Genetic Susceptibility to Lipid Levels and Lipid Change...
Transcript of Genetic Susceptibility to Lipid Levels and Lipid Change...
DOI: 10.1161/CIRCGENETICS.115.001096
1
Genetic Susceptibility to Lipid Levels and Lipid Change Over Time and Risk
of Incident Hyperlipidemia in Chinese Populations
Running title: Lu et al.; Genetics and incident hyperlipidemia
Xiangfeng Lu, PhD1; Jianfeng Huang, MD2; Zengnan Mo, MD3; Jiang He, PhD4; Laiyuan Wang,
PhD1,5; Xueli Yang, PhD1; Aihua Tan, MD3; Shufeng Chen, PhD1; Jing Chen, MD4; C. Charles
Gu, PhD6; Jichun Chen, MS;1 Ying Li, MD1; Liancheng Zhao, MD1; Hongfan Li, MS1; Yongchen
Hao, PhD1; Jianxin Li, MS1; James E. Hixson, PhD7; Yunzhi Li, PhD1; Min Cheng, MD1; Xiaoli
Liu, PhD8; Jie Cao, MS1; Fangcao Liu, PhD1; Chen Huang, PhD1; Chong Shen, PhD9; Jinjin Shen,
MD10; Ling Yu, MD11; Lihua Xu, MD12; Jianjun Mu, MD13; Xianping Wu, MD14; Xu Ji, MD15;
Dongshuang Guo, MD16; Zhengyuan Zhou, MD17; Zili Yang, MD18; Renping Wang, MD19; Jun
Yang, MD20; Weili Yan, PhD21; Xiaozhong Peng, PhD22; Dongfeng Gu, MD, PhD1
1State Key Laboratory of Cardiovascular Disease, 2Hypertension Division, Fuwai Hospital,
National Center for Cardiovascular Diseases, 22Institute of Basic Medical Sciences, Chinese
Academy of Medical Sciences & Peking Union Medical College, Beijing; 3Center for Genomic
& Personalized Medicine, Medical Scientific Research Center & Department of Occupational
Health & Environmental Health, School of Public Health, Guangxi Medical University, Nanning,
Guangxi, China; 4Department of Epidemiology, Tulane University School of Public Health &
Tropical Medicine, New Orleans, LA; 5National Human Genome Center at Beijing; 8Department
of Cardiology, Anzhen Hospital, Capital Medical University, Beijing, China; 6Division of
Biostatistics, Washington University School of Medicine, St. Louis, MO; 7Department of
Epidemiology, University of Texas School of Public Health, Houston, TX; 9Department of
Gu, PhD ; Jichun Chen, MS; Ying Li, MD ; Liancheng Zhao, MD ; Hongfan Li, MSM ; Yongche
Hao, PhD1; Jianxin Li, MS1; James E. Hixson, PhD7; Yunzhi Li, PhD1; Min Cheneneng,g,g, MDMDMD111;;; XiXiXiaoaoaolll
Liu,u,u, PPPhDhDhD888;;; JiJiJieee CaCaCaooo, MS1; Fangcao Liu, PhD1; Chhheene Huang, PhD1; Chonnng g g Shen, PhD9; Jinjin She
MMDMD10; Ling YYYuuu, MDMDMD1111; ; ; LiLiLihuhuhua a a XuXu, , , MDMDMD1222; Jianananjun MMMu, MDMM 131313;;; XiXianananpipingg WWWu, MDMDM 14414;;; XuXuXu JJJi,, MDMDMD1151 ;
DoDoDongngngshuaaangngng GGGuouoo, MDMDMD16; ZhZhhenenngygygyuauau n ZZZhohohou,u,u, MMMD1117; ; ZiZiilii YYYananang,g,g, MDMDMD18;;; ReReennpnpinnng g g WaWaW ngngng, MDMDMD19;; JuJuJunn
Yang, MD2000;; WeWeWeilililiii YaYaYan,n,n PPPhDhDhD212121; ; ; XiXiXiaoaoaozhzhzhononong g g PePePengngng, PhPhPhDDD222222;; DoDoDongngngfefefengngng GGGuuu, MD, PhD1
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
2
Epidemiology & Biostatistics, School of Public Health, Nanjing Medical University, Nanjing;10Yancheng Municipal Center for Disease Control & Prevention, Yancheng, Jiangsu;
11Department of Cardiology, Fujian Provincial People's Hospital, Fuzhou, Fujian; 12Department
of Cardiology, the Affiliated Hospital of Beihua University, Beihua University, Jilin; 13Department of Cardiology, First Affiliated Hospital of Medical College, Xi'an Jiaotong
University, Xi'an, Shannxi; 14Center for Chronic & Noncommunicable Disease Control &
Prevention, Sichuan Center of Disease Control & Prevention, Chengdu, Sichuan; 15Department
of Internal Medicine, Xinle Red Cross Hospital, Xinle, Hebei; 16Department of Internal Medicine,
Yuxian Renmin Hospital, Yuxian, Shanxi; 17Changshu Center of Disease Control & Prevention,
Changshu; 18Nantong Center of Disease Control & Prevention, Nantong, Jiangsu; 19Department
of Internal Medicine, Shijiazhuang Greatwall Hospital, Shijiazhuang, Hebei; 20Department of
Cardiology, Hanzhong People’s Hospital, Hanzhong, Shannxi; 21Department of Clinical
Epidemiology, Children’s Hospital of Fudan University, Shanghai, China
Correspondence:
Xiangfeng Lu, PhD Dongfeng Gu, MD, PhD
State Key Laboratory of State Key Laboratory of
Cardiovascular Disease, Fuwai Hospital, Cardiovascular Disease, Fuwai Hospital,
National Center for Cardiovascular Diseases, National Center for Cardiovascular Diseases,
Chinese Academy of Medical Sciences & Chinese Academy of Medical Sciences &
Peking Union Medical College, Peking Union Medical College,
Beijing, 100037, China Beijing, 100037, China
Tel: (8610) 60866599 Tel: (8610) 68331752
Fax: (8610) 88363812 Fax: (8610) 88363812
E-mail: [email protected] E-mail: [email protected]
Journal Subject Terms: Lipids and Cholesterol; Genetic, Association Studies; Genetics
gy, g p p , g, ; p
Epidemiology, Children’s Hospital of Fudan University, Shanghai, CChChinininayy
CCoC rrrrespondencee::
Xiiananangfgfgfeneneng g g LuLuLu, PPhPhDDD DoDoDongngngfefefengngng GGu,u,u, MMMD,D,D, PPPhDhDhD
State Key Laboratooryryry ooof f f StStStataa e e e KeKeKey y y LaLaLabbborororaatatororory y y ofofof
CaCardrdioiovavascsculularar DDisiseaeasese,,, FuFuwawaii HoHosspipip tatall,,, CaCardrdioiovavascsculularar DDisiseaeasese, , , FuFuwawai i HoHospsppititalal,,,
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
3
Abstract:
Background - Multiple genetic loci associated with lipid levels have been identified
predominantly in Europeans, and the issue of to what extent these genetic loci can predict blood
lipid levels increase over time and the incidence of future hyperlipidemia remains largely
unknown.
Methods and Results - We conducted a meta-analysis of genome-wide association studies of
lipid levels in 8,344 subjects followed by replication studies including 14,739 additional
individuals. We replicated seventeen previously reported loci. We also newly identified three
Chinese specific variants in previous regions (HLA-C, LIPG, and LDLR) with genome-wide
significance. Almost all the variants contributed to lipid levels change and incident
hyperlipidemia over 8.1-years follow-up among 6,428 individuals of a prospective cohort study.
The strongest associations for lipid levels change were detected at LPL, TRIB1,
APOA1-C3-A4-A5, LIPC, CETP, and LDLR (P range from 4.84 × 10-4 to 4.62 × 10-18), while
LPL, TRIB1, ABCA1, APOA1-C3-A4-A5, CETP, and APOE displayed significant strongest
associations for incident hyperlipidemia (P range from 1.20 × 10-3 to 4.67 × 10-16). The four
lipids genetic risk scores (GRS) were independently associated with linear increases in their
corresponding lipid levels and risk of incident hyperlipidemia. A C-statistics analysis showed
significant improvement in the prediction of incident hyperlipidemia on top of traditional risk
factors including the baseline lipid levels.
