当前位置：文档库 › Exome sequencing identifies rare LDLR and APOA5 alleles conferring risk for myocardial infarction

Exome sequencing identifies rare LDLR and APOA5 alleles conferring risk for myocardial infarction

LETTER

doi:10.1038/nature13917 Exome sequencing identifies rare LDLR and APOA5 alleles conferring risk for myocardial infarction

A list of authors and their affiliations appears at the end of the paper

Myocardial infarction(MI),a leading cause of death around the world, displays a complex pattern of inheritance1,2.When MI occurs early in life,genetic inheritance is a major component to risk1.Previously, rare mutations in low-density lipoprotein(LDL)genes have been shown to contribute to MI risk in individual families3–8,whereas com-mon variants at more than45loci have been associated with MI risk in the population9–15.Here we evaluate how rare mutations con-tribute to early-onset MI risk in the population.We sequenced the protein-coding regions of9,793genomes from patients with MI at an early age(#50years in males and#60years in females)along with MI-free controls.We identified two genes in which rare coding-sequence mutations were more frequent in MI cases versus controls at exome-wide significance.Atlow-density lipoprotein receptor(LDLR), carriers of rare non-synonymous mutations were at4.2-fold increased risk for MI;carriers of null alleles at LDLR were at even higher risk (13-fold difference).Approximately2%of early MI cases harbour a rare,damaging mutation in LDLR;this estimate is similar to one made more than40years ago using an analysis of total cholesterol16. Among controls,about1in217carried an LDLR coding-sequence mutation and had plasma LDL cholesterol.190mg dl21.At apo-lipoprotein A-V(APOA5),carriers of rare non-synonymous muta-tions were at2.2-fold increased risk for MI.When compared with non-carriers,LDLR mutation carriers had higher plasma LDL choles-terol,whereas APOA5mutation carriers had higher plasma triglyc-erides.Recent evidence has connected MI risk with coding-sequence mutations at two genes functionally related to APOA5,namely lipo-protein lipase15,17and apolipoprotein C-III(refs18,19).Combined, these observations suggest that,as well as LDL cholesterol,disordered metabolism of triglyceride-rich lipoproteins contributes to MI risk. The US National Heart,Lung,and Blood Institute’s exome sequenc-ing project(ESP)sought to use exome sequencing as a tool to identify genes and mechanisms contributing to heart,lung and blood disorders. Within this program,we designed a discovery study for the extreme phenotype of early-onset MI(Fig.1),as heritability is substantially greater when MI occurs early in life1,2.From eleven studies,we identified1,088 cases with MI at an early age(MI in males#50years old and in females #60years old).As a comparison group,we selected978participants from prospective cohort studies who were of advanced age(males$60 years old or females$70years old)and free of MI.

We sequenced cases and controls to high coverage by performing solution-based hybrid selection of exons followed by massively parallel sequencing(see Methods)20.We performed several quality control steps to identify and remove outlier samples and variants(see Methods and Supplementary Figs1–13).Characteristics of the discovery set of1,027 cases and946controls are provided in Supplementary Tables1–3.Across the autosomes,each participant had an average of43nonsense,7,828 missense,92splice-site,189insertion or deletion(indel)frameshift,366 indel non-frameshift,and103non-synonymous singleton variants. We first tested whether low-frequency coding variants(defined here as a single nucleotide variant(SNV)or indel with minor allele frequency (MAF)between1%and5%)are associated with risk for MI in the dis-covery sequencing study.We observed no significant association of MI status with any individual variant(Supplementary Fig.14).We next evaluated the hypothesis that rare alleles(defined here as a SNV or indel with MAF,1%)collectively within a gene contribute to risk for MI(see Methods).We tested for an excess(or deficit)in cases versus controls of rare,non-synonymous mutations by aggregating together SNVs and indels with MAF,1%(‘T1’test)in each gene and comparing the counts in cases and controls21.Empirical P values were obtained using permutation.

The need to aggregate rare variants requires consideration of which variants to be studied together.Ideally,one would aggregate only harm-ful alleles and ignore benign alleles.To enrich for harmful alleles,we considered three sets of variants:(1)non-synonymous only;(2)a‘dele-terious(PolyPhen)’set consisting of non-synonymous after excluding missense alleles annotated as benign by PolyPhen-2HumDiv software; and(3)‘disruptive’mutations only(nonsense,indel frameshift,splice-site;also referred to as‘null’mutations).To account for multiple test-ing,we set exome-wide significance for this study at P5831027,a Bonferroni correction for the testing of,20,000genes and three variant sets.When the T1test was applied across these three sets of alleles in the discovery sequencing study,no gene-based association signal deviated from what we expected by chance(Supplementary Figs15–22).

Age

Discovery exome sequencing

Figure1|Overall design for the early-onset myocardial infarction study within the US National Heart,Lung,and Blood Institute’s exome sequencing project(ESP).Whole exome sequencing was performed in

1,973individuals from the phenotypic extremes.To test the hypothesis that low-frequency variants confer risk for myocardial infarction(MI),we performed follow-up statistical imputation and array-based genotyping of single nucleotide variants.To test the hypothesis that a burden of rare mutations in a gene confers risk for MI,we performed targeted re-sequencing and additional exome sequencing.

00M O N T H2014|V O L000|N A T U R E|1

We followed up on discovery sequencing results in four ways:(1)statistical imputation;(2)array-based genotyping using the Illumina HumanExome Beadchip (‘Exome’chip);(3)targeted re-sequencing;and (4)additional exome sequencing (Fig.1).Imputation and array-based genotyping were used to mainly evaluate low-frequency variants,whereas targeted re-sequencing and exome sequencing were used to test the role of rare mutations.

With the first and second follow-up approaches:imputation (n 564,132)and array-based genotyping (n 515,936),respectively,we did not identify novel low-frequency variants associated with MI or coronary artery disease (CAD)(see Methods,Supplementary Tables 4–7and Supplementary Figs 23–27).The top association results for SNVs from array-based genotyping are shown in Supplementary Table 8.

In the third follow-up approach,we re-sequenced several genes in additional cases and controls (see Methods,Supplementary Table 9).After sequencing the exons of APOA5in 6,721cases and 6,711controls,we identified 46unique non-synonymous or splice-site SNVs or indel

frameshifts with allele frequency ,1%(Supplementary Table 10).Based on these variants,we observed 93alleles in cases and 42alleles in con-trols (P 5531027;Table 1,Fig.2and Supplementary Table 10).This burden of rare mutation signal was primarily driven by mutations seen in one or two study participants (Fig.2and Supplementary Table 10).Carriers of a rare APOA5mutation had a 2.2-fold higher risk for MI/CAD than non-carriers (Table 1).

According to a recent report,consideration of variant sets based on multiple protein prediction algorithms might yield stronger association signals 22.Therefore,we investigated two additional variant sets:(1)‘dele-terious (broad)’as defined by nonsense,splice-site,indel frameshift,

A315V Q313X Q295X T292I R282C R289C

Q283R V281M E279D L277P P272Q A262S R259S E255G E255G E255K

I246M P215L V213M G185C G175A E168D E99K

E98D Q97X R94W R40M R40M D37E D37E M92fs R93Q L60H R94W Q97X Q95X E98D E99K A105S Q145R G164R Q145R S197G R245C A222V P215L H321L P319L G334V R343C L363Q

E81K A315V H321L L363R

93 mutations found in 6,721 cases Unique to cases Unique to controls

Observed in cases and controls 42 mutations found in 6,711 controls

Figure 2|Apolipoprotein A-V (APOA5)mutations discovered after

sequencing of 13,432individuals.Individual mutations (non-synonymous,indel frameshift and splice-site variants with minor allele frequency less than 1%)are depicted according to the genomic position along the length of the APOA5gene starting at the 59end (top).The number of circles on the left and right represents the number of times that mutation is observed in cases or controls,respectively.Dashed lines across the gene connect the same mutation seen in both cases and controls.Mutations are shaded in red (observed in cases only),blue (observed in controls only)or yellow (observed in both cases and controls).

Table 1|Association of a burden of rare mutations in APOA5with risk for early-onset myocardial infarction or coronary artery disease

Mutation set

n cases/controls

T1cases

T1controls

Freq cases (%)

Freq control (%)

Non-synonymous

6,721/6,7119342 1.40.63 2.2531027Deleterious (PolyPhen)6,721/6,71163310.940.46 2.0631025Deleterious (broad)6,721/6,7116831 1.00.46 2.2231025Deleterious (strict)6,721/6,7111030.150.045 3.30.008Disruptive 6,721/6,711920.130.03 4.5

0.007

Summary allele counts and carrier frequencies are shown.Only SNVs and indels with minor allele frequency less than 1%were considered in burden analysis.Deleterious (PolyPhen)as defined by nonsense,splice-site,indel frameshift,and missense annotated as ‘possibly damaging’or ‘probably damaging’by PolyPhen-2HumDiv software;‘deleterious (broad)’as defined by nonsense,splice-site,indel frameshift,and missense annotated as deleterious by at least one of the five protein prediction algorithms of LRT score,MutationTaster,PolyPhen-2HumDiv,PolyPhen-2HumVar and SIFT;‘deleterious

(strict)’as defined by nonsense,splice-site,indel frameshift,and missense annotated as deleterious by all five protein prediction algorithms;Disruptive defined as nonsense,splice-site or indel frameshift;T1:alleles from SNVs or indels with minor allele frequency less than 1%;Freq (%):percentage of

cases or controls carrying a T1allele;OR:odds ratio.

