleotides of
protein-coding sequence and totaled 2,757,914 nucleotides overall.
We designed probes to capture 111,592 protein-coding nucleotides
(96.5% of coding sequence) and 966,607 nucleotides overall (35.1%
of the locus sequence), generating an average of 123,221,974 mapped
Identification of a rare coding variant in complement 3
associated with age-related macular degeneration
Xiaowei Zhan1,39, David E Larson2,39, Chaolong Wang1,3,39, Daniel C Koboldt2, Yuri V Sergeev4,
Robert S Fulton2, Lucinda L Fulton2, Catrina C Fronick2, Kari E Branham5, Jennifer Bragg-Gresham1,
Goo Jun1, Youna Hu1, Hyun Min Kang1, Dajiang Liu1, Mohammad Othman5, Matthew Brooks6,
Rinki Ratnapriya6, Alexis Boleda6, Felix Grassmann7, Claudia von Strachwitz8, Lana M Olson9,10,
Gabriëlle H S Buitendijk11,12, Albert Hofman12,13, Cornelia M van Duijn12, Valentina Cipriani14,15,
Anthony T Moore14,15, Humma Shahid16,17, Yingda Jiang18, Yvette P Conley19, Denise J Morgan20,
Ivana K Kim21, Matthew P Johnson22, Stuart Cantsilieris23, Andrea J Richardson23, Robyn H Guymer23,
Hongrong Luo24,25, Hong Ouyang24,25, Christoph Licht26, Fred G Pluthero27, Mindy M Zhang24,25,
Kang Zhang24,25, Paul N Baird23, John Blangero22, Michael L Klein28, Lindsay A Farrer29–33,
Margaret M DeAngelis20, Daniel E Weeks18,34, Michael B Gorin35, John R W Yates14–16, Caroline C W Klaver11,12,
Margaret A Pericak-Vance36, Jonathan L Haines9,10, Bernhard H F Weber7, Richard K Wilson2,
John R Heckenlively5, Emily Y Chew37, Dwight Stambolian38, Elaine R Mardis2,40, Anand Swaroop6,40 &
Goncalo R Abecasis1,40
A full list of author affiliations appears at the end of the paper.
Received 26 January; accepted 19 August; published online 15 September 2013; doi:10.1038/ng.2758
LETTERS
npg © 2013 Nature America, Inc. All rights reserved.
1376 VOLUME 45 | NUMBER 11 | NOVEMBER 2013 Nature GeNetics
LETTERS
bases of on-target sequence per individual (127.5× average depth
when counting bases with quality of >20 in reads with mapping quality
of >30 after duplicate read removal); 98.49% of sites with designed
probes were covered at >10× depth. We applied variant calling tools
and quality control filters similar to those used to analyze National
Heart, Lung, and Blood Institute (NHLBI) Exome Sequencing Project
(ESP) data22,23 (Supplementary Table 2). We identified an average of
1,714 non-reference sites in each sequenced individual. In total, we
identified 31,527 single-nucleotide variants, of which 18,956 were not
in dbSNP135. Discovered sites included 834 synonymous variants,
1,379 nonsynonymous variants and 43 nonsense variants, most of
which were extremely rare (Supplementar y Table 3). For 13 samples
sequenced in duplicate, genotype concordance was 99.82% (when
depth was >10×). For 908 samples previously examined with GWAS
arrays20, sequencing-based genotypes were 98.99% concordant with
array-based calls (again, when depth was >10×).
In an initial comparison of AMD cases and controls (Supplementary
Table 4), no rare coding variants with frequency of <1% reached
experiment-wide significance (P < 0.05/31,527 = 1.6 × 10−6 when
including all discovered variants or P < 0.05/1,422 = 3.5 × 10−5 when
considering only protein-altering variants), although several showed
encouraging patterns of association. For example, the rare variant
p.Arg1210Cys encoded in the CFH gene was observed in 23 of the 2,335
sequenced cases but in none of the 789 sequenced controls (exact test P =
0.0025). Common variants in several loci exhibited strong evidence
of association, including in CFH (peak variant rs9427642: fcase = 12%;
fcontrol = 27%; P value = 2.52 × 10−48), ARMS2 (rs10490924:
fcase = 33%; fcontrol = 18%; P value = 5.48 × 10−27), C3 (rs2230199:
fcase = 25%; fcontrol = 17%; P value = 3.94 × 10−9) and C2-CFB
(rs556679: fcase = 7%; fcontrol = 12%; P value = 1.32 × 10−10).