Conclusions - These findings identified some evidence for allelic heterogeneity in Chinese
compared with Europeans in relation to lipid associations. The individual variants and those
cumulative effects were independent risk factors for lipids increase and incident hyperlipidemia.
Key words: Genome Wide Association Study; lipids; incidence; incident hyperlipidemia, lipid level change
yp p y p g , p p y
The strongest associations for lipid levels change were detected at LPL, TRIB1,
APOA1-C3-A4- -A5- , LIPC, CETP, and LDLR (P( range from 4.84 × 10-4 to 4.62 × 100 1-18), while
LPPPLLL,,, TTRTRIBIBB111,,, ABABABCACACA1, APOA1-C3-A4- -A5- , CETP, aananddd APOE displaaayeyy d sisiigggnnificant strongest
aasassooociations fororr iiinncidddenee t t t hyhyhypepeperlrlrliipipididdemememiaiia (P( rraannnge fffroomom 111..20200 ××× 1000-3-- tto 4.44 66767 × 111000-1-1-1666).)) TTThehehe fofofoururur
iiipipp ddsds genetic rissk scorrresss (GRSRSRS)) wereree indndndepeendddentlylyly assssssooociaaateeed wwwithhh lineeearrr innncrcrreae sesss ininin tttheheeir
corresspopopondndndiing lilili iipiddd lelel vevellsls and rrisisiskkk of incncidididenenttt hypypererlililipipiiddede imimiaaa. A A CCC-ststatattisii tititicscs anaalylylysisis shhhowowededd
ignificant improvememmenenenttt ininin ttthehehe ppprereredididictctctioioion n n ofofof iiincncncididdenenentt hyhyhypepeerlrlrlipipipidididemememiaiaia ononon tttopopop oooff f traditional risk
fffa ttctors iiincllludididing ttthhhe bbbaselililine lilili iipiddd lllevellls.
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
4
Introduction
Plasma concentrations of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C),
high-density lipoprotein cholesterol (HDL-C) and triglycerides (TG) are the most important risk
factors for cardiovascular diseases and are targets for therapeutic intervention1-2. Genome-wide
association studies have successfully identified multiple genetic loci associated with blood
lipids3-7. However, almost all these loci were identified initially in European ancestry populations,
and few such genetic studies have assessed Asian populations8-10, especially Chinese populations.
Moreover, most of these studies mainly focused on cross-sectional data, and little data exist
regarding their genetic loci predicting lipid levels variation over time and hyperlipidemia
incidence. Lifestyle factors, including poor diet, obesity, and lack of exercise, are known to
increase the risk of hyperlipidemia. The differences in environmental exposures and genetic
background between Chinese and Europeans might suggest potential different pathways of blood
lipids. Some variants may be more common in specific ethnic groups, thereby providing greater
statistical power, or the effects of genetic variants on lipid levels may be enlarged in specific
ethnic groups11,12. Therefore, large scale studies in Chinese are needed not only to evaluate
whether the previous reported loci could be generalized to Chinese population but also to
identify new loci or Chinese specific variants for blood lipids.
Herein we conducted a genome-wide association study (GWAS) of blood lipid levels that
included a meta-analysis of GWAS from 8,344 samples at the discovery stage and additional
14,739 samples in an independent replication study, involving a total of 23,083 subjects from
Chinese Han ancestry. We subsequently investigated whether the identified variants would
contribute to lipids increase and incident hyperlipidemia in 6,428 individuals who participated in
a prospective cohort study.
ncidence. Lifestyle factors, including poor diet, obesity, and lack of exercise, areee knknknowowwn n n tototo
ncreasese tthe rrisi k ofof hyperlipidemia. The differencees s in environmental exxpopop sures and genetic
bababackkkground bbbetttweweweenn CCChihihinenenesess aaandndnd EEEurururopopopeaee nsnsns mmmighhht suggggegegeststst pppotenenentiiialalal ddifififfefeferentt pppatatathwhwhwayayysss ofofof bbblololoodo
iiipipp ddsds. Some variaiaiannts mamamay beee mmmore cococommmmmmon innn speeeciiificcc eeethtthnnniccc grouououpsp , thhherrrebbby y y prpp oviididiininggg gggreattterrr
tatistical power, or thththe ee effecttts s s ofoo ggenettic variants s s ono lippiddd llleveve ele s may yy bebebe enlarged in specific
etethnhnicic gggroroupuppss11,12.. ThThererefeforore,e,, llarargegeg sscacalele sstutudidieses iin n ChChininesese e arare e neneedededed nnotot oonlnly y y toto eevavaluluatate e
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
5
Material and Methods
Study population
The study comprised two-staged analyses carried out separately for TC, LDL-C, HDL-C, and TG.
A detailed description of the sample characteristics and phenotype measurements for each study
are provided in Table 1 and Supplementary material, Methods. The discovery stage included
meta-analysis of four independent GWAS in 8,344 individuals from Chinese populations: the
China atherosclerosis study (CAS)13, the Beijing atherosclerosis study (BAS)14, Genetic
Epidemiology Network of Salt-Sensitivity study (GenSalt)15, and the Guangxi Fangchenggang
Area Male Health and Examination Survey (FAMHES)16. In stage 2, replication analyses were
conducted in an independent sample with a total of 14,739 individuals from China
Cardiovascular Health Study (CCHS). CCHS was conducted to investigate the risk factors for
cardiovascular diseases in China since 2006. Overnight fasting blood samples were drawn by
venipuncture to measure lipid levels. Blood specimens were processed in the central clinical
laboratory at the Department of Population Genetics at Fuwai Hospital of the Chinese Academy
of Medical Sciences in Beijing. This laboratory participates in the Lipid Standardization Program
of the US Centers for Disease Control and Prevention and National Heart, Lung, and Blood
Institute.
A subset of CCHS individuals was from two prospective cohorts (ChinaMUCA17 and
InterAsia18) and used to investigate individual single nucleotide polymorphisms (SNPs) and
those cumulative effects on blood lipids increase and incident hyperlipidemia. The individuals
completed the baseline survey and examination in 1998, 2000 and 2001. For the present study,
the analyses were limited to participants for whom complete data were available for both
follow-up data and genetic risk score. These restrictions resulted in 6,428 individuals. After
conducted in an independent sample with a total of 14,739 individuals from Chinnana
Cardiovovasa culalar HeHealth Study (CCHS). CCHS was s coconducted to investigagatet the risk factors for
cacaardddiovascularaa dddisisiseaeasesesesss ininn CCChiinanana sssininincecee 2220000 6.6.6. OOOverrnniiightt fffasasastititingnn bbblloloododod ssamamamplpp ess wwwererere e e drdd awawawn n n bybyby
veveeniniipupp ncture to memm asuuureee lipiid dd lelel velss. .. BBBlooood spspspecimmmeeenss wwew reee ppproccessssed innn ttthee e ccecentralll ccclil nnnicccal
aboratory at the Deepapapartrtr ment ooof f f Poopulaatit on Genettticici ss at Fuwuwuwaiaa Hospip tataalll ofofof the Chinese Academy
ofof MMededicicalal SScicienencecess inin BBeieijijij ngngg.. ThThisis llababororatatorory y y papap rtrticicipippatateses iin n ththe e LiLipipip d d StStanandadardrdizizatatioion n PrProgoggraramm
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
6
exclusion of prevalent hyperlipidemia cases at baseline, 6,022, 5,814, 4,977, and 5,450
individuals for TC, LDL-C, HDL-C, and TG, respectively, were eligible for the present analysis
of lipids increase and incident hyperlipidemia.
The diagnosis criteria of hyperlipidemia were based on Detection, Evaluation, and
Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) report19:
dl), high levels of low-density lipoprotein cholesterol
(LDL- -density lipoprotein cholesterol (HDL-C < 40 mg/dl),
Each study obtained approval from the institutional review boards of local research
institutions. All participants in each study gave written informed consent.
Genotype imputation and genotyping of selected SNPs
Detailed descriptions of genotyping arrays and quality control filters applied to the four
discovery studies are provided in Supplementary Material, Table S1. To facilitate combining
results of genome-wide association scans based on the different genotyping platforms, we
imputed missing genotypes based on reference haplotypes from the phased CHB+JPT HapMap
data release 22 reference dataset using MACH20 or IMPUTE21. Only imputed SNPs with high
genotype information content were used for the association analysis. After
quality control, we obtained up to 2.5 million genotyped or imputed autosomal SNPs for
subsequent association analysis.