Y419X

R350X

S70fs Q33X Q102X c.313+1G>A

S99X E140X C155X

c.1358+1G>A

Y489X

W533X

c.1586+1G>A

c.1845+2T>C

C68IX

I687fs

Q770X L804fs

C143X

24 mutations found in 4,703 cases

Unique to cases Unique to controls

2 mutations found in 5,090 controls

L D L c h o l e s t e r o l (m g d l –1)

Variant class

100

150

200

250

300350

o n -c

a r r i e r s D e l e t e r i o u s (P o l y P h e n )D e l e t e r i o u s (B r o a d )D e l e t e r i o u s (S t r i c t )D i s r u p t i v e

Figure 3|Low-density lipoprotein receptor (LDLR )mutations discovered after sequencing 9,793individuals.a ,Individual disruptive mutations

(nonsense,indel frameshift,and splice-site variants with minor allele frequency less than 1%)are depicted according to the genomic position along the length of the LDLR gene starting at the 59end (top).The number of circles on the left and right represents the number of times that mutation is observed in cases or controls,respectively.Mutations are shaded in red if observed in cases only or blue if observed in controls only.b ,LDL cholesterol level as observed in different LDLR gene mutation annotation categories.Mean (height of bar)and 95%confidence intervals (error bars)are shown.Each individual is categorized based on mutation annotation as follows.Non-carriers:carriers without a missense or disruptive mutation;deleterious (PolyPhen)as defined by nonsense,splice-site,indel frameshift,and missense annotated as ‘possibly damaging’or ‘probably damaging’by PolyPhen-2HumDiv software;

‘deleterious (broad)’as defined by nonsense,splice-site,indel frameshift,and missense annotated as deleterious by at least one of five protein prediction algorithms (LRT score,MutationTaster,PolyPhen-2HumDiv,PolyPhen-2HumVar and SIFT);‘deleterious (strict)’as defined by nonsense,splice-site,indel frameshift,and missense annotated as deleterious by all five of the above protein prediction algorithms;disruptive:carriers of mutations that are nonsense,indel frameshift,or splice-site.

RESEARCH LETTER

2|N A T U R E |V O L 000|00M O N T H 2014

?2014

and missense annotated as damaging by at least one of five protein prediction algorithms;and(2)‘deleterious(strict)’as defined by non-sense,splice-site,indel frameshift,and missense annotated as damaging by all five protein prediction algorithms(see Methods).Carriers of a rare APOA5deleterious(strict)mutation had an even higher risk for MI/CAD(3.3-fold,P50.008).

A burden of rare mutations in APOA5explains about0.14%of the total variance for MI and roughly0.28%of the heritability(assuming that additive genetic factors explain,50%of the overall variance)(see Methods and Supplementary Table11).When compared with non-carriers,carriersofrare non-synonymous APOA5alleleshad higher plasma triglycerides(median in carriers was167mg dl21versus104mg dl21 for non-carriers,P50.007)and lower high-density lipoprotein cho-lesterol(mean in carriers was43mg dl21versus57mg dl21for non-carriers,P50.007),but similar LDL cholesterol(median in carriers was110mg dl21versus108mg dl21for non-carriers,P50.66)(Sup-plementary Table12).

In the fourth follow-up approach,we performed exome sequencing in additional early-onset MI/CAD cases and controls,bringing the total number of exomes analysed to9,793(Supplementary Tables13 and14).We tested for an excess(or deficit)in cases versus controls of rare mutations in any gene(Supplementary Fig.28and Supplementary Tables15–17).At this sample size,rare alleles collectively conferred risk for MI at exome-wide significance in only one gene,LDLR(Fig.3). After sequencing the exons of LDLR in4,703cases and5,090con-trols,we identified156unique non-synonymous,splice-site SNVs and indel frameshifts with allele frequency,1%(Table2and Supplemen-tary Table18).Of these variants,we observed285alleles in cases(6.1% of cases)and208alleles in controls(4.1%of controls)(1.5-fold effect size,P5431026)(Table2).When restricting analysis to the delete-rious(PolyPhen)set,3.1%of cases and1.3%of controls carried at least one such rare mutation,for a2.4-fold effect size(P5*******).A higher effect size of4.2-fold(P5*******)was observed when restrict-ing to the deleterious(strict)set.When restricting to disruptive alleles, 0.51%of cases and0.04%of controls carried at least one such rare disruptive mutation,for a13-fold effect size(P5931025)(Table2 and Fig.3).

Among controls,approximately1in217individuals carried an LDLR non-synonymous or disruptive mutation and had LDL cholesterol .190mg dl21;in contrast,among cases,approximately1in51indi-viduals carried an LDLR non-synonymous or disruptive mutation and had LDL cholesterol.190mg dl21.

A burden of rare mutations in LDLR explains about0.24%of the total variance for MI and roughly0.48%of the heritability(see Methods and Supplementary Table19).LDL cholesterol level differed based on functional class annotation with the greatest difference seen between carriers of disruptive mutations and those who did not carry any non-synonymous mutations(279mg dl21versus135mg dl21,Fig.3and Supplementary Table20).Approximately49%of the LDLR alleles dis-covered in this study(77of156)have been previously observed in LDLR familial hypercholesterolemia databases23(Supplementary Table21). Using these rare variant signals as a guide,we estimated sample sizes that will be required to make similar discoveries.A very large number of samples,at least10,000exomes,are required to achieve80%statistical power at an exome-wide level of statistical significance(Supplementary Figs29–31).

Here we show that a burden of rare alleles in two genes,LDLR and APOA5,contributes to risk for MI.These results suggest several conclu-sions regarding the inherited basis for MI and rare variant association studies.First,after a DNA sequence-based search across nearly all protein-coding genes in.9,700early-onset MI cases and controls,LDLR is the strongest association signal,with mutations in the gene accounting for about2%of cases.In1973,Goldstein and colleagues studied survivors of early MI and noted two common lipid abnormalities:hypercholes-terolemia and hypertriglyceridemia16.On the basis of a total cholesterol value exceeding,285mg dl21,it was estimated that4.1%of cases with MI prior to the age of60had familial hypercholesterolemia;this ori-ginal estimate is similar to ours based on direct sequencing.In contrast, the prevalence of harmful LDLR mutations in the general population is higher than the original estimate(,0.5in the present study versus0.1–0.2%by Goldstein).Second,the rare variant association signal presented here establishes APOA5as a bona fide MI gene.Initially discovered through comparative genomics analysis of a region harbouring several lipid regulators(that is,APOA1and APOC3),the APOA5locus har-bours common variants associated with plasma triglycerides24.Candi-date gene and genome-wide association studies have associated common variants at this locus also with MI risk(that is,21131T.C,APOA5 promoter region,rs662799,MAF of8%)25,26.However,because of ex-tensive linkage disequilibrium in this region,it had been previously uncertain which gene is responsible for the association with MI.The identification of multiple coding sequence variants within APOA5clar-ifies that this gene contributes to MI risk in the population.Third,these data point to a route to MI beyond LDL cholesterol,namely triglyceride-rich lipoproteins27and the lipoprotein lipase pathway.Genetic variation at two other proteins related to APOA5function,apolipoprotein C-III (refs18,19,28)and lipoprotein lipase15,17,has been associated with tri-glycerides and MI risk.Finally,the present study makes clear that rare variant discovery for complex disease will require the sequencing of thousands of cases and careful statistical analysis.Two reasons for the large sample size requirement are an inability to readily distinguish harmful from benign alleles and the extreme rarity of harmful alleles. Online Content Methods,along with any additional Extended Data display items and Source Data,are available in the online version of the paper;references unique to these sections appear only in the online paper.

Received9January;accepted3October2014.

Published online10December2014.

1.Marenberg,M.E.,Risch,N.,Berkman,L.F.,Floderus,B.&de Faire,U.Genetic

susceptibility to death from coronary heart disease in a study of twins.N.Engl.

J.Med.330,1041–1046(1994).

2.Lloyd-Jones,D.M.et al.Parental cardiovascular disease as a risk factor for

cardiovascular disease in middle-aged adults:a prospective study of parents and offspring.J.Am.Med.Assoc.291,2204–2211(2004).

3.Lehrman,M.A.et al.Mutation in LDL receptor:Alu–Alu recombination deletes

exons encoding transmembrane and cytoplasmic domains.Science227,

140–146(1985).

4.Brown,M.S.&Goldstein,J.L.A receptor-mediated pathway for cholesterol

homeostasis.Science232,34–47(1986).

5.Soria,L.F.et al.Association between a specific apolipoprotein B mutation and familial

defective apolipoprotein B-100.Proc.Natl https://www.wendangku.net/doc/3c17519564.html,A86,587–591(1989).