A key requirement for establishing significance of rare disease-
associated variants is the availability of sufficient numbers of control
samples. To increase power, we sought to identify additional controls
and focused on samples from NHLBI ESP23, which sequenced 15,336
genes across 6,515 individuals. Sequence data for our samples and the
NHLBI ESP samples were analyzed with the same analysis pipeline,
which minimized potential differences due to heterogeneity in analy-
sis tools and parameters. To further avoid artifacts from sequencing
and variant calling, we restricted our analysis to sites within regions
targeted in both sequencing experiments, genotyped and covered with
>10 reads in >90% of the samples examined in each project and >5 bp
away from insertion-deletion polymorphisms catalogued by the 1000
Genomes Project24. Because careful matching of genetic ancestry is criti-
cal for rare variant association studies24,25, we selected an ancestry-
matched subset of our samples and of samples from NHLBI ESP. We
used principal-component analysis (PCA) to construct a genetic ances-
try map of the world with samples from the Human Genome Diversity
Project, each genotyped at 632,958 SNPs26. If GWAS array genotypes
were available for our samples and for the NHLBI ESP samples,
it would be straightforward to place the samples directly on this genetic
ancestry map. Using targeted sequence data, however, the analysis is
more challenging: targeted regions include too few variants to accu-
rately represent global ancestry, and off-target regions are covered too
poorly, precluding estimation of the accurate genotypes needed for
standard PCA. Thus, we relied on the new LASER algorithm (C.W.,
X.Z., J.B.-G., H.M.K., D.S. et al., unpublished data) to localize each
sequenced sample on a predefined genetic ancestry map of the world.
The method can accurately place individuals on this worldwide ances-
try map with <0.05× average coverage of the genome and is thus ideal
for targeted sequence data, such as ours and the NHLBI ESP data,
which have average off-target coverage of ~0.23× and ~0.90×, respec-
tively (see Supplementary Fig. 1a,b,e,f, which show that the PCA
coordinates inferred using 0.10× genome coverage or using GWAS
array genotypes are highly similar). We focused on samples where
PCA coordinates could be estimated confidently (Procrustes similarity
larger than 0.95; Online Methods) and used a greedy algorithm to
match cases and controls on the basis of estimated genetic ancestry.
As shown in the Online Methods, alternative matching algorithms did
not alter our conclusions. After matching, we focused on a set of 2,268
AMD cases and 2,268 controls that were ancestry-matched one to one
(Supplementary Fig. 1c,g). Because AMD phenotype information
was not available for most controls, we expect that a small proportion
may eventually develop disease; however, this should not affect power
substantially27 . After matching case-control samples, we excluded
1 variant with Hardy-Weinberg equilibrium test P value < 1 × 10−6
and focused our analysis on 430 protein-changing variants in regions
that were targeted and deeply sequenced in both experiments as well
as being far away from insertion-deletion polymorphisms.
In this expanded analysis (Table 1), common variant signals at
all loci increased in significance (in comparison to what is shown in
Supplementary Table 4). In addition, two rare coding variants exhibited
association with P < 0.01. The first variant was p.Arg1210Cys encoded
in the CFH gene (observed in 1 control and 23 cases; OR = 23.11; exact
P = 2.9 × 106), providing strong support for the original report16.
Table 1 Summary association results for 2,268 sequenced AMD cases and 2,268 sequenced controls
Frequency (alt allele)
SNP Chromosome Position (bp) Nearest gene Consequence Alleles (ref/alt) Cases Controls OR P value Conditional P valuea
Common variant hits
rs1061170 1 196659237 CFH p.His402Tyr C/T 0.478 0.623 0.555 1.01 × 10−43
rs438999 6 31928306 SKIV2L p.Gln151Arg A/G 0.058 0.098 0.566 1.26 × 10−12
rs10490924 10 124214448 ARMS2 p.Ala69Ser G/T 0.329 0.197 1.990 1.04 × 10−45
rs2230199 19 6718387 C3 p.Arg102Gly G/C 0.253 0.206 1.300 1.58 × 10−7
Rare variant hits (MAF < 1%; marginal and conditional P < 0.01 after conditioning on nearby common variants)
rs121913059 1 196716375 CFH p.Arg1210Cys C/T 0.005 0.000 23.11 2.9 × 10−6 6.0 × 10−4
(rs1061170)
rs147859257 19 6718146 C3 p.Lys155Gln T/G 0.011 0.004 2.68 2.7 × 10−4 2.8 × 10−5
(rs2230199)
Samples in this expanded analysis include our sequenced AMD samples and genetically matched controls, sequenced by us or by the NHLBI ESP. The top coding variant in each
locus is included in this table when P < 1 × 10−6. Rare coding variants are included when the corresponding P value for conditional or marginal analysis was less than 1 × 10−4.
All P values were calculated using exact logistic regression. Ref, reference; alt, alternative; MAF, minor allele frequency.
aFor rare variants, we re-evaluated statistical significance after adjusting for the top common variant in the locus to avoid shadow signals driven by linkage disequilibrium. The variant used for
conditioning is named (in parentheses).
npg © 2013 Nature America, Inc. All rights reserved.
Nature GeNetics VOLUME 45 | NUMBER 11 | NOVEMBER 2013 1377
LETTERS
The second variant was p.Lys155Gln encoded in the C3 gene
(observed in 18 controls and 48 cases; OR = 2.68; exact P = 2.7 × 104;
see Supplementary Fig. 1d,h for carrier ancestry distribution). When
controlling for a previously described common variant signal nearby,
rs2230199 (fcontrol = 20.63%; fcase = 25.26%; marginal exact P = 1.8 ×
10−7; OR = 1.31), the evidence for association with p.Lys155Gln
increased slightly (conditional OR = 2.91; exact P = 2.8 × 10−5).