After genome-wide association analyses for each of the four discovery studies and
meta-analysis in the combined samples, SNPs representing the independent association signal
were taken forward to replication if they showed potential association (P < 5.0 × 10-5) for lipid
levels in the discovery meta-analysis. If a SNP could not be genotyped, alternative tagging SNPs
nstitutions. All participants in each study gave written informed consent.
Genotytyypepep impmpututata ion and genotyping of selecteded SNPs
DDDetatatailed descrcrc ipipiptititionoo s s s ofofof gggenenenottypypypininingg g ararrrararaysyy aaandndnd quaaaliiity cccononontrtrtroloo ffilililtttersrsrs aapppppplililied ttooo thththe e e fofof ururur
didiiscscscovoo ery studiess aaare ppprooovidededed in Suuupppp lllemmennntaaary MaMaMateeriririalalal, TaTaTable S11. To faaacilillitittataa e cooommmbiiiniining
esults of genome-wiwiwidedede assoccciaiaiatitit on scans based ooon nn tht e difffffferererenent geg nootytyypipipingg pplatforms, we
mmpupup teted d mimissssining g g gegeg nonotytyypepep ss babasesed d onon rrefefererenencece hhapapplolotytyypepep ss frfromom tthehe ppphahasesed d CHCHB+B JPJPT T HaHapMpMp apapp
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
7
were considered. In total, twenty-seven SNPs were selected and genotyped using iPLEX
Sequenom MassARRAY platform (Sequenom).
Statistical analysis
TG values were log transformed prior to analysis. Population stratification was estimated by a
principal component approach, as implemented by EIGENSTRAT software22. Within each of the
four discovery studies, continuous lipid levels were adjusted for age, age-squared, gender, body
mass index (BMI), the first two principal components, and lipid lowering drug prescriptions (if
applicable) in linear regression models under an additive model. We used allele dosages from
imputation to account for uncertainty in imputed genotypes. A fixed-effects inverse
variance-weighted meta-analysis implemented in METAL23 was used to combine the four studies
in the discovery stage and to obtain results for each SNP. A quantile-quantile plot generated using
R was used to evaluate the overall significance of the GWAS results and the potential impact of
2 statistic divided by 0.456. We detected the associations of SNPs in the replication populations,
and additionally, we carried out meta-analysis of the discovery and replications.
Genetic risk score (GRS)
We assessed the cumulative effect of the SNPs by using lipids GRS, which was a weighted sum
across the SNPs combining doses of the lipid increasing alleles (lipid decreasing alleles for
HDL-C) and the -coefficient) for lipid levels change over time in a prospective
cohort. Four scores were calculated individually for each lipid level (TC, LDL-C, HDL-C, and
TG). We included in the GRS only those SNPs significantly associated with the corresponding
lipid levels in our combined discovery and replication studies with genome-wide significance or
P < 5.0 × 10-5 for the SNPs located in known lipid genes. There was no linkage disequilibrium
variance-weighted meta-analysis implemented in METAL23 was used to combineee ttthehehe fffououour r r stststudududies
n the ddisiscoveeryy sstat ge and to obtain results for eacch h SNP. A quantile-quanantit le plot generated using
R R R wwwas used tto oo evevevalalaluaaatetete ttthehehe oooveeerararallllll sssigigignininififificaaancncnce ee of tthhhe GGGWAWAWASSS resususulttts s s ananandd d thththe popopotetetentntntiaiaial imimimpapapactctct ooof f
2 statistic divided bbby y y 0.0.0 456. WWWe ee ded tectede the assococociaations ofofof SNPN s in ttthehehe repplication populations,
anand d adaddidititiononalallylyy, , , wewe ccararririeded ooutut mmeteta-a ananalalysysy isis oof f ththe e didiscscovoverery y y anand d rereplplp icicatatioionsns. .
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
8
(LD) between SNPs included in the GRS. A 14-SNP GRS was constructed for TC, an 8-SNP
GRS for LDL-C, a 6-SNP GRS for HDL-C, and a 9-SNP GRS for TG. Missing genotype data
for each SNP were imputed using the average risk allele frequency. However, if more than two
SNP genotypes were missing for a given individual, the GRS was set as missing for that
individual. The GRS was modeled as a continuous variable (per one standard deviation increase)
and as quartiles of TC (7.85 to 19.79, 19.79 to 22.28, 22.28 to 24.77, and 24.77 to 35.40), LDL-C
(2.95 to 10.02, 10.02 to 11.60, 11.60 to 13.20, and 13.20 to 17.90), HDL-C (-13.38 to -9.18,
-9.18 to -8.25, -8.25 to -7.11, and -7.11 to -1.34), and TG (0.05 to 0.14, 0.14 to 0.16, 0.16 to 0.19,
and 0.19 to 0.26). Associations of the GRS with their corresponding lipid levels increase and
incident hyperlipidemia were analyzed using linear and logistic regression, respectively, in
models adjusting for sex, age, body mass index (BMI), and the baseline lipid levels
corresponding to GRS. We tested the null hypothesis of no linear trends effect over the quintiles.
We determined the effect estimates for quintiles of GRS with the bottom quintile serving as the
reference group. P values are reported for the linear trends across the quintiles. To evaluate the
improvement in risk discrimination by using the genetic information, we compared C-indices24
for the models with and without the GRS.
Results
The discovery meta-analysis and replication study
The discovery stage included four independent GWAS in 8,344 individuals from Chinese
populations (Table 1). The discovery meta-analysis evaluated separately associations for TC,
LDL-C, HDL-C, and TG, with 2,573,667 genotyped or imputed autosomal SNPs. All genotyped
and imputed autosomal SNPs passed quality control filters in each of the four datasets prior to
conducting the meta-analysis (Supplementary Material, Table S1). Quantile-quantile plots for
ncident hyperlipidemia were analyzed using linear and logistic regression, respeecectititivevev lylyly, , , ininin
modelss adjdjjusstit ngngg ffor sex, age, body mass index ((BMBMI), and the baseline llipipid levels
cocoorrrrespondinnnggg tototo GGGRSRSS. WeWeWe tttessteteted dd thththe e nununullll hhhypypypotoo hesis offf nnno oo lililinenn ararar trererendndn sss efefeffectt ooovevever r r ththt eee quququininintititileleles.s
WWWe dddetermined tthheee effffeccct esttimimimates ffforoo quuuintiiileees offf GRGRGRSS S wwwittth the booottommm qqquininntitiile serrrvivivingngg aaas thhhe
eference group. P vvvalalalueueu s aree rererepopop rted for the lineaeaear trendsss aaacrcrc oso s the e quququinini tiles. P To evaluate the
mmprprp ovovememenent t inin rrisisk k didiscscririmiminanatitionon bybyy uusisingngg tthehe gggeneneteticic iinfnforormamatitionon,,, wewe ccomompapap rered d C-C inindidicecess24
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
9
lipid levels are presented in Supplementary Material, Figure S1. The genomic control inflation
- 30~1.038). As shown in the Manhattan plots
of the -log10 P values for lipid levels in Supplementary Material, Figure S2, the meta-analysis
identified nine well-established loci (GCKR, HMGCR, LPL, ABO APOA1-C3-A4-A5, LIPC,
CETP, LIPG, and LDLR) at genome-wide significance (defined as P < 5.0 × 10-8). We then
selected 27 SNPs that were associated with TC, LDL-C, HDL-C, and/or TG at P < 5.0 × 10-5 in
the discovery analysis and genotyped them in an independent sample comprising 14,739 Chinese
individuals. In replication analysis, of these 27 SNPs, 18 SNPs showed significant association
with lipid after adjustment for multiple testing (P < 0.0018 = 0.05/27), including 11 SNPs with
genome wide significance). Additionally, two SNPs showed nominal significance (P < 0.05)
(Supplementary Material, Table S2).