Table2|Association of a burden of rare mutations in LDLR with risk for early-onset myocardial infarction or coronary artery disease Mutation set n cases/controls T1cases T1controls Freq cases(%)Freq controls(%)OR P

Non-synonymous4,703/5,090285208 6.1 4.1 1.5431026 Deleterious(PolyPhen)4,703/5,09014867 3.1 1.3 2.41310211 Deleterious(broad)4,703/5,090243158 5.2 3.1 1.7931028 Deleterious(strict)4,703/5,0909023 1.90.45 4.23310211 Disruptive4,703/5,0902420.510.03913.0931025 Summary allele counts and carrier frequencies are shown.Only SNVs and indels with minor allele frequency less than1%were considered in burden analysis.Deleterious(PolyPhen)as defined by nonsense, splice-site,indel frameshift,and missense annotated as‘possibly damaging’or‘probably damaging’by PolyPhen-2HumDiv software;‘deleterious(broad)’as defined by nonsense,splice-site,indel frameshift, and missense annotated as deleterious by at least one of the five protein prediction algorithms of LRT score,MutationTaster,PolyPhen-2HumDiv,PolyPhen-2HumVar and SIFT;‘deleterious(strict)’as defined by nonsense,splice-site,indel frameshift,and missense annotated as deleterious by all five protein prediction algorithms;Disruptive defined as nonsense,splice-site or indel frameshift;T1:alleles from SNVs or indels with minor allele frequency less than1%;Freq(%):percentage of cases or controls carrying a T1allele;OR:odds ratio.

LETTER RESEARCH

00M O N T H2014|V O L000|N A T U R E|3

?2014

6.Garcia,C.K.et al.Autosomal recessive hypercholesterolemia caused by

mutations in a putative LDL receptor adaptor protein.Science292,1394–1398 (2001).

7.Berge,K.E.et al.Accumulation of dietary cholesterol in sitosterolemia caused by

mutations in adjacent ABC transporters.Science290,1771–1775(2000).

8.Abifadel,M.et al.Mutations in PCSK9cause autosomal dominant

hypercholesterolemia.Nature Genet.34,154–156(2003).

9.McPherson,R.et al.A common allele on chromosome9associated with coronary

heart disease.Science316,1488–1491(2007).

10.Samani,N.J.et al.Genomewide association analysis of coronary artery disease.

N.Engl.J.Med.357,443–453(2007).

11.Helgadottir,A.et al.A common variant on chromosome9p21affects the risk of

myocardial infarction.Science316,1491–1493(2007).

12.Kathiresan,S.et al.Genome-wide association of early-onset myocardial infarction

with single nucleotide polymorphisms and copy number variants.Nature Genet.

41,334–341(2009).

13.Schunkert,H.et https://www.wendangku.net/doc/3c17519564.html,rge-scale association analysis identifies13new susceptibility

loci for coronary artery disease.Nature Genet.43,333–338(2011).

14.Coronary Artery Disease(C4D)Genetics Consortium.A genome-wide association

study in Europeans and South Asians identifies five new loci for coronary artery disease.Nature Genet.43,339–344(2011).

15.The CARDIoGRAMplusC4D Consortium et https://www.wendangku.net/doc/3c17519564.html,rge-scale association analysis

identifies new risk loci for coronary artery disease.Nature Genet.45,25–33(2013).

16.Goldstein,J.L.,Schrott,H.G.,Hazzard,W.R.,Bierman,E.L.&Motulsky,A.G.

Hyperlipidemia in coronary heart disease.II.Genetic analysis of lipid levels in176 families and delineation of a new inherited disorder,combined hyperlipidemia.

J.Clin.Invest.52,1544–1568(1973).

17.Varbo,A.et al.Remnant cholesterol as a causal risk factor for ischemic heart

disease.J.Am.Coll.Cardiol.61,427–436(2013).

18.The TG and HDL Working Group of the Exome Sequencing Project,National Heart,

Lung and Blood Institute et al.Loss-of-function mutations in APOC3,triglycerides, and coronary disease.N.Engl.J.Med.371,22–31(2014).

19.J?rgensen,A.B.,Frikke-Schmidt,R.,Nordestgaard,B.G.&Tybjaerg-Hansen,A.

Loss-of-function mutations in APOC3and risk of ischemic vascular disease.

N.Engl.J.Med.371,32–41(2014).

20.Gnirke,A.et al.Solution hybrid selection with ultra-long oligonucleotides for

massively parallel targeted sequencing.Nature Biotechnol.27,182–189(2009).

21.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).

22.Purcell,S.M.et al.A polygenic burden of rare disruptive mutations in

schizophrenia.Nature506,185–190(2014).

23.Leigh,S.E.,Foster,A.H.,Whittall,R.A.,Hubbart,C.S.&Humphries,S.E.Update and

analysis of the University College London low density lipoprotein receptor familial hypercholesterolemia database.Ann.Hum.Genet.72,485–498(2008).

24.Pennacchio,L.A.et al.An apolipoprotein influencing triglycerides in humans and

mice revealed by comparative sequencing.Science294,169–173(2001).

25.Triglyceride Coronary Disease Genetics Consortium and Emerging Risk Factors

Collaboration et al.Triglyceride-mediated pathways and coronary disease:

collaborative analysis https://www.wendangku.net/doc/3c17519564.html,ncet375,1634–1639(2010).

26.Teslovich,T.M.et al.Biological,clinical and population relevance of95loci for

blood lipids.Nature466,707–713(2010).

27.Do,R.et https://www.wendangku.net/doc/3c17519564.html,mon variants associated with plasma triglycerides and risk for

coronary artery disease.Nature Genet.45,1345–1352(2013).

28.Pollin,T.I.et al.A null mutation in human APOC3confers a favorable plasma lipid

profile and apparent cardioprotection.Science322,1702–1705(2008). Supplementary Information is available in the online version of the paper. Acknowledgements The authors wish to acknowledge the support of the National Heart,Lung,and Blood Institute(NHLBI)and the National Human Genome Research Institute(NHGRI)of the US National Institutes of Health(NIH)and the contributions of the research institutions,study investigators,field staff and study participants in creating this resource for biomedical research.Funding for the exome sequencing project(ESP)was provided by NHLBI grants RC2HL-103010(HeartGO),RC2

HL-102923(LungGO)and RC2HL-102924(WHISP).Exome sequencing was performed through NHLBI grants RC2HL-102925(BroadGO)and RC2HL-102926 (SeattleGO).Exome sequencing in the ATVB,PROCARDIS,and Ottawa studies was supported by NHGRI5U54HG003067-11to E.S.L.and S.G.Cleveland Clinic GeneBank was supported by NIH grants P01HL076491and P01HL098055.S.K.is supported by a Research Scholar award from the Massachusetts General Hospital(MGH),the Howard Goodman Fellowship from MGH,the Donovan Family Foundation,

R01HL107816,and a grant from Fondation Leducq.R.D.is supported by a Banting Fellowship from the Canadian Institutes of Health Research.N.O.S.is supported,in part, by a career development award from the NIH/NHLBI K08HL114642and by The Foundation for Barnes-Jewish Hospital.N.O.S.was supported by award number

T32HL007604from the NHLBI.G.M.P was supported by award number

T32HL007208from the NHLBI.The content is solely the responsibility of the authors and does not necessarily represent the official views of the NHLBI,NHGRI,or NIH. The Italian ATVB Study was supported by a grant from RFPS-2007-3-644382.A full listing of acknowledgements is provided in the Supplementary Information.

Author Contributions R.Do,N.O.S.,H.-H.W.,A.B.J.,and A.K.carried out the primary data analyses.R.Do,N.O.S.,L.A.L.,G.M.P.,P.L.A.,J.E.R.,B.M.P.,D.M.H.,J.G.W.,S.S.R.,M.J.B., R.P.T.,L.A.C.,S.L.H.,H.A.,J.A.S.,S.C.,C.S.C.,C.K.,R.D.J.,E.B.,G.R.A.,S.M.S.,D.S.S.,D.A.N., S.R.S.,C.J.O.,D.Altshuler,S.G.,and S.K.contributed to the design and conduct of the discovery exome sequencing study.S.G.,D.N.F.,and M.A.D.enabled the exome sequencing,variant calling,and annotation.R.Do,N.O.S.,H.-H.W.,A.B.J.,S.D.,P.A.M.,M.F.,A.G.,I.G.,R.A.,D.G.,N.M.,O.O.,R.R.,A.F.R.S.,D.S.,J.D.,S.E.E.,S.S.,G.K.H.,J.J.K.,N.J.S., H.S.,J.E.,S.H.S.,W.E.K.,C.T.J.,R.A.H.,O.Z,E.H.,W.M.,M.N.,J.W.,A.H.,R.C.,D.F.R,W.Y., M.E.K.,J.H.,A.D.J.,M.L.,G.L.B.,M.G.,Y.L.,T.L.A.,G.H.,E.M.L.,A.R.F.,H.A.T.,M.A.R.,P.D., D.J.R.,M.P.R.,J.H.,W.W.H.T.,A.P.R.,D.Ardissino,D.Altshuler,R.M.,A.T.-H.,H.W.,and S.K. contributed to the design and conduct of the imputation-based validation, genotyping-based validation,and/or the re-sequencing based validation study.S.S. supervised the analysis of exome sequencing data and power analysis.R.Do,N.O.S., H.-H.W.,S.D.,P.A.M.,M.F.,A.G.,R.A.,E.S.L.,R.M.,H.W.,D.Ardissino,S.G.,and S.K. contributed to the design and conduct of the replication exome sequencing study.S.G., E.S.L.,S.K.,D.A.N.,and D.Altshuler obtained funding.D.Altshuler,D.A.N.,S.S.R.,R.D.J., and https://www.wendangku.net/doc/3c17519564.html,prised the executive committee of the NHLBI Exome Sequencing Project.C.J.O.and S.K.led the Early-Onset Myocardial Infarction study team within the NHLBI Exome Sequencing Project.R.Do,N.O.S.,H.-H.W.and S.K.wrote the manuscript. Author Information DNA sequences have been deposited with the NIH dbGAP repository under accession numbers phs000279and phs000814.Reprints and permissions information is available at https://www.wendangku.net/doc/3c17519564.html,/reprints.The authors declare no competing financial interests.Readers are welcome to comment on the online version of the paper.Correspondence and requests for materials should be addressed to S.K.(skathiresan@https://www.wendangku.net/doc/3c17519564.html,).