Inspection of the raw read data showed that the variant was well sup-
ported and was unlikely to be an artifact of sequencing or alignment,
a result further confirmed by Sanger sequencing (Supplementary
Figs. 2–4). Finally, in an examination of our sequenced samples and
available whole-genome sequences (Online Methods), we observed
no additional variants in strong linkage disequilibrium with the
mutation encoding p.Lys155Gln that might account for the associa-
tion signal. Analysis with burden tests, which jointly evaluate evi-
dence for association with rare variants at each gene, identified no
significant association signals (Supplementary Fig. 5)28–30.
To confirm the signal corresponding to p.Lys155Gln, we genotyped
additional samples totaling 4,526 cases and 3,787 controls and, again,
observed strong association (fcontrol = 0.5%; fcase = 1.3%; follow-up
P = 7.7 × 10−7; combined P = 1.1 × 10−9; Table 2). In addition, we
genotyped 471 families with multiple AMD cases to identify 18
nuclear families where the mutation encoding p.Lys155Gln segre-
gates. These families included 49 affected individuals, with at least
1 individual carrying an allele encoding p.Lys155Gln, and, adjusting
for ascertainment, we estimated that 75% of
the first-degree relatives of a p.Lys155Gln
carrier who also had AMD would carry the
variant, consistent with an OR of ~3 (Online
Methods and Supplementary Table 5).
Further strong evidence for association of
this variant with macular degeneration is
provided in independent work by deCODE
Genetics31 examining 1,143 Icelandic macular
degeneration cases and 51,435 Icelandic con-
trols (fcontrol = 0.55%; OR = 3.45; deCODE
P = 1.1 × 10−7; combined P = 1.6 × 10−15).
In 1,606 directly genotyped cases of macu-
lar degeneration from AREDS2 (ref. 32), the
variant had a frequency of 1.77%, similar to
our sequenced AMD cases (1.10%) and our
follow-up AMD cases (1.30%) and notably
higher than our sequenced controls (0.30%),
our genotyped controls (0.50%), NHLBI ESP
participants with primarily European ancestry
(0.40%) and deCODE controls (0.55%).
We found no evidence of the p.Lys155Gln
variant in a small sample of individuals
with atypical hemolytic uremic syndrome (aHUS; n = 53), a rare
disorder whose genetic risk factors partially overlap with those of
macular degeneration.
We next investigated the potential functional consequences of the
p.Lys155Gln variant in silico. On the basis of protein crystallography,
the model in Figure 1 shows that CFH variant p.Arg1210Cys
(OR = 23.11), C3 variant p.Lys155Gln (OR = 2.91) and C3 variant
p.Arg102Gly (OR = 1.31) all map near the surface where CFH and
C3b interact, suggesting that they might affect binding of complement
factor H to C3b. CFH inhibits C3b and limits the immune responses
mediated by the alternative complement pathway. We hypothesize that
p.Lys155Gln and p.Arg102Gly affect binding of the first macroglobular
domain of C3 to CFH and thus interfere with inactivation of the alter-
native complement pathway, a hypothesis that must be confirmed
Sushi-20
Sushi-19
Sushi-18
MG-2
MG-1
Arg1210C
Arg102G
Lys155Q
Figure 1 C3 variants p.Arg102Gly and p.Lys155Gln and CFH variant
p.Arg1210Cys are in the interaction domains of the first α-macroglobular
domains of C3b and CFH, respectively. A fragment of the crystal structure
of the four Sushi domains of CFH (purple; one not shown for clarity) in a
complex with complement fragment C3b (Protein Data Bank (PDB) 2wii)
was used to explore the effect of disease-associated nonsynonymous
changes. CFH residues 987–1230 were used to generate the structure
with the first four Sushi domains from 2wii serving as a structural
template (light purple, with cysteine residue side chains in yellow).
The C-terminal Sushi domains were docked to the binding site in C3b.
The first two α-macroglobulin domains of C3b, MG-1 and MG-2, are
shown in green and cyan, respectively. The locations of the p.Arg102Gly,
p.Lys155Gln and p.Arg1210Cys alterations are marked in red.
Table 2 Follow-up genotyping summary and meta-analysis summary
Controls Cases
Sample set NMAF NMAF P value
Discovery sample
Sequenced samples (N = 4,536) 2,268 0.004 2,268 0.011 2.7 × 10−4
Follow-up samples
Germany: University of Regensburg (N = 2,976) 1,147 0.006 1,829 0.016 1.7 × 10−3
United States: Vanderbilt/Miami (N = 1,819) 726 0.004 1,093 0.007 3.5 × 10−1
Netherlands: Rotterdam Study (N = 1,409) 1,280 0.005 129 0.031 1.5 × 10−4
UK: Cambridge AMD Study (N = 1,279) 423 0.006 856 0.015 6.2 × 10−2
United States: University of California,
Los Angeles/University of Pittsburgh (N = 830)
211 0.004 619 0.017 8.3 × 10−4
deCODE study
deCODE discovery sample (N = 52,578) 51,435 0.005 1,143 –a1.1 × 10−7
Meta-analysis
All follow-up samples (N = 8,313) 3,787 0.005 4,526 0.013 7.7 × 10−7
Discovery and all follow-up samples (N = 12,849) 6,055 0.005 6,794 0.013 1.1 × 10−9
Discovery, all follow-up and deCODE samples (N = 65,427) 57,490 0.005 7,937 −a1.6 × 10−15
The table includes the number of cases and controls in each comparison, the corresponding allele frequency for the
allele encoding p.Lys155Gln in each set of samples and the P value for a comparison of allele frequencies in cases
and controls. Meta-analysis P values were calculated using Stouffer’s method.
aMAF values are unavailable for imputed cases from the deCODE study.
npg © 2013 Nature America, Inc. All rights reserved.