Associations at previously reported loci and Chinese-specific variants
We evaluated the evidence of association for these SNPs by combined results of the discovery
and replication study. We confirmed 14 previously reported loci of TC, LDL-C, HDL-C, and/or
TG (PCSK9, ANGPTL3, GCKR, HMGCR, MLXIPL, LPL, TRIB1, ABCA1, ABO,
APOA1-C3-A4-A5, ALDH2, LIPC, CETP, and APOE) at the level of genome-wide significance
(P < 5.0 × 10-8). Moreover, suggestive associations were replicated for APOB, HPR and ABCA8
loci (Table 2). We also identified three novel ethnic-specific variants in previous reported
regions (HLA-C, LIPG and LDLR) in European populations at the level of genome-wide
significance. These SNPs were not in LD (r2 < 0.2 in JPT+CHB) with the previously reported
SNPs. These include SNPs in HLA-C (TC, P = 1.50 × 10-10; LDL-C, P = 3.28 × 10-10), LIPG
(HDL-C, P = 2.35 × 10-9), and LDLR (TC, P = 2.39 × 10-16; LDL-C, P = 1.44 × 10-24).
genome wide significance). Additionally, two SNPs showed nominal significancecee (((PPP( <<< 000.0.005)5)5) P
Supppplplplemememenenentatataryy MMMaterial, Table S2).
AAAsssosociations aaat t t prprprevvvioioiousususlylyly repepeporororteteteddd lololocicc aaandndnd Chhhinnnesesee-spspspecececifii iccc vvvararariaiaiantntntss
WWWe e eveve aluated d d theee eeevidddennnce ooof ff aasassociciiataa iooonnn forrr thhheseee SSSNPNPPss s bybyy cccombbibinnned reeessultttsss ooof theee dddisi ccocovev ryyy
and replication studddy.y.y. WWWee cococonfnfnfiririrmememed d d 141414 prprprevevevioioiousususlylyly rrrepepeporoo teteteddd loloocicici ofofof TCTCTC,,, LDLDLDLLL-CCC, HDL-C, and/or
TGTGTG (((PCPCPCSKSKSK999((( , ANANANGPGPGPTLTLTL333, GCGCGCKRKRKR, HMHMHMGCGCGCRRR, MLMLMLXIXIXIPLPLPL, LPLPLPLLL, TRTRTRIBIBIB111, ABABABCACACA111, ABABABOOO,
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
10
Single variants and lipid levels change and incident hyperlipidemia
Of these 20 SNPs in Chinese shown in Table 2, 17 were associated in at least two lipid
phenotypes, which resulted in a total of 37 associations of those SNPs with the corresponding
lipid levels change and incident hyperlipidemia (Supplementary Material, Table S3). For lipid
levels change, all the SNPs displayed a positive association in a direction consistent with their
effect on lipid. We compared the effect sizes for lipids change over time to those observed in
GWAS shown in Table 2 and found a high degree of correlation (r values from 0.83 to 0.92)
(Supplementary Material, Figure S3). Of these 37 associations, 24 (64.9%) showed nominal
significance (P < 0.05), while 10 (27.0%) displayed significance even at a Bonferroni-corrected
threshold (P range from 4.84 × 10-4 to 4.62 × 10-18 < 0.05/37) in LPL, TRIB1, APOA1-C3-A4-A5,
LIPC, CETP, and LDLR genes (Table 3). For incident hyperlipidemia, directions of effect were
consistent for all SNPs but the association of LIPG with hypercholesterolemia. 21 (56.8%)
displayed were significantly associated with the corresponding incident hyperlipidemia at
nominal significance (P < 0.05). ABCA1, APOA1-C3-A4-A5, and CETP with low levels of
high-density lipoprotein cholesterol, and LPL, TRIB1, APOA1-C3-A4-A5, and APOE with
hypertriglyceridemia displayed significant associations after accounting for multiple testing (P
range from 1.20 × 10-3 to 4.67 × 10-16 < 0.05/37).
Genetic risk score and the corresponding lipid levels change and incident hyperlipidemia
Over the mean follow-up period of 8.1 year, 404, 221, 476, and 807 individuals developed
incident hypercholesterolemia, high levels of low-density lipoprotein cholesterol, low levels of
high-density lipoprotein cholesterol, and hypertriglyceridemia, respectively. For each lipid
outcome (TC, LDL-C, HDL-C, and TG), the SNPs associated with the corresponding lipid levels
(Supplementary Material, Table S3) were used to calculate GRS. The GRS for four lipids were
hreshold (P( range from 4.84 × 10-4 to 4.62 × 10-18 < 0.05/37) in LPL, TRIB1, APPPOAOAOA111--C3C3C3--A4A4A4-- -AAA5-
LIPC, CECETPP,,, and d LDL LR genes (Table 3). For incidedent hyperlipidemia, didirer ctions of effect were
cocoonsnsnsistent forrr aallllll SSSNPNPPsss bububuttt thtt e e asasassososocicc atatatioioion ofofof LILL PGGG wwwithhh hhhypypyperee chhhooolesesestetet rororolelelemia.a.a. 222111 (5(5(56..888%)%)%)
didiispspsplalal yed were sssigggnifiicacaantly asssssociatetet ddd wwwith thhhe cooorrrrespopoponndinining innnciiided nt hhhyyypeererlililipidemimimia aat
nominal significancecee (((PPP((( < 0.050505).).). ABCA1, APOA11--C3C -A4- -A5A5A5- ,,, ana d CETPTPTP wiww th low levels of
hihighghg -ddenensisitytyy llipippopopproroteteinin cchoholeleststererolol,,, anand d LPLPLL,,, TRTRIBIB11,,, APAPOAOA11-C3C3-A4A4- -A5A5- ,,, anandd APAPOEOE wiwithth
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
11
independently associated with linear increases in their corresponding lipid levels over time (P
range from 6.55 × 10-11 to 1.42 × 10-28) and incident hyperlipidemia (P range from 5.33 × 10-4 to
5.33 × 10-18) (Table 4). For example, individuals in the highest quartile of risk score had an
increase of 9.930 mg/dl in TC levels, and 93.5% increased risk for incident hypercholesterolemia,
compared with those in the lowest quartile. An increase of one standard deviation of the GRS
was associated with increases of 3.724 mg/dl (P = 6.59 × 10-22) in TC, 2.252 mg/dl (P = 4.72 ×
10-12) in LDL-C, and 0.033 mg/dl (P = 5.71 × 10-36) in log-transformed TG, and a decrease of
1.699mg/dl (P = 2.06 × 10-28) in HDL-C. Figure 1 showed the distribution of GRS by their
corresponding incident hyperlipidemia. There was significant difference of GRS between
individuals with hyperlipidemia and without hyperlipidemia (P from 4.87 × 10-5 to 1.57 × 10-27).
Each standard deviation increase of the GRS resulted in 28.6%-48.2% increased risk for the
corresponding incident hyperlipidemia (P values range from 8.46 × 10-4 to 2.00 × 10-21).
Adding GRS to the models including traditional risk factors significantly improved risk
discrimination of incident hypercholesterolemia (C-index change = 1%; P = 0.035),
hypertriglyceridemia (C-index change = 2%; P = 2.0 × 10-4) and low levels of high-density
lipoprotein cholesterol (C-index change = 2%; P = 5.0 × 10-3) (Supplementary Material, Table
S4), while C-index analysis showed no improvement in the prediction of incident high levels of
low-density lipoprotein cholesterol (P = 0.122).
Discussion
The present study systematically investigated genetic susceptibility to lipid levels and its
relevance to lipid change over time and risk of incident hyperlipidemia in Chinese populations.
Our GWAS for lipids in Chinese replicated seventeen loci previously identified in populations of
European. We also identified Chinese specific variants at three loci (HLA-C, LIPG, and LDLR).
ndividuals with hyperlipidemia and without hyperlipidemia (P( from 4.87 × 10-5 tto o 11.1 575757 ××× 101010--2227).
Each statandn arrd d deeviv ation increase of the GRS resultltede in 28.6%-48.2% inncrc eased risk for the
cocoorrrrespondinnnggg ininincicicideentntnt hhhypypyperee liiipipipidededemimimiaaa (((aaa PPP((( vavavalululues rrannnge fffrororom m m 8.464646 ××× 10101 -4 tototo 2.0000 0 0 ××× 101010-21).).).