Ron Do1,2,3,4*,Nathan O.Stitziel5,6*,Hong-Hee Won1,2,3,4*,Anders Berg J?rgensen7, Stefano Duga8,Pier Angelica Merlini9,Adam Kiezun4,Martin Farrall10,Anuj Goel10,Or Zuk4,Illaria Guella8,Rosanna Asselta8,Leslie https://www.wendangku.net/doc/3c17519564.html,nge11,Gina M.Peloso1,2,3,4,Paul L. Auer12,NHLBI Exome Sequencing Project{,Domenico Girelli13,Nicola Martinelli13, Deborah N.Farlow4,Mark A.DePristo4,Robert Roberts14,Alexander F.R.Stewart14, Danish Saleheen15,John Danesh15,Stephen E.Epstein16,Suthesh Sivapalaratnam17, G.Kees Hovingh17,John J.Kastelein17,Nilesh J.Samani18,Heribert Schunkert19, Jeanette Erdmann20,Svati H.Shah21,22,William E.Kraus22,Robert Davies23,Majid Nikpay23,Christopher T.Johansen24,Jian Wang24,Robert A.Hegele24,25,Eliana Hechter4,Winfried Marz26,27,28,Marcus E.Kleber26,Jie Huang29,Andrew D. Johnson30,Mingyao Li31,Greg L.Burke32,Myron Gross33,Yongmei Liu34, Themistocles L.Assimes35,Gerardo Heiss36,Ethan https://www.wendangku.net/doc/3c17519564.html,nge11,37,Aaron R.Folsom38, Herman A.Taylor39,Oliviero Olivieri13,Anders Hamsten40,Robert Clarke41,Dermot F. Reilly42,Wu Yin42,Manuel A.Rivas43,Peter Donnelly43,44,Jacques E.Rossouw45, Bruce M.Psaty46,47,David M.Herrington48,James G.Wilson49,Stephen S.Rich50, Michael J.Bamshad51,52,53,Russell P.Tracy54,L.Adrienne Cupples55,Daniel J. Rader56,Muredach P.Reilly57,John A.Spertus58,Sharon Cresci5,59,Jaana Hartiala60, W.H.Wilson Tang61,Stanley L.Hazen61,Hooman Allayee60,Alex P.Reiner12,62, Christopher S.Carlson12,Charles Kooperberg12,Rebecca D.Jackson63,Eric Boerwinkle64,Eric https://www.wendangku.net/doc/3c17519564.html,nder4,Stephen M.Schwartz12,62,David S.Siscovick62,65,Ruth McPherson23,Anne Tybjaerg-Hansen7,66,Goncalo R.Abecasis67,Hugh Watkins10,43, Deborah A.Nickerson53,Diego Ardissino68,Shamil R.Sunyaev4,69,Christopher J. O’Donnell29,David Altshuler1,4,Stacey Gabriel4&Sekar Kathiresan1,2,3,4

1Center for Human Genetic Research,Massachusetts General Hospital,Boston,

Massachusetts02114,USA.2Cardiovascular Research Center,Massachusetts General Hospital,Boston,Massachusetts02114,USA3Department of Medicine,Harvard Medical School,Boston,Massachusetts02114,USA.4Program in Medical and Population Genetics,Broad Institute,7Cambridge Center,Cambridge,Massachusetts02142,USA. 5Cardiovascular Division,Department of Medicine,Washington University School of

Medicine,St Louis,Missouri63110,USA.6Division of Statistical Genomics,Washington University School of Medicine,St Louis,Missouri63110,USA.7Department of Clinical Biochemistry KB3011,Section for Molecular Genetics,Rigshospitalet,Copenhagen University Hospitals and Faculty of Health Sciences,University of Copenhagen, Copenhagen1165,Denmark.8Dipartimento di Biotecnologie Mediche e Medicina Traslazionale,Universita`degli Studi di Milano,Milano20122,Italy.9Division of Cardiology,Ospedale Niguarda,Milano20162,Italy.10Department of Cardiovascular Medicine,The Wellcome Trust Centre for Human Genetics,University of Oxford,Oxford OX12J,UK.11Department of Genetics,University of North Carolina,Chapel Hill,North Carolina27599,USA.12Public Health Sciences Division,Fred Hutchinson Cancer Research Center,Seattle,Washington98109,USA.13University of Verona School of Medicine,Department of Medicine,Verona37129,Italy.14John&Jennifer Ruddy Canadian Cardiovascular Genetics Centre,University of Ottawa Heart Institute,Ottawa, Ontario K1Y4W7,Canada.15Department of Public Health and Primary Care,University of Cambridge,Cambridge CB21TN,UK.16MedStar Health Research Institute, Cardiovascular Research Institute,Hyattsville,Maryland20782,USA.17Department of Vascular Medicine,Academic Medical Center,Amsterdam1105AZ,The Netherlands. 18Department of Cardiovascular Sciences,University of Leicester,and Leicester NIHR Biomedical Research Unit in Cardiovascular Disease,Glenfield Hospital,Leicester LE3 9QP,UK.19DZHK(German Research Centre for Cardiovascular Research),Munich Heart Alliance,Deutsches Herzzentrum Mu¨nchen,Technische Universita¨t Mu¨nchen,Berlin 13347,Germany.20Medizinische Klinik II,University of Lu¨beck,Lu¨beck23562,Germany. 21Center for Human Genetics,Duke University,Durham,North Carolina27708,USA. 22Department of Cardiology and Center for Genomic Medicine,Duke University School of Medicine,Durham,North Carolina27708,USA.23Division of Cardiology,University of Ottawa Heart Institute,Ottawa,Ontario K1Y4W7,Canada.24Department of Biochemistry, Schulich School of Medicine and Dentistry,Robarts Research Institute,University of Western Ontario,London,Ontario N6A3K7,Canada.25Department of Medicine,Schulich School of Medicine and Dentistry,Robarts Research Institute,University of Western Ontario,London,Ontario N6A3K7,Canada.26Medical Faculty Mannheim,Mannheim Institute of Public Health,Social and Preventive Medicine,Heidelberg University,Ludolf Krehl Strasse7-11,Mannheim D-68167,Germany.27Clinical Institute of Medical and Chemical Laboratory Diagnostics,Medical University of Graz,Graz8036,Austria.

RESEARCH LETTER

4|N A T U R E|V O L000|00M O N T H2014

?2014

28Synlab Academy,Mannheim68259,Germany.29The National Heart,Lung,Blood Institute’s Framingham Heart Study,Framingham,Massachusetts01702,USA.

30National Heart,Lung,and Blood Institute Center for Population Studies,The Framingham Heart Study,Framingham,Massachusetts01702,USA.31Department of Biostatistics and Epidemiology,School of Medicine,University of Pennsylvania, Philadelphia,Pennsylvania19104,USA.32Department of Epidemiology,University of Alabama–Birmingham,Birmingham,Alabama35233,USA.33Department of Laboratory Medicine and Pathology,School of Medicine,University of Minnesota,Minneapolis, Minnesota55455,USA.34School of Medicine,Wake Forest University,Winston–Salem, North Carolina27106,USA.35Department of Medicine,Stanford University School of Medicine,Stanford,California94305,USA.36Department of Epidemiology,University of North Carolina,Chapel Hill,North Carolina27599,USA.37Carolina Center for Genome Sciences,University of North Carolina,Chapel Hill,North Carolina27599,USA.38Division of Epidemiology and Community Health,University of Minnesota School of Public Health, Minneapolis,Minnesota55455,USA.39University of Mississippi Medical Center,Jackson, Mississippi39216,USA.40Atherosclerosis Research Unit,Department of Medicine,and Center for Molecular Medicine,Karolinska Institutet,Stockholm17177,Sweden.