1378 VOLUME 45 | NUMBER 11 | NOVEMBER 2013 Nature GeNetics
LETTERS
experimentally33. Interestingly, the three variants (p.Arg102Gly and
p.Lys155Gln in C3 and p.Arg1210Cys in CFH) all involve the replace-
ment of a positively charged residue.
In summary, our work and that described in the companion paper
identify p.Lys155Gln as a rare C3 variant associated with ~2.91-fold
increased risk of macular degeneration. Together with rare CFH
variant p.Arg1210Cys and previously described common C3 variant
p.Arg102Gly, p.Lys155Gln may reduce binding of CFH to C3b, inhib-
iting the ability of CFH to inactivate the alternative complement path-
way. Clarifying the mechanistic impact of p.Lys155Gln is likely to be
challenging, as illustrated by contradictory results from previous func-
tional follow-up studies of AMD-associated loci34–36, but functional
studies of complement activity suggest potential next steps33,37. Our
work relied on targeted sequencing of GWAS-identified loci, genetic
ancestry matching of our sequenced samples to additional sequenced
controls analyzed with the same variant calling and filtering tools,
focused analysis of regions deeply sequenced in both our project and
previously sequenced controls, and avoidance of common calling
artifacts near insertion-deletion polymorphisms. The use of publicly
available samples to augment control sets may be useful in many tar-
geted sequencing studies, but the strictness of matching and variant
filtering required to prevent false positive findings due to population
stratification and/or sequence analysis artifacts are areas deserving
of further study. As the number of sequenced human genomes and
exomes grows, we expect that the usefulness of the approach will
grow, making it possible to match multiple controls to each case and
to focus on progressively finer ancestry matches. Although our results
emphasize that large sample sizes will be required for rare variant
studies of complex human traits, they also show the promise of these
studies for clarifying disease biology.
URLs. LASER software for estimation of genetic ancestry can be
obtained from http://genome.sph.umich.edu/wiki/LASER. UMAKE
and GotCloud tools for variant calling can be obtained from http://
genome.sph.umich.edu/wiki/UMAKE and http://genome.sph.
umich.edu/wiki/GotCloud, respectively. The QPLOT tool for asses-
ing sequence quality can be obtained from http://genome.sph.umich.
edu/wiki/QPLOT.
METHODS
Methods and any associated references are available in the online
version of the paper.
Note: Any Supplementary Information and Source Data files are available in the
online version of the paper.
ACKNOWLEDGMENTS
We thank all study participants for their generous volunteering. We thank B. Li,
W. Chen, C. Sidore, T. Teslovich, L. Fritsche and M. Boehnke for useful discussion
and suggestions. This project was supported by grants from the US National
Institutes of Health (National Eye Institute, National Human Genome Research
Institute; grants EY022005, HG007022, HG005552, EY016862, U54HG003079
and EY09859); the Medical Research Council, UK (grant G0000067); the Deutsche
Forschungsgemeinschaft (grant WE1259/19-2); the Alcon Research Institute;
The UK Department of Health’s National Institute for Health Research (NIHR)
Biomedical Research Centre for Ophthalmology at Moorfields Eye Hospital and
the UCL Institute of Ophthalmology; Research to Prevent Blindness (New York);
the Thome Memorial Foundation; the Harold and Pauline Price Foundation; and
the National Health and Medical Research Council of Australia (NHMRC) Clinical
Research Excellence (grant 529923, NHMRC practitioner fellowship 529905 and
NHMRC Senior Research Fellowship 1028444). The study was also supported
by the Intramural Research Program (Computational Medicine Initiative) of the
National Eye Institute. The Centre for Eye Research Australia (CERA) receives
operational infrastructure support from the Victorian Government. The views
expressed in the publication are those of the authors and not necessarily those of
their employers or the funders.
AUTHOR CONTRIBUTIONS
R.K.W., J.R.H., E.Y.C., D.S., E.R.M., A.S. and G.R.A. conceived, designed and
supervised the experiments. X.Z. and G.R.A. wrote the initial version of the
manuscript. X.Z., D.E.L., C.W. and D.C.K. analyzed the data. D.E.L., D.C.K.,
R.S.F., L.L.F. and C.C.F. supervised data generation. C.W. developed statistical
methodology. Y.V.S. analyzed protein structures. K.E.B. supervised sample and
data collection. J.B.-G., G.J., Y.H., H.M.K. and D.L. contributed data and analysis
tools. M.B., R.R. and A.B. assisted in laboratory experiments. M.O. and F.G.