Adding g GRGRGRS tooo ttthe mmmodododels innnclclududdinggg trrraditttiooonaaall l riririskskk fffactooorsss signnnififfiiicananantltltly immprprrovovveddd risssk
discrimination of incicicideded nt hyhyypepepercrr holestterolemia (C(C(C-index chchchanangeg = 1%;%;%; PPP = 0.035),
hyhyypepep rtrtririglglg ycycy ererididememiaia (((CC-inindedex x chchanangegeg = 22%%;;; PP = 22.0.0 × 1100PP -444) ) ) anand d lolow w lelevevelsls oof f hihighghg -dedensnsitity y y
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
12
The study is the first investigation to address effect of variants on both lipid changes over time
and risk of incident hyperlipidemia in East Asians. In a sample of 6,418 Chinese participants
from a prospective cohort study, during more than 8.1 years of follow-up, we found that this
genetic information in Chinese could improve the prediction of incident hyperlipidemia beyond
baseline lipid levels.
Recent GWAS have identified 157 significantly lipid-associated loci in individuals of
European ancestry, and ancestry-specific analyses identified some associated SNPs clearly
distinct from the original GWAS variants3. As these studies were conducted almost exclusively in
populations of European descent, studies in non-European populations would allow us to assess
the relevance of the findings to other ethnic groups. In the present study, we found both shared
and population-specific lipid susceptibility was commonly present. We not only replicated 17
previously reported loci but also identified three Chinese specific variants (HLA-C, LIPG, and
LDLR). Of those, the association of LIPG with HDL-C has been found not to be generalized to
diverse populations including African, Indians, Mexican and Hispanics.12 We observed that the
reported lead SNPs at these three loci in Europeans were monomorphic or had low minor allele
frequency (MAF < 0.1) in the Chinese Han population, whereas these SNPs were quite
polymorphic in European populations (Supplementary Material, Table S5). For example, in
Europeans a prominent association was reported for rs6511720, which is not polymorphic in the
Chinese, whereas the strongest LDL-C association at LDLR (rs7258950) was detected in the
Chinese. These data suggest that genetic heterogeneity in these loci may be due to different LD
structure and minor allele frequency. As expected, the ethnic specific variants we detected in
Chinese were not in LD with the initially reported lead SNPs in Europeans. Conversely, we also
investigated whether the 3 Chinese specific variants identified in our samples were associated
he relevance of the findings to other ethnic groups. In the present study, we foundndnd bbbototth h h shshshararar deded
and popopupup lationo -spsppece ific lipid susceptibility was comommonly present. We nonot only replicated 17
prprprevvviously repepeporororteteted loloocicici bbbututut alslssooo idididenene tititifififiedee thththrereree CCChiiinessse e e spspspecececifficicic vvvararariaiaiantntntsss (HLLLAAA( -C-C-C,, LILL PGPGPG,, anananddd
LDLDDLLRLR). Of thoseee, ttthe aaasssociaaatitiioono of LILL PGPGPG wiiithhh HDDDLLL-CCC hahahas bebebeen fffoouound nnnot tooo bbbe gennnerereralllizzzede ttto
diverse populations inininclclc uding g g AfAfAfrican, Indians, MeMeMexican ananand d d Hisppaniciccs.s.s 12 We observed that the
eepopop rtrteded lleaead d SNSNPsPs aat t ththesese e ththreree e lolocici iin n EuEuroropepep anans s wewerere mmononomomororphphp icic oor r hahad d lolow w miminonor r alallelelele
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
13
with lipid in Europeans, using the results from the Global Lipids Genetics Consortium (GLGC),
a meta-analysis of 188,577 individuals.3 The rs12970066 at LIPG showed suggestive significant
association with HDL-C in the population of European ancestry (P = 4.77 × 10-7), while
rs9357121 at HLA-C and rs7258950 at LDLR showed no associations with lipid levels. These
differences in European and Chinese populations may facilitate the fine mapping of common
causal variants. We used HaploReg25 to search for evidence of the functional role for three
Chinese-specific variants. Rs9357121 (HLA-C) and rs12970066 (LIPG) were in strong LD with
rs1131151 (r2=0.91 in 1000 Genomes ASN) and rs2000813 (r2=0.81), respectively, which were
non-synonymous variants. The SNP rs7258950 in LDLR lies within the promoter and enhancer
histone marks as well as DNase hypersensitive sites. To gain further understanding of the lipid
susceptibility loci, we tested their associations with the traditional risk factors of cardiovascular
diseases in the replication samples. After Bonferroni correction for 20 independent tests, LPL
showed significant associations (P < 2.5 × 10-3 = 0.05/20) with diastolic blood pressure and
hypertension, while GCKR was significantly associated with plasma glucose (Supplementary
Material, Table S6).
We further assessed the individual effect of lipids-related SNPs on increases in lipids and
risk of incident hyperlipidemia. All the SNPs indicated a directionality-consistent association
with increases in lipids for lipid-raising alleles. These SNPs were associated with mild increases
in TC (range, 0.456~2.849 mg/dl per allele), LDL-C (range, 0.664~2.464 mg/dl per allele) and
log-transformed TG (range, 0.005~0.036 mg/dl per allele), and decrease in HDL-C (range,
0.460~1.959 mg/dl per allele). As expected, almost all of these SNPs were also observed to
increase the risk of incident hyperlipidemia. Although each SNP exerts a modest effect, a
combination of SNPs, in aggregate, can have a substantial influence on incident hyperlipidemia.
histone marks as well as DNase hypersensitive sites. To gain further understandinngng oof f f thththe e e lililipipipiddd
uscepepptitibib litytyy locci,i,, we tested their associations withth the traditional risk fafactc ors of cardiovascular
dididiseeeases in thhheee rerereplplp iccatatatioioion n n sasasampmpmpleleles.s.s. AAAftftftererer BBBonononfeff rroooniii coorrrrrrececectititionoo fffooror 222000 ininindededependndndenenent t t tetet sttts,s,s, LPLPLPLL L
hhhowwwed significaaanntt assssoccciatiooonsnsns (P( << 2..5 × 1000P -333 = 000.0005//202020))) wwwittth diiiassstolic bbbloododod pressssururure anannd
hypertension, while e GCGCGCKR wawawasss signg ificcantly y assooocicc ata ed wwwititith hh plp asma ggglululucocc se ((Supplementary
MaMateteririalal,,, TaTablble e S6S6).).)
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
14
Consistent with single variants finding, the variants in aggregate were significantly associated
with linear increases in lipids and risk of incident hyperlipidemia. These associations were
largely independent of lipid levels measured at baseline. By accumulation, for example,
individuals in the top compared to bottom quintiles of GRS differed by an increase of 9.930
mg/dl in TC and a 93.5% increased risk for incident hypercholesterolemia in a follow-up period
of 8.1 year. We also found GRS improved risk discrimination of incident hypercholesterolemia,
hypertriglyceridemia, and low levels of high-density lipoprotein cholesterol over the traditional
risk factors. Several European studies also showed the lipid-associated SNPs were associated
with the longitudinal lipid changes26-29. Tibor et al.26 further demonstrated that the ability of the
157 establish lipid-loci to predict incident dyslipidemia was modest in the Swedish populations
with 10-year follow-up. The C-index values were increased by 2-4%, which were slightly higher
than those (1-2%) reported by our study, given only 20 variants were involved in our GRS.
Although the variants have modest effects on lipids, their presence may act over the entire life
course and translate into comparatively large effects. It has been shown that joint effect of
LDL-C related SNPs was an independent risk factor for incident cardiovascular diseases in
Europeans30,31. However, the results might not be generalized to populations with genetic
backgrounds different from that of European populations. Thus, it is necessary to further evaluate
the predictive abilities of the genetic predisposition to higher lipids for risk of cardiovascular
diseases in Chinese population.
The major strength of our study includes the use of a population-based cohort for the
assessment of effect of variants on both lipid change over time and risk of incident
hyperlipidemia. The effect size estimated in published GWAS is the most frequently used to
create the GRS in published studies of genetic variants and risk prediction, but it may be
157 establish lipid-loci to predict incident dyslipidemia was modest in the Swedisissh h h popoopupupulalalatititioono s
with 10-0 yeyey ar ffollolow-w up. The C-index values were inincreased by 2-4%, whhici h were slightly higher
hhhannn those (1-2%2%%))) reepopoportrtrtededed bbby yy ououour r r stststududdy,y,y, givivivenenen onllly y 20 vvvararariaiaiantnn s wewewererere iinvnvnvoloo ved d d ininin oooururu GGGRSRSRS.. .