41Clinical Trial Service Unit and Epidemiological Studies Unit,University of Oxford,Oxford OX12JD,UK.42Merck Sharp&Dohme Corporation,Rahway,New Jersey08889,USA. 43The Wellcome Trust Centre for Human Genetics,University of Oxford,Oxford OX12JD, UK.44Department of Statistics,University of Oxford,Oxford OX12JD,UK.45National Heart,Lung,and Blood Institute,Bethesda,Maryland20824,USA.46Cardiovascular Health Research Unit,Departments of Medicine,Epidemiology,and Health Services, University of Washington,Seattle,Washington98195,USA.47Group Health Research Institute,Group Health Cooperative,Seattle,Washington98101,USA.48Section on Cardiology,and Public Health Sciences,Wake Forest School of Medicine,Winston–Salem, North Carolina27106,USA.49Jackson Heart Study,University of Mississippi Medical Center,Jackson State University,Jackson,Mississippi39217,USA.50Center for Public Health Genomics,University of Virginia,Charlottesville,Virginia22904,USA.51Division of Genetic Medicine,Department of Pediatrics,University of Washington,Seattle, Washington98195,USA.52Seattle Children’s Hospital,Seattle,Washington98105,USA. 53Department of Genome Sciences,University of Washington,Seattle,Washington

98195,USA.54Department of Biochemistry,University of Vermont,Burlington,Vermont 05405,USA.55Department of Biostatistics,Boston University School of Public Health, Boston,Massachusetts02118,USA.56Perelman School of Medicine,University of Pennsylvania,Philadelphia,Pennsylvania19104,USA.57Cardiovascular Institute, Perelman School of Medicine at the University of Pennsylvania,Philadelphia, Pennsylvania19104,USA.58St Luke’s Mid America Heart Institute,University of Missouri-Kansas City,Kansas City,Missouri64111,USA.59Department of Genetics, Washington University in St Louis,Missouri63130,USA.60Department of Preventive Medicine and Institute for Genetic Medicine,University of Southern California Keck School of Medicine,Los Angeles,California90033,USA.61Cardiovascular Medicine,Cleveland Clinic,Cleveland,Ohio44195,USA.62Department of Epidemiology,University of Washington,Seattle,Washington98195,USA.63Ohio State University,Columbus,Ohio 43210,USA.64Human Genetics Center,The University of Texas Health Science Center at Houston,Houston,Texas77030,USA.65Department of Medicine,School of Medicine, University of Washington,Seattle,Washington98195,USA.66Faculty of Health and Medical Sciences,University of Copenhagen,Blegdamsvej3B,2200K?benhavn N, Denmark.67Center for Statistical Genetics,Department of Biostatistics,University of Michigan,Ann Arbor,Missouri48109,USA.68Department of Cardiology,Parma Hospital, Parma43100,Italy.69Division of Genetics,Brigham and Women’s Hospital,Harvard Medical School,Boston,Massachusetts02115,USA.

*These authors contributed equally to this work.

{A list of authors and their affiliations appears in the Supplementary Information.

LETTER RESEARCH

00M O N T H2014|V O L000|N A T U R E|5

METHODS

General overview of the Exome Sequencing Project(ESP).Details of the study design of the National Heart,Lung and Blood Institute’s GO exome sequencing project(NHLBI ESP)have been published previously29.Briefly,the goal of the NHLBI ESP was to discover rare coding variation in genes contributing to heart, lung and blood disorders using next-generation sequencing of the protein-coding regions of the genome(exome sequencing).The study includes five primary groups including:Seattle GO(University of Washington,Seattle,Washington);Broad GO (Broad Institute,Cambridge,Massachusetts);WHISP(Ohio State University Medical Center,Columbus,Ohio);Lung GO(University of Washington,Seattle, Washington);Heart GO(University of Virginia Health System,Charlottesville, Virginia)and two collaborating groups,WashU GO(Washington University,St Louis)and CHARGE-S GO(University of Texas Health Sciences Center,Houston, Texas).

We included samples from several studies:Women’s Health Initiative(WHI); Framingham Heart Study(FHS);Jackson Heart Study(JHS);Multi-Ethnic Study of Atherosclerosis(MESA);Atherosclerosis Risk in Communities(ARIC);Coro-nary Artery Risk development in Adults(CARDIA);Cardiovascular Health Study (CHS);Lung Health Study(LHS);COPD genetic epidemiology(COPD Gene); severe asthma research project(SARP);pulmonary arterial hypertension(PAH); acute lung injury(ALI);cystic fibrosis(CF);Cleveland Clinic GeneBank(CCGB); Massachusetts General Hospital premature coronary artery disease study(MGH PCAD);Heart Attack Risk in Puget Sound(HARPS);Translational Research In-vestigating Underlying Disparities in Acute Myocardial Infarction Patients’Health Status(TRIUMPH)and the PennCath study.

General overview of the ESP early-onset myocardial infarction study.Within the NHLBI ESP,we designed an exome sequencing experiment specifically to study early-onset myocardial infarction(EOMI).We selected EOMI cases and controls from eleven studies,including:ARIC,MESA,CCGB,FHS,HARPS,MGH PCAD, PennCath,TRIUMPH,WHI,CHS,and JHS(Supplementary Tables1–3).Samples were selected based on the extreme tails of the phenotypic distribution,in order to enrich for a genetic contribution to disease.EOMI cases were defined as individuals who had an MI at an age of#50for men and#60for women.Controls were selected as individuals with no history of MI at baseline or during follow-up to at least age60 for men and70for women.The study samples,along with case and control defini-tions,are briefly described below and shown in Supplementary Tables1–3. Study and phenotype descriptions for ESP EOMI

The HeartGO consortium.HeartGO is a multiethnic consortium consisting of six NHLBI population-based cohorts of men and women:ARIC,CHS,FHS,CARDIA, JHS,and MESA.The age range of participants in these six cohorts spans the spec-trum from early adulthood to old age,providing a broad age representation.Each participating cohort in HeartGO has completed ascertainment of multiple pheno-types,including all of the major cardiovascular risk factors(blood pressure,lipids, diabetes status),biomarkers including measures of blood cell counts,subclinical disease imaging,and cardiovascular and lung outcomes including MI and stroke. Participants in all six cohorts provided written informed consent.The NIH data-base of genotypes and phenotypes(dbGaP)site contains further details regarding the phenotypes accessible for each individual HeartGO cohort.

Cleveland Clinic GeneBank(CCGB).The CCGB study is a single-centre prospec-tive cohort-based study that enrolled patients undergoing elective diagnostic cor-onary angiography between2001and2006.

Heart Attack Risk in Puget Sound(HARPS).The HARPS study is a population-based case-control study that enrolled cases with incident MI presenting to a net-work of hospitals in the metropolitan Seattle–Puget Sound region of Washington State between1998and2002.

The Massachusetts General Hospital premature coronary artery disease(MGH PCAD)study.The MGH PCAD study is a hospital-based case-control study that enrolled cases hospitalized with early MI at MGH between1999and2004. PennCath.The PennCath study is a catheterization-lab based cohort study from the University of Pennsylvania Medical Center and enrolled subjects at the time of cardiac catheterization and coronary angiography between1998and2003.Persons undergoing cardiac catheterization at either the Hospital of the University of Pennsylvania or Penn Presbyterian Medical Center consented for the PennCath study to identify genetic and biochemical factors related to coronary disease. The Translational Research Investigating Underlying Disparities in Acute Myo-cardial Infarction Patients’Health Status(TRIUMPH).The TRIUMPH study is a large,prospective,observational cohort study of consecutive patients with acute MI presenting to24US hospitals from April2005to December2008.MI was diagnosed using contemporary definitions30and all patients had an elevated tro-ponin blood test.

Women’s Health Initiative(WHI).WHI is a major research program that has been ongoing for over20years to address the most common causes of death,disability and poor quality of life in postmenopausal women—cardiovascular disease,cancer, and osteoporosis.

Studies involved in follow-up statistical imputation,array-based genotyping, targeted re-sequencing and additional exome sequencing

Statistical imputation.We performed statistical imputation of single nucleotide variants(SNVs)discovered in the exomes of the first786samples.We imputed exonic SNVs into64,132independent samples in16studies to test for association of coding SNVs with MI or CAD.The studies are described in Supplementary Table5. Array-based genotyping.We performed follow-up array-based genotyping using the Illumina HumanExome Beadchip(‘exome chip’)array in15,936independent samples from seven studies.The studies are described in Supplementary Table7. Targeted re-sequencing.We performed targeted re-sequencing of the APOA5gene in an additional11,414individuals from five cohorts.The studies are described in Supplementary Table9.

Exome sequencing-based follow-up.We performed exome sequencing in addi-tional individuals from three cohorts.The studies are described in Supplementary Table13.

Detailed methods for the processing and analysis of samples for the various stages of the project are described below.We describe methods for the different stages of the project,including discovery exome sequencing,follow-up imputation,array-based genotyping,targeted re-sequencing and additional exome sequencing. Laboratory methods for discovery exome sequencing in the ESP EOMI Project. Exome sequencing.Exome sequencing was performed at the Broad Institute.Sequenc-ing and exome capture methods have been previously described29.A brief descrip-tion of the methods is provided below.

Receipt/quality control of sample DNA.Samples were shipped to the Biological Samples Platform laboratory at the Broad Institute of MIT and Harvard.DNA con-centration was determined by the Picogreen assay(Invitrogen)before storage in 2D-arcoded0.75ml Matrix tubes at220u C in the SmaRTStore(RTS,Manchester, UK)automated sample handling system.We performed initial quality control(QC) on all samples involving sample quantification(PicoGreen),confirmation of high-molecular weight DNA and fingerprint genotyping and gender determination (Illumina iSelect).Samples were excluded if the total mass,concentration,integ-rity of DNA or quality of preliminary genotyping data was too low.