carried out experimental studies (genotyping and data analysis) for the Michigan
and Regensburg samples, respectively. C.v.S. recruited the family members of
sporadic AMD cases and controls and collected peripheral blood samples for the
Regensburg study. L.M.O., M.A.P.-V. and J.L.H. provided results and analysis for
the Vanderbilt/Miami samples. G.H.S.B., A.H., C.M.v.D. and C.C.W.K. provided
results and analysis for samples from the Rotterdam Study, Erasmus Medical
Center. V.C., A.T.M., H.S. and J.R.W.Y. provided results and analysis for the
Cambridge AMD Study samples. Y.J., Y.P.C., D.E.W. and M.B.G. provided results
and analysis for the University of California, Los Angeles/University of Pittsburgh
samples. D.J.M., I.K.K., L.A.F. and M.M.D. provided results and analysis for the
Utah samples. M.P.J., J.B. and M.L.K. provided results and analysis for the Oregon
Health Sciences Center samples. S.C., A.J.R., R.H.G. and P.N.B. provided results
and analysis for the University of Melbourne samples. H.L., H.O., M.M.Z. and
K.Z. provided results and analysis for the University of California, San Diego
samples. C.L. and F.G.P. provided results and analysis for a cohort of individuals
with aHUS. B.H.F.W. was involved in the design and planning of the Southern
Germany AMD Study. B.H.F.W. participated in study coordination and critically
read the manuscript. All authors have critically commented on this manuscript.
COMPETING FINANCIAL INTERESTS
The authors declare competing financial interests: details are available in the online
version of the paper.
Reprints and permissions information is available online at http://www.nature.com/
reprints/index.html.
1. Priya, R.R., Chew, E.Y. & Swaroop, A. Genetic studies of age-related macular
degeneration: lessons, challenges, and opportunities for disease management.
Ophthalmology 119, 2526–2536 (2012).
2. Swaroop, A., Chew, E.Y., Rickman, C.B. & Abecasis, G.R. Unraveling a multifactorial
late-onset disease: from genetic susceptibility to disease mechanisms for age-related
macular degeneration. Annu. Rev. Genomics Hum. Genet. 10, 19–43 (2009).
3. Friedman, D.S. et al. Prevalence of age-related macular degeneration in the United
States. Arch. Ophthalmol. 122, 564–572 (2004).
4. Haines, J.L. et al. Complement factor H variant increases the risk of age-related
macular degeneration. Science 308, 419–421 (2005).
5. Edwards, A.O. et al. Complement factor H polymorphism and age-related macular
degeneration. Science 308, 421–424 (2005).
6. Klein, R.J. et al. Complement factor H polymorphism in age-related macular
degeneration. Science 308, 385–389 (2005).
7. Jakobsdottir, J. et al. Susceptibility genes for age-related maculopathy on
chromosome 10q26. Am. J. Hum. Genet. 77, 389–407 (2005).
8. Rivera, A. et al. Hypothetical LOC387715 is a second major susceptibility gene
for age-related macular degeneration, contributing independently of complement
factor H to disease risk. Hum. Mol. Genet. 14, 3227–3236 (2005).
9. Yates, J.R. et al. Complement C3 variant and the risk of age-related macular
degeneration. N. Engl. J. Med. 357, 553–561 (2007).
10. Gold, B. et al. Variation in factor B (BF) and complement component 2 (C2) genes is
associated with age-related macular degeneration. Nat. Genet. 38, 458–462 (2006).
11. Fagerness, J.A. et al. Variation near complement factor I is associated with risk of
advanced AMD. Eur. J. Hum. Genet. 17, 100–104 (2009).
12. Maller, J.B. et al. Variation in complement factor 3 is associated with risk of age-
related macular degeneration. Nat. Genet. 39, 1200–1201 (2007).
13. Fritsche, L.G. et al. Seven new loci associated with age-related macular degeneration.
Nat. Genet. 45, 433–439 (2013).
14. Arakawa, S. et al. Genome-wide association study identifies two susceptibility
loci for exudative age-related macular degeneration in the Japanese population.
Nat. Genet. 43, 1001–1004 (2011).
15. Nejentsev, S., Walker, N., Riches, D., Egholm, M. & Todd, J.A. Rare variants of
IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes.
Science 324, 387–389 (2009).
16. Raychaudhuri, S. et al. A rare penetrant mutation in CFH confers high risk of
age-related macular degeneration. Nat. Genet. 43, 1232–1236 (2011).
17. Józsi, M. et al. Factor H and atypical hemolytic uremic syndrome: mutations in the
C-terminus cause structural changes and defective recognition functions. J. Am.
Soc. Nephrol. 17, 170–177 (2006).
npg © 2013 Nature America, Inc. All rights reserved.
Nature GeNetics VOLUME 45 | NUMBER 11 | NOVEMBER 2013 1379
LETTERS
18. Manuelian, T. et al. Mutations in factor H reduce binding affinity to C3b and heparin
and surface attachment to endothelial cells in hemolytic uremic syndrome. J. Clin.
Invest. 111, 1181–1190 (2003).
19. Ferreira, V.P. et al. The binding of factor H to a complex of physiological polyanions
and C3b on cells is impaired in atypical hemolytic uremic syndrome. J. Immunol.
182, 7009–7018 (2009).
20. Chen, W. et al. Genetic variants near TIMP3 and high-density lipoprotein–associated
loci influence susceptibility to age-related macular degeneration. Proc. Natl. Acad.
Sci. USA 107, 7401–7406 (2010).