AlAlAlthhhouo gh the varrriaaants haaave mmmoododest efefe feeecttts onnn lllipiddds,, theheheiririr prrresssenccce mam y acacct oovoveeer thee eeennntiiiree e lifeee
course and translatee ininintotot commpapapararr tit velyy largeg effececectss. It hasasas bbbeeen showwwn n n ththt at jjoint effect of
LDLDLL-CC rerelalateted d SNSNPsPs wwasas aan n inindedepepep ndndenent t ririsksk ffacactotor r fforor iincncididenent tffff cacardrdioiovavascsculularar ddisiseaeasesess iin n
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
15
overestimated due to winner’s curse. In the present study, we constructed the GRS using
coefficient as determined in our prospective cohort to increase the accuracy of the GRS. Our
results should be interpreted in the context of potential limitations. Firstly, our discovery sample
size is modest relative to other previous consortia of European descent. This means that some
signals especially with a low minor allele frequency and/or weaker genetic effects may have
been missed. Secondly, although our genetic predisposition score captured the combined
information from the established genetic variants for lipids in Chinese thus far, it may account
for only a small proportion of lipid variation. For the 20 significantly associated variants in
Chinese, trait variance explained was 5.0% for TC, 4.7% for LDL-C, 6.9% for TG and 6.1% for
HDL-C, respectively, which was small when compared to 12%–15% explained by ~157 variants
discovered in European populations.3-4 The identification of the variants with small effect in
large-scale studies would be expected to capture additional genetic variance. Finally, our study
was undertaken in individuals of Chinese, and hence the results might not be generalized to
populations with genetic backgrounds different from that of our population.
In conclusion, we replicated seventeen loci previously identified in populations of
European for lipid levels. At HLA-C, LIPG, and LDLR, we also identified some evidence for
allelic heterogeneity in Chinese compared to Europeans. The individual SNPs and those
aggregate effects were independent risk factors for lipid increase and incident hyperlipidemia.
Funding Sources: This study was funded by the National Basic Research Program of China (973 Plan) (2011CB503901) and the High-Tech Research and Development Program of China (863 Plan) (2012AA02A516) from the Ministry of Science and Technology of China and the National Science Foundation of China (91439202, 81422043, 81370002). This study was also supported by Beijing Natural Science Foundation (7142138).
Conflict of Interest Disclosures: None.
HDL-C, respectively, which was small when compared to 12%–15% explained bybyby ~~~15155777 vavavariririants
discovere ede in n Euuroropean populations.3-4 The identifificac tion of the variantss wwith small effect in
aaargrggee-scale stttudududieieiesss wowooulululd d d bebebe expxpxpecececteteted dd tototo capapaptututurer adddddditiooonananal ll gegegeneeetititic vavavaririiananancecc . FiFiFinananalllllly,y,y, oooururur ssstututudydydy
wawawas s unu dertaken iiinn indidiivvviduallls ofo Chihiineneeseee, annnd hennnceee thhheee rressulllts mmmiggght nooot beee gggene eraaalizizzeddd ttto
populations with genenneteteticic backgkgkgroror unu ds ddifferent ffrororom that ooof f f ouour pop pupulalalatititionoo .
IIn n coconcnclulusisionon, , , wewe rereplplp icicatateded ssevevenenteteenen llococi i prprp evevioioususlylyy iidedentntififieied d inin pppopoppululatatioionsns oof f
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
16
References:
1. Castelli WP. Cholesterol and lipids in the risk of coronary artery disease--the Framingham Heart Study. Can J Cardiol. 1988;4 Suppl A:5A-10A.
2. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ, et al. Heart disease and stroke statistics--2014 update: a report from the American Heart Association. Circulation.2014;129:e28-e292.
3. Willer CJ, Schmidt EM, Sengupta S, Peloso GM, Gustafsson S, Kanoni S, et al. Discovery and refinement of loci associated with lipid levels. Nat Genet. 2013;45:1274-1283.
4. Teslovich TM, Musunuru K, Smith AV, Edmondson AC, Stylianou IM, Koseki M, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature.2010;466:707-713.
5. Willer CJ, Sanna S, Jackson AU, Scuteri A, Bonnycastle LL, Clarke R, et al. Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet.2008;40:161-169.
6. Kathiresan S, Melander O, Guiducci C, Surti A, Burtt NP, Rieder MJ, et al. Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet. 2008;40:189-197.
7. Kathiresan S, Willer CJ, Peloso GM, Demissie S, Musunuru K, Schadt EE, et al. Common variants at 30 loci contribute to polygenic dyslipidemia. Nat Genet. 2009;41:56-65.
8. Kamatani Y, Matsuda K, Okada Y, Kubo M, Hosono N, Daigo Y, et al. Genome-wide association study of hematological and biochemical traits in a Japanese population. Nat Genet.2010;42:210-215.
9. Kim YJ, Go MJ, Hu C, Hong CB, Kim YK, Lee JY, et al. Large-scale genome-wide association studies in East Asians identify new genetic loci influencing metabolic traits. Nat Genet. 2011;43:990-995.
10. Zhou L, He M, Mo Z, Wu C, Yang H, Yu D, et al. A genome wide association study identifies common variants associated with lipid levels in the Chinese population. PLoS One.2013;8:e82420.
11. Chang MH, Ned RM, Hong Y, Yesupriya A, Yang Q, Liu T, et al. Racial/ethnic variation in the association of lipid-related genetic variants with blood lipids in the US adult population. Circ Cardiovasc Genet. 2011;4:523-533.
12. Dumitrescu L, Carty CL, Taylor K, Schumacher FR, Hindorff LA, Ambite JL, et al. Genetic determinants of lipid traits in diverse populations from the population architecture using genomics and epidemiology (PAGE) study. PLoS Genet. 2011;7:e1002138.
oci that influence lipid concentrations and risk of coronary artery disease. Nat Genenenettt.2008;40:161-169.
6. Kathihireesan n S, MMelander O, Guiducci C, Surti A,, BuB rtt NP, Rieder MJ,,, eet al. Six new loci asssososociciciatateddd wwwititth blblbloood low-density lipoprotein chohoh lell sssterol, high-dennnsity lllipipipoprotein cholesterol orrrrigigiglllycerides ininin hhhumumu ananans.s.s. NaNaNat t GeGeGenenenettt. 20200080808;4440:0:0:1811 9-1919197.
7.7.. KKKata hiresan S, WWWilleerr CCCJ, PPPelloloso GGGMM,M, DDDemmmissssie S,S,S, MMMusususunurururu KKK, SSSchadddt EEEE, etee al. CComomommmmon variririananantststs aaatt t 303030 llocococi cococontntntriiibububute ttto o o popopolylylygenininic c c dydydyslsllipipipidididemmmiaiaia. . NaNaNat tt GeGeGeneeett. . 202020090909;4;4;41:11 565656-6-6-65.5.5.
8. Kamatani Y, Matttsususudadada KKK, OkOkOkadadadaa a Y,Y,Y, KKKubububooo M,M,M, HHHosososonononooo N,N,N, DDDaiaiaigogogo YYY,,, etetet aaal.ll. GGGenenenome-wide asassosociciatatioion n ststududy y y ofof hhemematatolologoggicicalal aandnd bbioiochchememicicalal ttraraititss inin aa JJapappananesese e popop pupup lalatitionon. . NaNat t GeGenenett..20201010;4;42:2:212100-212155
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
17
13. Lu X, Wang L, Chen S, He L, Yang X, Shi Y, et al. Genome-wide association study in Han Chinese identifies four new susceptibility loci for coronary artery disease. Nat Genet.2012;44:890-894.
14. Hou L, Chen S, Yu H, Lu X, Chen J, Wang L, et al. Associations of PLA2G7 gene polymorphisms with plasma lipoprotein-associated phospholipase A2 activity and coronary heart disease in a Chinese Han population: the Beijing atherosclerosis study. Hum Genet.2009;125:11-20.
15. The GenSalt Collaborative Research Group. GenSalt: rationale, design, methods and baseline characteristics of study participants. J Hum Hypertens. 2007;21:639-646.
16. Tan A, Sun J, Xia N, Qin X, Hu Y, Zhang S, et al. A genome-wide association and gene-environment interaction study for serum triglycerides levels in a healthy Chinese male population. Hum Mol Genet. 2012;21:1658-1664.