Library construction and in-solution hybrid selection.Starting with3m g of geno-mic DNA,library construction and in-solution hybrid selection were performed as described previously31.A subset of samples,however,was prepared using this protocol with some slight modifications.Initial genomic DNA input into shearing was reduced from3m g to100ng in50m l of solution.In addition,for adaptor liga-tion,Illumina paired-end adapters were replaced with palindromic forked adap-ters with unique8base index sequences embedded within the adaptor. Preparation of libraries for cluster amplification and sequencing.After in-solution hybrid selection,libraries were quantified using qPCR(KAPA Biosystems) with probes specific to the ends of the adapters.This assay was automated using Agilent’s Bravo liquid handling platform.Based on qPCR quantification,libraries were normalized to2nM and then denatured using0.1N NaOH using Perkin-Elmer’s MultiProbe liquid handling platform.A subset of the samples prepared using forked,indexed adapters was quantified using qPCR,normalized to2nM using Perkin-Elmer’s Mini-Janus liquid handling platform,and pooled by equal volume using the Agilent Bravo.Pools were then denatured using0.1N NaOH. Denatured samples were diluted into strip tubes using the Perkin-Elmer MultiProbe. Cluster amplification and sequencing.Cluster amplification of denatured tem-plates was performed according to the manufacturer’s protocol(Illumina)using either Genome Analyzer v3,Genome Analyzer v4,or HiSeq2,000v2cluster chemistry and flowcells.After cluster amplification,SYBR green dye was added to all flowcell lanes,and a portion of each lane visualized using a light microscope,in order to confirm target cluster density.Flowcells were sequenced either on Genome Analyzer II using v3and v4Sequencing-by-Synthesis Kits,then analysed using RTA v1.7.48, or on HiSeq2,000using HiSeq2,000v2Sequencing-by-Synthesis Kits,then ana-lysed using RTA v1.10.15.All samples were run on76cycle,paired end runs.For samples prepared using forked,indexed adapters,Illumina’s Multiplexing Sequenc-ing Primer Kit was also used.

Read mapping and variant analysis.Samples were processed from real-time base-calls(RTA1.7software[Bustard],converted to qseq.txt files,and aligned to a human reference(hg19)using Burrows–Wheeler Aligner(BWA,see ref.32).Aligned reads duplicating the start position of another read were flagged as duplicates and not analysed(‘duplicate removal’).Data was processed using the Genome Analysis ToolKit(GATK v1.1.3,ref.33).Reads were locally realigned(GATK IndelRealigner) and their base qualities were recalibrated(GATK TableRecalibration).Variant detection and genotyping were performed on both exomes and flanking50base pairs of intronic sequence using the UnifiedGenotyper(UG)tool from the GATK. Variant data for each sample was formatted(variant call format(VCF))as‘raw’calls for all samples.SNVs and indel sites were flagged using the variant filtration

RESEARCH LETTER

?2014

walker(GATK)to mark sites of low quality that were likely false positives.SNVs were marked as potential errors if they exhibited strong strand bias(SB$0.10), low average quality(quality per depth of coverage(QD),5.0),or fell in a homo-polymer run(HRun.4).Indels were marked as potential errors for low quality (quality score(QUAL),30.0),low average quality(QD,2.0),or if the site exhib-ited strong strand bias(SB.21.0).Samples were considered complete when exome targeted read coverage was$203over$80%of the exome target. Data analysis QC.Fingerprint concordance between sequence data and finger-print genotypes was evaluated.Variant calls were evaluated on both bulk and per-sample properties:novel and known variant counts,transition–transversion(TS–TV) ratio,heterozygous–homozygous non-reference ratio,and deletion/insertion ratio. Both bulk and sample metrics were compared to historical values for exome sequenc-ing projects at the Broad Institute.No significant deviation of the ESP variants or ESP samples from historical values was noted.

Data processing,quality control and association analysis of discovery exome sequencing

Variant calling.Variants(SNVs and indels)were identified and genotyped from recalibrated BAM files34using the multi-sample processing mode of the Unified Genotyper tool from the GATK.Variants were first identified and genotyped in random batches of100samples.The batches were then merged into a single VCF file using the GATK CombineVariants tool.

Variant annotation.Variants(SNVsandindels)were annotatedusing theGRCh37.64 database using the SNP effect predictor tool(SnpEff,see ref.35)and the GATK VariantAnnotator.The primary SnpEff genomic effects that were annotated include: splice-site acceptor,splice-site donor,indel frameshift,indel non-frameshift,non-sense,non-synonymous and synonymous variants.For variants that have different annotations due to multiple transcripts of the gene,the highest impact effect for each variant was taken.

Sample level quality control.We performed several quality control steps to iden-tify and remove outlier samples(Supplementary Figs1–8).First,we required that each sample had a minimum of20-fold coverage for at least80%of the targeted bases. Second,we compared self-reported ancestry with that inferred from the sequence data and removed discordant samples.Third,we removed samples with high degree of heterozygosity and low number of singleton counts as this pattern suggests DNA contamination across samples.Fourth,we removed samples with an extremely high number of variants or singletons as this can suggest low quality DNA.Finally, we removed samples exhibiting a mismatch between the reported gender and that inferred from sequence data.Of2,066cases and controls sequenced across the exome, we removed93samples due to these exclusion criteria.

Variant level quality control.QC measures were also performed to remove low quality variants.We assessed population genetics metrics including the TS–TV ratio, the ratio of the number of heterozygous changes to the number of homozygous non-reference changes,and the number of non-synonymous to the number of synonym-ous changes.This analysis can help filter false positive calls since we expect the true TS–TV to be around,3.2in European populations33,while a set of random SNVs (or false positive variants)should give a random expectation of0.5.Variants with low depth of coverage(DP)and high percent missingness generally had low TS–TV and heterozygous–homozygous non-reference ratios.Variants were removed if there was DP,8average per sample and.2%missingness(Supplementary Figs 9–12).Distribution of allele frequencies of the SNVs is shown in Supplementary Fig.13.

Common variant association analysis.We performed single variant association analysis in our exome sequencing data set.For SNVs with MAF greater than5%, we ran logistic regression,after adjusting for10principal components while for SNVs with MAF less than5%,we ran Fisher’s Exact test.We performed association analysis in European Americans and African Americans separately and then per-formed sample size weighted meta-analysis using METAL36.The association results are shown in Supplementary Fig.14.

Rare variant association analysis.To test whether rare mutations contribute to MI,we performed burden of rare variant analysis on the,2,000ESP EOMI exome samples.We performed a variant of the Combined Multivariate Collapsing test21, that groups the count of alleles of SNVs in cases and controls.Phenotype labels were permuted100,000times to assign a statistical significance.We accounted for ethnicity by permuting phenotype labels within each ethnicity.Association ana-lysis was performed using PLINK/SEQ.

We collapsed variants based on computational predictions from PolyPhen-2 HumDiv37.Minor allele frequencies were calculated from all available samples sequenced in each study in order to obtain the most accurate MAF estimates. Therefore,calculation of MAF for ESP EOMI1and2was performed on a larger set of exome samples that were sequenced at the Broad Institute as part of ESP (n5970exomes for ESP EOMI1and n53,014for ESP EOMI2).For our burden of rare variant association analysis,we use a MAF threshold of1%(T1).Further-more,we use three different types of variant groupings when collapsing by gene.These variant groups are:(1)non-synonymous only;(2)a deleterious set consisting of non-synonymous after excluding missense alleles annotated as benign by PolyPhen-2 HumDiv software;and(3)disruptive(nonsense,indel frameshift,splice-site)muta-tions only.We also performed the T1test after collapsing all non-synonymous mutations by KEGG pathways(Supplementary Figs21and22).

Methods for follow-up statistical imputation

Construction of reference panels and targeted imputation panels.Exome impu-tations were performed using two reference panels and16targeted imputation panels.

A total of697ESP samples(436African Americans and261European Americans) were used for the first reference panel while89samples from the1000Genomes Project38were drawn for the second reference panel.For the ESP reference panel, all samples from ARIC(n5212),JHS(n5119),MGH PCAD or HARPS(n5151) and WHI studies(n541)were genotyped using commercially available Affymetrix 6.0arrays.Samples from the FHS(n5174)were genotyped using the Affymetrix 5.0array.The second reference panel was comprised of samples from the1000 Genomes Project that had genotype data for both low coverage sequencing and high coverage exome sequencing data38.A total of89samples were selected from6 diverse populations(23African Ancestry in Southwest US(ASW),9Utah residents with Northern and Western European ancestry(CEU),12Colombian in Medellin, Colombia(CLM),25Mexican Ancestry in Los Angeles,CA(MXL),17Toscani in Italia(TSI)and3Yoruba in Ibadan,Nigeria(YRI)samples).Low-coverage whole genome sequencing,high-coverage exome sequencing and targeted exome capture were performed based on standard protocols at the Broad Institute.Details of the sequencing methods and samples have been described previously38.Imputation was performed into16independent study samples with genome-wide genotype data. Study samples were genotyped using commercially available Affymetrix or Illumina genotyping arrays.Further details are described in Supplementary Table5. Reference panels were created by merging genotypes from SNVs that span the entire genome(hence,providing a haplotype‘scaffold’),with genotypes from SNVs from ESP exome sequencing data.The first reference panel was generated using genotypes from both genome-wide SNV arrays obtained from dbGAP and exome sequencing data.The second reference panel was generated using genotype data for both low coverage sequencing and high coverage exome sequencing data.Both the reference panel and targeted genome-wide panel were phased using the‘best guess haplotypes’option in IMPUTE2(ref.39).Haplotype phasing were performed in5megabase chunks as recommended by the software tutorial39.