21. Age-Related Eye Disease Study Research Group. Risk factors associated with age-
related macular degeneration. A case-control study in the age-related eye disease
study: Age-Related Eye Disease Study Report Number 3. Ophthalmology 107,
2224–2232 (2000).
22. Tennessen, J.A. et al. Evolution and functional impact of rare coding variation from
deep sequencing of human exomes. Science 337, 64–69 (2012).
23. Fu, W. et al. Analysis of 6,515 exomes reveals the recent origin of most human
protein-coding variants. Nature 493, 216–220 (2013).
24. 1000 Genomes Project Consortium. An integrated map of genetic variation from
1,092 human genomes. Nature 491, 56–65 (2012).
25. Mathieson, I. & McVean, G. Differential confounding of rare and common variants
in spatially structured populations. Nat. Genet. 44, 243–246 (2012).
26. Li, J.Z. et al. Worldwide human relationships inferred from genome-wide patterns
of variation. Science 319, 1100–1104 (2008).
27. Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000
cases of seven common diseases and 3,000 shared controls. Nature 447, 661–678
(2007).
28. Li, B. & Leal, S.M. Methods for detecting associations with rare variants for common
diseases: application to analysis of sequence data. Am. J. Hum. Genet. 83,
311–321 (2008).
29. Price, A.L. et al. Pooled association tests for rare variants in exon-resequencing
studies. Am. J. Hum. Genet. 86, 832–838 (2010).
30. Wu, M.C. et al. Rare-variant association testing for sequencing data with the
sequence kernel association test. Am. J. Hum. Genet. 89, 82–93 (2011).
31. Helgason, H. et al. A rare nonsynonymous sequence variant in C3 confers high
risk of age-related macular degeneration. Nat. Genet. doi:10.1038/ng.2740
(15 September 2013).
32. AREDS2 Research Group. et al. The Age-Related Eye Disease Study 2 (AREDS2):
study design and baseline characteristics (AREDS2 report number 1). Ophthalmology
119, 2282–2289 (2012).
33. Heurich, M. et al. Common polymorphisms in C3, factor B, and factor H collaborate
to determine systemic complement activity and disease risk. Proc. Natl. Acad. Sci.
USA 108, 8761–8766 (2011).
34. Fritsche, L.G. et al. Age-related macular degeneration is associated with an unstable
ARMS2 (LOC387715) mRNA. Nat. Genet. 40, 892–896 (2008).
35. Kanda, A. et al. A variant of mitochondrial protein LOC387715/ARMS2, not HTRA1,
is strongly associated with age-related macular degeneration. Proc. Natl. Acad. Sci.
USA 104, 16227–16232 (2007).
36. Dewan, A. et al. HTRA1 promoter polymorphism in wet age-related macular
degeneration. Science 314, 989–992 (2006).
37. Sánchez-Corral, P. et al. Structural and functional characterization of factor H
mutations associated with atypical hemolytic uremic syndrome. Am. J. Hum. Genet.
71, 1285–1295 (2002).
1Department of Biostatistics, Center for Statistical Genetics, University of Michigan School of Public Health, Ann Arbor, Michigan, USA. 2The Genome Institute,
Washington University School of Medicine, St. Louis, Missouri, USA. 3Department of Biostatistics, Harvard School of Public Health, Boston, Massachusetts,
USA. 4Ophthalmic Genetics and Visual Function Branch, National Eye Institute, Bethesda, Maryland, USA. 5Department of Ophthalmology and Visual Sciences,
University of Michigan Kellogg Eye Center, Ann Arbor, Michigan, USA. 6Neurobiology–Neurodegeneration and Repair Laboratory, National Eye Institute, US National
Institutes of Health, Bethesda, Maryland, USA. 7Institute of Human Genetics, University of Regensburg, Regensburg, Germany. 8Southwest Eye Center, Stuttgart,
Germany. 9Center for Human Genetics Research, Vanderbilt University Medical School, Nashville, Tennessee, USA. 10Department of Molecular Physiology and
Biophysics, Vanderbilt University Medical School, Nashville, Tennessee, USA. 11Department of Ophthalmology, Erasmus Medical Center, Rotterdam, The Netherlands.
12Department of Epidemiology, Erasmus Medical Center, Rotterdam, The Netherlands. 13Netherlands Consortium for Healthy Aging, Netherlands Genomics Initiative,
The Hague, The Netherlands. 14UCL Institute of Ophthalmology, University College London, London, UK. 15Moorfields Eye Hospital, London, UK. 16Department of
Medical Genetics, Cambridge Institute for Medical Research, University of Cambridge, Cambridge, UK. 17Cambridge University Hospitals National Health Service
(NHS) Foundation Trust, Cambridge, UK. 18Department of Biostatistics, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.