17. Wu Y, Liu X, Li X, Li Y, Zhao L, Chen Z, et al. Estimation of 10-year risk of fatal and nonfatal ischemic cardiovascular diseases in Chinese adults. Circulation. 2006;114:2217-2225.
18. Gu D, Reynolds K, Wu X, Chen J, Duan X, Reynolds RF, et al. Prevalence of the metabolic syndrome and overweight among adults in China. Lancet. 2005;365:1398-1405.
19. National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002;106:3143-3421.
20. Li Y, Willer CJ, Ding J, Scheet P, Abecasis GR. MaCH: using sequence and genotype data to estimate haplotypes and unobserved genotypes. Genet Epidemiol. 2010;34:816-834.
21. Marchini J, Howie B, Myers S, McVean G, Donnelly P. A new multipoint method for genome-wide association studies by imputation of genotypes. Nat Genet. 2007;39:906-913.
22. Price AL1, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet.2006;38:904-909.
23. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics. 2007;26:2190-2191.
24. Antolinia L, Namb BH, D'Agostico RB. Inference on correlated discrimination measures in survival analysis: a nonparametric approach. Commun Stat Theory Methods. 2004;33:2117-2135.
25. Ward LD, Kellis M. HaploReg: a resource for exploring chromatin states, conservation, and
17. Wu Y, Liu X, Li X, Li Y, Zhao L, Chen Z, et al. Estimation of 10-year risk of fffatatatal aandnd ffffffnonfatal ischemic cardiovascular diseases in Chinese adults. Circulation. 2006;11114:4:4:222222171717-2-2-2222222555.
18. Guu DD,,, Reeynynololdsd K, Wu X, Chen J, Duan X, Reeynyy olds RF, et al. Prevvala ence of the metabolic yyyndndndrororome aaandndnd ooveveverweight among adults in China. LaLaancet. 2005;36665:55 13989898-1405.
1191 . National Choooleeesterrrolll EEduduuccacationnn PPProoogrrramm (NNNCEPEPEP) ExExExpepeperrtrt PPPaneeel on DDetttectttioioion, EEvavavalulul aaatiooon, aannndTrTrTreaaatmt ent of Higgghhh Blooooood Chhholollese teroool l innn AAduuulttts (AAAdduultt TTTrrreaattmmmenttt PPPanel IIII). TThihh rd RRReepepooortt of tttheee Natititionononalalal CCChohoholeleestststerrrololol EEEdududucacc tiiononon PPProroroggrammm (((NCNCNCEPEPEP))) ExExExpepepertrtrt PPPanananelelel onn n DeDeDeteteeccctioioon,n,n EEEvavavalululuatatatiooon,n,n, aaandndnd Treatment of High BlBlBlooooo d Cholololesese teerol in Adults (A(AAdud lt Treeeatatatmement Panelele IIIII)) final report. Circulation. 2002;111060606:3:3:3141443-3-3 343434212121...
2020 LiLi YY WiWillllerer CCJJ DDiningg JJ SSchcheeeett PP AAbebecacasisiss GRGR MaMaCHCH:: ususiningg seseququenencece aandnd ggenenototypypee dadatata ttoo
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
18
regulatory motif alterations within sets of genetically linked variants. Nucleic Acids Res.2012;40:D930-934.
26. Varga TV, Sonestedt E, Shungin D, Koivula RW, Hallmans G, Escher SA, et al. Genetic determinants of long-term changes in blood lipid concentrations: 10-year follow-up of the GLACIER study. PLoS Genet. 2014;10:e1004388
27. Lutsey PL, Rasmussen-Torvik LJ, Pankow JS, Alonso A, Smolenski DJ, Tang W, et al. Relation of lipid gene scores to longitudinal trends in lipid levels and incidence of abnormal lipid levels among individuals of European ancestry: the Atherosclerosis Risk in Communities (ARIC) study. Circ Cardiovasc Genet. 2012;5:73-80.
28. Costanza MC, Beer-Borst S, James RW, Gaspoz JM, Morabia A. Consistency between cross-sectional and longitudinal SNP: blood lipid associations. Eur J Epidemiol.2012;27:131-138.
29. Lu Y, Feskens EJ, Boer JM, Imholz S, Verschuren WM, Wijmenga C, et al. Exploring genetic determinants of plasma total cholesterol levels and their predictive value in a longitudinal study. Atherosclerosis. 2010;213:200-205.
30. Isaacs A, Willems SM, Bos D, Dehghan A, Hofman A, Ikram MA, et al. Risk scores of common genetic variants for lipid levels influence atherosclerosis and incident coronary heart disease. Arterioscler Thromb Vasc Biol. 2013;33:2233-2239.
31. Kathiresan S, Melander O, Anevski D, Guiducci C, Burtt NP, Roos C, et al. Polymorphisms associated with cholesterol and risk of cardiovascular events. N Engl J Med.2008;358:1240-1249.
genetic determinants of plasma total cholesterol levels and their predictive value in n n a aa loongnggititududini atudy. Atherosclerosis. 2010;213:200-205.
30. Isaaaacss A, , WiWilllleme s SM, Bos D, Dehghan A, Hoofmfman A, Ikram MA, etet al. Risk scores of coommmmmmono geneneneeeticcc vvvariants for lipid levels influenceee aaatherosclerosis aaand incncnciiident coronary heart dididiseeease. Artererrioooscscsclell r r ThThThrororombmbm VVVasasasc c c BiBiBiololol. 2000131313;3;; 3:22222233-2-2-22323239.9.9.
31311. KaKK thiresan SS, MMMelaaannnder OOO, Anevvvskss iii DDD, GGGuiiducccciii C,C,, BBBurtttt NP,,, RRRoos C,C,C, et t t alall. Polyyymmomorprpphih smmmsssassosoociciciatatatededed wwwiitith h h chchcholololesesesteteerororol andndnd rrisisisk k k of cccararrdididiovovovasasa cucuculaaarr r evevevenenentststs. N NN EnEnEnglglgl JJJ MMMededed.2008;358:1240-1249499... by guest on M
ay 29, 2018http://circgenetics.ahajournals.org/
Dow
nloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
19
Table 1: Baseline characteristics of the participants
Stage of studyDiscovery Replication
CAS BAS GenSalt FAMHES CCHS
sample size 3,998 466 1,881 1,999 14,739
Male/Female 1870/2128 363/103 993/888 1999/0 4222/10517
Age (yrs) 52.53±7.69 51.07±9.53 38.72±9.53 37.5 ±11.1 52.49±8.70
BMI (kg/m2) 25.18±3.79 24.64±3.36 23.35±3.18 23.3±3.4 24.77±3.69
TC (mg/dl) 185.84±35.79 195.34±37.05 170.70±33.65 220.42 ±40.22 183.17±34.93
HDL-C (mg/dl) 49.92±12.42 49.78±12.38 50.97±11.31 54.52 ±12.76 51.20±12.42
TG (mg/dl) 159.24±115.52 128.04±98.84 123.47±78.20 137.24 ±158.49 149.20±110.61
LDL-C (mg/dl) 105.58±30.35 119.79±32.94 95.25±27.30 114.46 ±30.94 103.82±29.73
Cigarette Smoking (%) 32.74 62.23 34.13 50.8 16.29
Alcohol consumers (%) 24.71 46.78 29.24 82.6 20.24
CAS, China atherosclerosis study; BAS, Beijing atherosclerosis study; GenSalt, the Genetic Epidemiology Network of Salt-Sensitivity study; FAMHES, Guangxi Fangchenggang Area Male Health and Examination Survey; CCHS, China Cardiovascular Health Study (CCHS) project; s.d., standard deviation; TC; total cholesterol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; TG, triglycerides; BMI, body mass index.