Data processing,quality control and association analysis.Imputation of the exome was performed using IMPUTE2.We imputed approximately400,000cod-ing SNVs from the reference panels into28,068cases and36,064controls from16 different study samples with genome-wide data.Descriptions for the study samples have been reported elsewhere(Supplementary Table5for references).We filtered SNVs with MAF,1%and imputation quality(INFO),0.5from further analysis. The distribution of imputation qualities of the SNVs is shown in Supplementary Figs23and24.Association testing for CAD/MI was performed using the score method and assuming an additive model in SNPTEST40.Age,sex and the first two principal components were used as covariates when appropriate.We did not observe any indication of excess inflation of test statistics in any of the study samples (Supplementary Table22).Meta-analysis of study-specific P values for imputed SNVs was performed using the Z-score method weighted by sample size in METAL. Beta and standard errors were estimated based on an inverse-weighted meta-analysis. The distribution of association results for the imputation results is shown in Sup-plementary Fig.25and top association results in Supplementary Table6. Methods for follow-up array-based genotyping

Laboratory methods.DNA samples were sent to the Broad Institute Genetic Ana-lysis Platform for genotyping and were placed on96-well plates for processing using the Illumina HumanExome v1.0SNP array.Genotypes were assigned using GenomeStudio v2010.3using the calling algorithm/genotyping module version 1.8.4along with the custom cluster file StanCtrExChp_CEPH.egt.Only samples passing an overall call rate of98%criteria and standard identity check were released from the genetic analysis platform.

Data processing,quality control and association analysis.To identify single low-frequency SNVs associated with MI or CAD,we performed array-based genotyping using the Illumina Human Exome Beadchip.We genotyped83,680sites identified from exome sequencing in1,027early-onset MI cases and946controls.The sam-ples for genotyping were drawn from the cohorts listed in Supplementary Table7 and have been previously described.The functional effect of each variant was predicted using the SeattleSeq Annotation server.For variants having more than one functional class,the most deleterious class was retained.

Several quality control processes were employed to ensure high quality geno-types and samples were used in the association analysis.Samples were excluded for the following criteria:greater than5%missing genotypes;discordance between inferred gender based on genotype and self-reported gender;inbreeding coefficient less than20.2or greater than0.2;duplicated samples;or proportion of genotypes

LETTER RESEARCH

identical by descent.0.2.In addition,principal components were calculated using Eigenstrat4.2(ref.41)and samples were removed if they were found to be statistical population outliers.Variants were removed for the following criteria:MAF50%; significant difference between missingness in cases compared with controls;extreme deviation from Hardy–Weinberg equilibrium(P,131026);or significant asso-ciation with genotyping plate assignment.All quality control filtering were per-formed using PLINK42and R(The R Project for Statistical Computing,Vienna, Austria).

Association testing for CAD/MI was performed within each study separately using logistic regression with ten principal components of ancestry as covariates.An inverse standard-error weighted meta-analysis was performed to combine results across studies.The association testing was performed using PLINK42and the meta-analysis was performed using METAL.There was no indication of an inflation of test statistics across studies(Supplementary Table23).The stability of logistic regres-sion was assessed by examining the standard error of the beta estimate as a function of minor allele frequency(see Supplementary Fig.32).As shown,logistic regression is unstable for a MAF,0.05%.Fisher’s Exact test was used for variants with MAF ,0.05%.The top association results are shown in Supplementary Table8. Methods for follow-up re-sequencing

Selection of genes.We first selected six associated genes(based on biologic and/or statistical evidence with T1P,0.005;APOA5,CHRM5,SMG7,LYRM1,APOC3, NBEAL1)for replication sequencing in the ATVB study(Supplementary Table24) where all cases had suffered an MI before age of46.We also pursued the same six genes in the Ottawa Heart Study with552cases and586controls(Supplementary Table25).One of the genes(APOA5)continued to show significant results and was sequenced in three additional studies(Table1and Supplementary Table26). In total,we performed follow-up sequencing of APOA5in six study samples,includ-ing the Verona heart study(VHS),Ottawa heart study(OHS),additional exomes from atherosclerosis,thrombosis,and vascular biology Italian study group(ATVB), additional exomes from the ESP EOMI study(ESP EOMI2),Precocious Coronary Artery Disease Study(PROCARDIS),and the Copenhagen City Heart Study and Copenhagen Ischaemic Heart Disease Study(CCHS/CIHDS).

Laboratory methods.For the VHS study,genomic DNA was extracted from white blood cells using the salting-out method.The protein-coding regions correspond-ing to the RefSeq transcripts NM_052968for APOA5and NM_012125for CHRM5 were sequenced using in-house designed primers(available on request)and the BigDye Terminator Cycle Sequencing Kit v1.1on an ABI-3130XL Genetic Analyzer (Applied Biosystems,Foster City,CA).SNVs were called using the Variant Reporter software v1.1(Applied Biosystems).

For the OHS study,PCR primers were designed,tested and optimized to target the exons and flanking non-coding sequences for each gene.Sequencing reactions were performed using big dye terminator chemistry and chromatograms obtained with an Applied Biosystems ABI3730XL capillary sequencer.Chromatograms were base-called by using Phred,assembled into contigs by using Phrap,and scanned for SNVs with PolyPhred43to identify polymorphic sites.Each read was trimmed to remove low-quality sequence(Phred score,25),resulting in analysed reads with an average Phred quality of40.After assembly and variant calling,each polymorphic site was reviewed by a data analyst using Consed44to ensure the quality and accuracy of the variant calls.This process generates sequence-based SNV genotypes with accuracy.99.9%.

For the PROCARDIS study,a single long range PCR product(LRPCR)was amplified to provide coverage of the APOA5exonic,intronic and flanking sequences (human reference sequence NCBI build37chromosome11:116,659,905–116,664,331). The LRPCR products were tagged with unique sequence(barcode)adaptors,and processed into56short amplicons(Reflex reactions,http://www.populationgenetics. com)and pooled for multiplex next-generation sequencing(NGS).NGS was per-formed on a MiSeq personal sequencer to.203coverage across95%of the APOA5 target region on1,385MI cases and1,499controls.Paired-end reads were mapped to NCBI build37using the BWA and SMALT aligners;variants were identified by the GATK unified genotyper(v1.6.13)and annotated using SnpEff v2.0.5and the GRCh37.64database.

For the CCHS/CIHDS study,lightscanner screening and re-sequencing were performed.Genomic DNA was isolated from frozen whole blood(QiaAmp4DNA blood mini kit;QIAGEN,Hilden,Germany).Six PCR fragments were amplified covering the three coding exons and adjacent splice-sites(approximately20base pairs upstream and downstream each exon)of APOA5.Mutational analysis of the PCR products was performed by high-resolution melting curve(HRM)analysis using the Lightscanner system(Idaho Technology,Salt Lake City,Utah).PCR fragments showing heteroduplex formation by HRM analysis were subsequently sequenced on an ABI3730DNA analyser(Applied Biosystems,Foster City,CA). Data processing,quality control and association analysis.After sequencing, variants were annotated using SnpEff or Annovar45.For each study,only non-synonymous SNVs with MAF,1%were analysed.Rare variant burden testing was performed using the T1test.Meta-analysis was performed to combine evidence across study specific P values using the sample size weighted Z-score method, implemented in METAL.Association results and a listing of APOA5mutations discovered from sequencing are described in Table1and Supplementary Table10. P values for association between APOA5mutation carrier status and lipid traits were performed using the Mann–Whitney ranksum test.Results are shown in Sup-plementary Table12.

Methods for follow-up exome sequencing

Laboratory methods.We performed follow-up exome sequencing in additional samples from three other studies.Sequencing was performed at the Broad Institute, using the same protocols described above for the NHLBI ESP Project.

Data processing,quality control and association analysis.Variant calling and annotations were performed as described above for the NHBLI ESP EOMI.Quality control of samples was performed using the following steps.To detect mismatched samples,we calculated discordance rates between genotypes from exome sequenc-ing with genotypes from array-based genotyping.We removed samples with dis-cordance rate.0.02.We tested for sample contamination using verifyBamID46, which examines the proportion of non-reference bases at reference sites.We removed samples with FREEMIX or CHIPMIX scores.0.2.Furthermore,we removed out-lier samples with too many or too few SNVs(.700or,5singletons,.400or,5 doubletons,.16,000(.20,000for African)or,10,000total SNVs),and those with too high or low TS–TV(.4or,3)and heterozygosity(heterozygote to homo-zygote non-reference ratio.6or,2).Finally,we removed samples with high missingness(.0.1).In total,202samples were removed.For quality control of variants,we removed SNVs and indels that had low recalibration scores after run-ning GATK VariantRecalibrator.We also removed SNVs with low coverage(DP ,140,000and quality over depth(QD),2)and high missingness(frequency of missing genotypes.0.02).For quality control of indels,we removed indels that had excessive strand bias(Fisher Strand.200),high proportion of alternate alleles seen near the ends of reads(ReadPosRankSum,220),deviation from Hardy–Weinberg equilibrium(InbreedingCoeff,20.8)and low coverage(QD,3).Rare variant association analysis was performed using EPACTS.We performed burden of rare variant analysis using the Efficient Mixed-Model Association eXpedited (EMMAX)Combined Multivariate and Collapsing(CMC)test47.This approach uses a kinship matrix to take into account population structure.We restricted ana-lyses to SNVs and indels with minor allele frequency,0.01.Furthermore,we re-stricted analyses to three different sets of variants:(1)non-synonymous only;(2)a deleterious set consisting of non-synonymous after excluding missense alleles anno-tated as benign by PolyPhen-2HumDiv software;and(3)disruptive(nonsense, indel frameshift,splice-site)mutations only.