19Department of Health Promotion and Development, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA. 20Department
of Ophthalmology and Visual Sciences, John A. Moran Eye Center, University of Utah, Salt Lake City, Utah, USA. 21Retina Service and Ophthalmology, Harvard
Medical School, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts, USA. 22Texas Biomedical Research Institute, San Antonio, Texas, USA. 23Centre
for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, East Melbourne, Victoria, Australia. 24Department of Ophthalmology,
Shiley Eye Center, University of California, San Diego, La Jolla, California, USA. 25Institute for Genomic Medicine, University of California, San Diego, La Jolla,
California, USA. 26Department of Pediatrics, The Hospital for Sick Children, Toronto, Ontario, Canada. 27Program in Cell Biology, The Hospital for Sick Children,
Toronto, Ontario, Canada. 28Macular Degeneration Center, Casey Eye Institute, Oregon Health & Science University, Portland, Oregon, USA. 29Section on Biomedical
Genetics, Department of Medicine, Boston University Schools of Medicine and Public Health, Boston, Massachusetts, USA. 30Department of Epidemiology, Boston
University School of Public Health, Boston, Massachusetts, USA. 31Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts,
USA. 32Department of Neurology, Boston University School of Medicine, Boston, Massachusetts, USA. 33Department of Ophthalmology, Boston University School
of Medicine, Boston, Massachusetts, USA. 34Department of Human Genetics, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania,
USA. 35Department of Ophthalmology, Jules Stein Eye Institute, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, California,
USA. 36John P. Hussman Institute for Human Genomics, University of Miami Miller School of Medicine, Miami, Florida, USA. 37Division of Epidemiology and Clinical
Applications, National Eye Institute, US National Institutes of Health, Bethesda, Maryland, USA. 38Department of Ophthalmology and Human Genetics, University
of Pennsylvania Medical School, Philadelphia, Pennsylvania, USA. 39These authors contributed equally to this work. 40These authors jointly directed this work.
Correspondence should be addressed to G.R.A. (goncalo@umich.edu).
npg © 2013 Nature America, Inc. All rights reserved.
Nature GeNetics doi:10.1038/ng.2758
ONLINE METHODS
Study samples. Macular degeneration cases and controls were recruited at
ophthalmology clinics at the University of Michigan and the University of
Pennsylvania and through the AREDS, as previously described. For replica-
tion, we contacted members of the International AMD Genetics Consortium;
their samples are described in Fritsche et al.13. All participants provided
informed consent allowing for the collection of genetic data, and all data
contributors obtained approval from their local institutional review boards
before generating genetic data. Our discovery sample, with ~2,350 sequenced
cases and ~750 sequenced controls, provides 90% power to discover variants
with a frequency of 0.1% and an associated relative risk of 19.2 or greater
(similar to the p.Arg1210Cys variant in CFH) at significance level
α
= 0.00005,
which corresponds to an adjustment for the analysis of 1,000 independent
coding variants.
Sequence production and quality control. Illumina multiplexed libraries
were constructed according to the manufacturer’s protocol with modifications:
(i) DNA was fragmented using a Covaris E220 DNA Sonicator to range in size
between 100 and 400 bp, (ii) Illumina adaptor-ligated library fragments were
amplified in 4 50-µl PCR runs for 18 cycles, and (iii) Solid-Phase Reversible
Immobilization (SPRI) bead cleanup was used for enzymatic purification
and final library size selection targeting 300- to 500-bp fragments. Samples
were pooled in groups of 4–24 before hybridization. A custom targeted probe
set of 150-bp probes was designed (Agilent Technologies) and captured
0.97 Mb of sequence. The concentration of each captured library pool was deter-
mined through quantitative PCR (Kapa Biosystems) to produce cluster counts
appropriate for the Illumina Genome Analyzer IIx and HiSeq 2000 platforms.
We generated approximately 1.7 Gb of sequence per sample, covering 80% of
the targeted space at a depth of >20×. Reads were aligned to the NCBI37/hg19
reference sequence using Burrows-Wheeler Aligner (BWA)38 . Where pre-
existing genotype information was available, sample identity was confirmed
by comparing sequence data with pre-existing array data.
Quality control and variant calling. Quality control steps for all BAM files
included removal of duplicated reads; recalibration of base qualities39; gen-
eration of diagnostic graphs and evaluation of sequencing quality (QPLOT;
see URLs); and checks for DNA contamination40. After removing samples
with high contamination, unexpected relatedness or high discordance rate,
we retained 2,335 cases and 789 controls for an initial round of analysis. We
calculated the sequencing depth using reads with mapping quality of >30 and
bases with quality of >20. Across the 966,607-bp target region, we retained an
average of 123,221,974 bases per individual (127.5× average coverage). Within
targeted regions, 98.49% of the protein-coding exons had coverage of >10×.
We performed the variant calling step using UMAKE23. Genotype calling
and polymorphism discover y were attempted across the original target
± 50 bp. To remove low-quality variants, we excluded (i) sites with average depth of
<0.5 or >500; (ii) sites with evidence of strand bias or cycle bias; (iii) sites
within 5 bp of a 1000 Genomes Project indel; and (iv) sites with excess hetero-
zygosity. These filters excluded 15,219 low-quality variants. The transition-
transversion ratio (Ts/Tv) for the remaining 31,527 sites was 2.10. Concordance
rates between sequencing-based genotypes in 13 duplicates were 99.82% when
depth was >10×. Concordance with array-based genotypes20 was 98.99% when
depth was >10×.
Overall, 59.8% of discovered variants were newly identified (compared to
dbSNP135 and the 1000 Genomes Project). On average, each sample carried 40
synonymous variants, 34 nonsynonymous variants and 1 nonsense variant.