Estimation of heritability explained by a burden of rare mutations in the APOA5and LDLR genes.We calculated the heritability explained by a burden of rare mutations in the APOA5and LDLR genes using the following assumptions. We assumed that the alleles come from a mixture of two distributions:harmless alleles,with no effect on the trait,and null alleles,which destroy the function of the gene and have an(constant)effect on the trait.We assumed different values for the fraction of null alleles,a(our current expectation for most genes for a is around one-third to one-half for missense alleles and here we clump missense alleles together with nonsense alleles,which should slightly increase a).The variance explained is sensitive to this parameter.We assumed a liability-threshold model for disease, with an underlying(un-observed)continuous trait representing risk for MI,and MI occurring if risk is above a certain threshold.We assume all null alleles have effect b(in units of standard deviations)on the liability scale.We assumed different values for the prevalence(denoted k)for early MI(3%to5%).Results are some-what sensitive to prevalence;higher prevalence will slightly increase heritability estimates.Given the prevalence,the number of carriers in cases and controls gives us the allele frequency in the population(which is very close to the allele frequency in controls).

We fitted the effect size(b on liability scale)and alleles for different values of a and k.Results for APOA5are shown in Supplementary Table11and results for LDLR are shown in Supplementary Table19.For APOA5,b is moderate(up to roughly one standard deviation),with variance explained between0.08%and 0.17%of the total phenotypic variance(on the liability scale).If we assume the heritability of MI is50%,a burden of rare mutations in the APOA5gene may explain0.16–0.34%of the heritability.For LDLR,for all values,variance explained is between0.13%and0.32%of the total phenotypic variance(on the liability scale) and0.26–0.64%of the heritability.

Sample size extrapolations and power calculations for burden of rare variants. We evaluated the sample size that is needed to reach genome-wide significance levels(P52.531026)for the T1test.Our calculations relied on the following assumptions.We assumed that all allelic variants with population frequency less than1%are causal and have identical effect sizes.We also assumed that all alleles with frequency greater than1%were benign.

RESEARCH LETTER

?2014

Our calculations differentiate between the allele frequency of a SNV in our exome samples with its true allele frequency in a population.The T1test compares the number of carriers of an allele for a SNV with sample(rather than population) allele frequency less than1%among cases and controls.We considered three factors when extrapolating to larger sample sizes.First,we assumed our sample is comprised of50%cases and50%controls.As the prevalence of EOMI is esti-mated to be5%,the sample frequency of causal alleles is likely to be higher than the population frequency.Second,some alleles with population frequency below1% may,by chance,have sampling frequency greater than1%and therefore be excluded from the test.Third,the true allele frequency of the SNVs in the population is unknown.In contrast to earlier work that relied on population genetics modelling48, we provide an update on the power needed to detect rare variant signal after con-sidering the three factors above.We calculated liberal and conservative estimates for our sample size extrapolations and power calculations.The conservative esti-mate was based on the estimate of the total population frequency of all causal alleles(below1%)that would be unlikely to be excluded from the T1test due to the sampling frequencies exceeding1%.Because allele frequency distribution is domi-nated by rare alleles,for an allele with population frequency^x,expected popu-lation allele frequency is smaller than x.

(E x j^x

eTv^x)e1TTherefore,the expected total population frequency of all alleles below frequency x is smaller than the total sampling frequency of alleles below sampling frequency^x. However,setting^x at1%would result in a liberal rather than conservative estimate because alleles with population frequency below1%may be excluded from the T1 test as having sampling frequency above1%.This occurs due to oversampling cases (our sample has50%of cases at disease prevalence of5%)and sampling variance. For example,assuming only one causal allele per gene,the power of the T1test is maximal for the population allele frequency close to0.5%for a sample of1,000 cases and1,000controls.For a sample of10,000individuals,the chance that a risk allele with population frequency of0.5%would be excluded from the T1test is below1023,making this threshold even more conservative.Therefore,for a con-servative estimate,we have assumed that the total population frequency of all causal alleles per gene would equal the total sampling frequency of alleles below0.5%in the ESP sample.Our liberal estimate assumed that all causal alleles will be included in the T1test.We assumed that the total population frequency of all causal alleles per gene would equal the total sampling frequency of alleles below1%in the ESP sample.

Once we extrapolated the number of mutation carriers to20,000samples,we then performed power calculations to see how many samples would be needed to reach a genome-wide significance level for the T1test(P52.531026after cor-recting for20,000genes).Power calculations were performed by first sampling a genotype at random from the pool of20,000simulated samples.Based on the T1 carrier status of the drawn sample,we simulated the phenotype based on a calcu-lated probability.The phenotype was simulated based on a prevalence rate of5% for disease,carrier status of the random sample and assumed relative risk of2.0of the mutation.For T1carriers,the probability of being a case was calculated as rel-ative risk(RR)of T1carrier multiplied by prevalence rate of disease(RR*prevalence rate).For non-carriers,the probability of being a case was simply the prevalence rate.The case-control ratio was1:1.We performed sample size extrapolations for genes with varying number of T1mutations(25th percentile,median and75th per-centile of carriers with a T1mutation for all genes discovered in the exome,Sup-plementary Figs29–31).

29.Tennessen,J.A.et al.Evolution and functional impact of rare coding variation from

deep sequencing of human exomes.Science337,64–69(2012).

30.Antman,E.et al.Myocardial infarction redefined—a consensus document of The

Joint European Society of Cardiology/American College of Cardiology Committee for the redefinition of myocardial infarction.J.Am.Coll.Cardiol.36,959–969

(2000).

31.Fisher,S.et al.A scalable,fully automated process for construction of sequence-

ready human exome targeted capture libraries.Genome Biol.12,R1(2011). 32.Li,H.&Durbin,R.Fast and accurate short read alignment with Burrows–Wheeler

transform.Bioinformatics25,1754–1760(2009).

33.DePristo,M.A.et al.A framework for variation discovery and genotyping using

next-generation DNA sequencing data.Nature Genet.43,491–498(2011). 34.Li,H.et al.The Sequence Alignment/Map format and SAMtools.Bioinformatics25,

2078–2079(2009).

35.Cingolani,P.et al.A program for annotating and predicting the effects of single

nucleotide polymorphisms,SnpEff:SNPs in the genome of Drosophila

melanogaster strain w(1118);iso-2;iso-3.Fly6,80–92(2012).

36.Willer,C.J.,Li,Y.&Abecasis,G.R.METAL:fast and efficient meta-analysis of

genomewide association scans.Bioinformatics26,2190–2191(2010).

37.Sunyaev,S.et al.Prediction of deleterious human alleles.Hum.Mol.Genet.10,

591–597(2001).

38.1000Genomes Projects Consortium et al.A map of human genome variation from

population-scale sequencing.Nature467,1061–1073(2010).

39.Howie,B.N.,Donnelly,P.&Marchini,J.A flexible and accurate genotype

imputation method for the next generation of genome-wide association studies.

PLoS Genet.5,e1000529(2009).

40.Marchini,J.,Howie,B.,Myers,S.,McVean,G.&Donnelly,P.A new multipoint

method for genome-wide association studies by imputation of genotypes.Nature Genet.39,906–913(2007).

41.Price,A.L.et al.Principal components analysis corrects for stratification in

genome-wide association studies.Nature Genet.38,904–909(2006).

42.Purcell,S.et al.PLINK:a tool set for whole-genome association and population-

based linkage analyses.Am.J.Hum.Genet.81,559–575(2007).

43.Stephens,M.,Sloan,J.S.,Robertson,P.D.,Scheet,P.&Nickerson,D.A.Automating

sequence-based detection and genotyping of SNPs from diploid samples.Nature Genet.38,375–381(2006).

44.Gordon,D.,Abajian,C.&Green,P.Consed:a graphical tool for sequence finishing.

Genome Res.8,195–202(1998).

45.Wang,K.,Li,M.&Hakonarson,H.ANNOVAR:functional annotation of genetic

variants from high-throughput sequencing data.Nucleic Acids Res.38,e164

(2010).

46.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).

47.Kang,H.M.et al.Variance component model to account for sample structure in

genome-wide association studies.Nature Genet.42,348–354(2010).

48.Kryukov,G.V.,Shpunt,A.,Stamatoyannopoulos,J.A.&Sunyaev,S.R.Power of

deep,all-exon resequencing for discovery of human trait genes.Proc.Natl Acad.

https://www.wendangku.net/doc/3c17519564.html,A106,3871–3876(2009).

LETTER RESEARCH