Initial analyses. We first performed single-variant association tests using
Fisher’s exact test. This analysis confirmed strong association for common
variants near the CFH, C2, ARMS2 and C3 genes. An initial examination of
rare variants suggested that some signals were shadows of common variants
with larger effects, so we focused on those signals where association remained
significant after accounting for nearby common variants. Conditional signals
were evaluated by exact logistic regression41,42. Three coding variants had
conditional exact P values < 0.01 (all also had marginal P values < 0.01).
Augmenting our sample. We sought ancestr y-matched controls among
samples sequenced in ESP. First, we used genome-wide reads to infer sample
ancestries on a worldwide population map. Briefly, we first generated a genetic
ancestry PCA space using genotyped reference samples (such as those from
the Human Genome Diversity Panel). Then, we generated a series of sample-
specific genetic ancestry PCA data that were calibrated to the exact sequencing
depth and coverage pattern of each sample and included the reference samples
together with a single sequenced sample. Finally, we transformed sample-
specific PCA coordinates onto the original map using Procrustes analysis.
This procedure generates a metric (Procrustes similarity) that summarizes the
similarity of reference sample placements using array genotypes to placements
using sequencing data, and we only considered samples where this metric was
>0.95 as candidates for matching. Second, we used a procedure inspired by
propensity score matching to pair cases and controls43. Briefly, this procedure
uses logistic regression to predict the probability that an individual is a case
using the four principal components of ancestry as predictors and disease
status as the outcome. This estimated probability of being a case for each
sample is a propensity score and can be used to match cases and controls. For
matching, we used a greedy algorithm to match cases and controls, allowing
matches when the respective propensity scores differed by <0.0001. An alter-
native matching algorithm that matched cases and controls mapping close
together in principal-component space according to the Euclidean distance
between them gave similar results (association at p.Lys155Gln had OR = 2.68;
exact P = 4.5 × 10−5 using Fisher’s exact test).
To avoid artifacts from variant calling, we applied very stringent filters to
both the AMD study and ESP study call sets. For both studies, we examined
only sites with call rates of >90% and Phred-scaled variant quality scores
of >30 that passed all study-specific quality control filters, had depth of
>10× for >90% of the samples in the AMD or ESP call sets and were >5 bp
from a 1000 Genomes Project indel. Primers used to confirm the presence
of the mutation encoding p.Lys155Gln by Sanger sequencing are given in
Supplementar y Table 6.
Analyses using the combined AMD and ESP data set. As in our initial
analysis, we first applied Fisher’s exact test for association with all vari-
ants. Next, we examined variants with frequency of <1% for which signal
remained significant after adjusting for common variants. This analysis
highlighted p.Arg1210Cys encoded by CFH and p.Lys155Gln encoded
by C3 (Fig. 1).
Linkage disequilibrium analysis. To search for variants that might explain the
signal encoding p.Lys155Gln, we evaluated linkage disequilibrium between
the variant encoding p.Lys155Gln and all variants within 1 Mb, both within
the samples sequenced for this experiment and also in preliminary whole-
genome sequence data for 600 individuals (300 macular degeneration cases
and 300 controls; A.S., D.S. and G.R.A., unpublished data). This analysis did not
find variants in strong linkage disequilibrium in the nearby region. The variant
was only present in one 1000 Genomes Project sample, which did not allow
for reliable estimates of linkage disequilibrium.
Segregation analysis. In a segregation analysis, one identifies probands who
carry p.Lys155Gln and then evaluates the probability that they transmit the
variant to affected relatives (under the null hypothesis, we would expect to
find the variant in 50% of the first-degree relatives of a carrier). We geno-
typed 471 pedigrees with multiple affected individuals. In each pedigree where
p.Lys155Gln was found in more than one affected individual, we selected the
nuclear family with the largest number of affected individuals. We recorded the
number of affected individuals (N) and the number of carriers of p.Lys155Gln
(C). Then, to average over possible choices of proband, we assigned each fam-
ily a weight of C/N (this is the probability that a randomly selected proband
in the family carries p.Lys155Gln) and then scored the number of affected
first-degree relatives (N – 1) and carriers among those (C – 1). The estimated
fraction of carriers among affected first-degree relatives of a proband was then
calculated by summing C/N × (C – 1) and C/N × (N – 1) over families and
taking the ratio of the two quantities.
npg © 2013 Nature America, Inc. All rights reserved.
Nature GeNetics
doi:10.1038/ng.2758
38. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler
transform. Bioinformatics 25, 1754–1760 (2009).
39. McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing
next-generation DNA sequencing data. Genome Res. 20, 1297–1303 (2010).
40. Jun, G. et al. Detecting and estimating contamination of human DNA samples in
sequencing and array-based genotype data. Am. J. Hum. Genet. 91, 839–848 (2012).
41. Cox, D.R. & Shell, E.J. Analysis of Binary Data 2nd edn. (CRC Press, New York,
1989).
42. Hirji, K.F., Mehta, C.R. & Patel, N.R. Computing distributions for exact
logistic-regression. J. Am. Stat. Assoc. 82, 1110–1117 (1987).
43. Rosenbaum, P.R. & Rubin, D.B. The central role of the propensity score in
observational studies for causal effects. Biometrika 70, 41–55 (1983).
npg © 2013 Nature America, Inc. All rights reserved.