/Female 1870/2128 363/103 993/888 1999/0 424242222222/1/1/10505051717
yrs) 52.53±7.69 51.07±9.53 38.72±9.53 37.5 ±11.1 52.49±8.70
(kkg/g/g/mmm222)) 22525.18±3.79 24.64±3.36 23233.35±3.18 232 .3±333 44.4 24.77±3.69
mmmg/dddl) 1885.848484±3±33555.797979 11919555.344±3±337.05 17770.70700±±3±33.66565 222 0.424242 ±±40.222222 1811 33..1777±±±34.4.4.9333
-C-CC (m(mmg/g dl) 4999.92±±±1222.42 49.7.778±±±1222.38 5000.97±7±7±11111.3111 545 .52 ±±±12.76766 55151.2220±±±12.4442
mg/dl) 159.2424±115.52 128.04±98.84 123.47±78..2020 137.244 ±158.49 149.20±110.61
C (mg/dl) 105.5.5.585858±3±3±30.0.353535 1111119.9.9 79797 ±3±3±322.2.949494 959595.2.2255±±2777.3.3.3000 1111114.4.4.464646 ±±±303030 9.9.944 4 103.82±29.73
eetttte e SmSmokokining g (%(%)) 3232.7.744 6262.2.233 3434.1.133 5050.8.8 1616.2.299
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
20
Table 2: Association results with lipid levels in the combined discovery and replication studies
Gene SNP CHR BP Code/other allele Code alleleFrequency Trait Beta(S.E.) N P
Previous reported loci in GWAS
PCSK9 rs7525649 1 55271744 T/C 0.72 TC 2.212(0.34) 22612 1.62 × 10-10
LDL-C 1.954(0.29) 22376 2.69 × 10-11
ANGPTL3 rs12042319 1 62822407 G/A 0.81 TC 2.229(0.404) 23028 4.85 × 10-08
TG 0.02(0.003) 23027 8.33 × 10-12
APOB rs312949 2 21187788 C/G 0.28 TC 1.594(0.355) 23031 6.92 × 10-06
LDL-C 1.415(0.303) 22788 2.87 × 10-06
GCKR rs1260333 2 27602128 A/G 0.52 TC 1.773(0.32) 22873 2.94 × 10-08
TG 0.022(0.002) 22873 2.16 × 10-21
HMGCR rs6871667 5 74640498 A/G 0.55 TC 2.214(0.324) 22668 7.99 × 10-12
LDL-C 2.014(0.278) 22427 7.41 × 10-13
MLXIPL rs13231516 7 72501185 T/G 0.87 TG 0.03(0.004) 23039 9.65 × 10-17
LPL rs12678919 8 19888502 A/G 0.91 HDL-C -1.954(0.204) 22648 1.86 × 10-21
TG 0.044(0.004) 22649 3.86 × 10-25
TRIB1 rs2954029 8 126560154 A/T 0.41 TC 2.218(0.324) 23008 1.16 × 10-11
TG 0.023(0.002) 23008 1.86 × 10-20
ABCA1 rs2575876 9 106705560 G/A 0.79 TC 3.224(0.392) 23027 1.17 × 10-16
HDL-C 1.006(0.14) 23027 4.03 × 10-13
ABO rs579459 9 135143989 C/T 0.20 TC 3.038(0.480) 18474 2.72 × 10-10
LDL-C 2.853(0.392) 18259 5.70 × 10-13
APOA1-C3-A4-A5 rs662799 11 116168917 G/A 0.28 HDL-C -2.516(0.126) 23021 1.84 × 10-85
TG 0.081(0.003) 23021 4.18 × 10-213
PTL3 rs12042319 1 62822407 G/A 0.81 TC 2.229(0.404004)4)4) 232323020202888 4.4.4.8588 ×TG 0.02(0.000303) 232323020202777 8.8.8 3333 ×
B rs312949 2 21187788 C/G 0.28 TC 1.594(0.355)5) 23031 6.92 ×LDL-CCC 1.415(0.303) 22788 2.87 ×
R rs1260333 2 27602128 A//GGG 0.52 TC 1.773(0.32) 22873 2.94 ×TGTGTG 0.02222(2(2(00.0.000002)2)2) 2222228878 333 2.2.2.1666 ×
CCCR rs688871667 555 746400049998 A/GGG 0.0 555 TC 2.2221444(0.324) 222 6666888 7.9999 ×LDL-CCC 2.0101014(((0.0 278))) 224442777 7.4441 ×
PL rsss13131323232315151 161616 77 772501111858585 T/T/T/GGG 000.878787 TGTGTG 00.0 030 (0(00 00.0040404))) 232323030303999 9.99 656565 ×××rs1267898989191919 88 19191988888885858502022 AAA/GGG 00.0.919191 HDHDHDLL-C-C-C -1-1-1.9.9.95455 (0.204) 22648 1.86 ×
TGTGTG 0.0.0.040404444(0.004) 22649 3.86 ×111 rs292929545454020202999 888 121212656565606060151515444 A/A/A/TTT 000.414141 TCTCTC 222.2121218(8(8(000.3232324)4)4) 232323000000888 111.161616 ×××
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
21
ALDH2 rs11066280 12 111302166 T/A 0.83 TG 0.017(0.003) 22589 9.79 × 10-9
LIPC rs1800588 15 56510967 T/C 0.38 TC 2.719(0.33) 22845 8.45 × 10-16
HDL-C 1.462(0.118) 22844 1.90 × 10-34
TG 0.018(0.002) 22845 2.52 × 10-14
CETP rs3764261 16 55550825 A/C 0.17 TC 4.439(0.447) 20576 9.69 × 10-23
HDL-C 3.353(0.159) 20575 2.00 × 10-97
HPR rs12927205 16 70582578 A/G 0.74 TC 1.784(0.362) 23005 6.94 × 10-07
LDL-C 1.462(0.309) 22763 8.02 × 10-07
ABCA8 rs12453914 17 64650473 A/C 0.42 TC 1.237(0.279) 23032 1.67 × 10-06
LDL-C 0.993(0.250) 22788 3.66 × 10-05
APOE rs157582 19 50088059 T/C 0.18 TG 0.078(0.008) 12320 3.92 × 10-22
Novel Chinese-specific variants at the previous reported lociHLA-C rs9357121 6 31348458 T/G 0.85 TC 2.062(0.330) 23001 1.50 × 10-10
(tagging rs3177928 r2=0.02) LDL-C 1.842(0.301) 22759 3.28 × 10-10
LIPG rs12970066 18 45361150 G/C 0.28 HDL-C 0.784(0.128) 22850 2.35 × 10-09
(tagging rs7241918 r2=0.017) LDLR rs7258950 19 11111139 G/A 0.78 TC 3.168(0.383) 22495 2.39 × 10-16
(tagging rs6511720 r2=0) LDL-C 3.387(0.326) 22252 1.44 × 10-24
SNP IDs and chromosomal positions are based on NCBI Build 36 of the genome. CHR, chromosome; r2 are obtained from 1000 Genomes data in JPT+CHB
LDL-C 1.462(0.3030309)9)9) 222222767676333 8.8.8.0200 ×8 rs12453914 17 64650473 A/C 0.42 TC 1.237(0.2727279)9)) 2323230303032 2 2 1.1.1.676767 ××
LDL-C 0.993(0.250) 22788 3.66 ×E rs1515575757 82 19 50088059 T/T/T/CCC 0.18 TG 0.078(0.008) 12320 3.92 ×
ChChiininese-specific vararariaiaiantntnts atatt ttthehehe ppprrereviviviouo ss rererepopoportrtrtededed llococociiiCCC rs93335777121 6 313313488845558 T/GGG 0.0 885 TC 2.00606222(0.330))) 232 0000111 1.505050 ×
(t( aggggiing rs33317777928 rrr2=0=0=0.0. 2) LDL-CCC 1.8484842(2((0.301)) 222 7775999 3.2228 ×rsss121212979797000000666666 18 4445361111505050 G/G/G/CCC 0.0.0 282828 HDHDHDLLL-C-C-C 0.0.0 78784(4(4(0.0.0.1212128)8)8) 22222285858500 0 2.2.2 353535 ×××(tagginggg rrrs7s7s7242419188 r222=0=0=0.0.0.0177) )rs72589595950 0 0 191919 11111111111111111 393939 G/G/G AAA 0.0.0.78788 TCTCTC 3.3.3.161616888(0.383) 22495 2.39 ×(((tatatagggggginininggg rrrs6s6s6515151171717202020 rrr2=0=00))) LDLDLDLLL-C-CC 333.3838387(7(7(000.3232326)6)6) 222222252525222 111.444444 ×××
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
22
Table 3: Significant variants associated with lipid levels change and incident hyperlipidemia
SNP IDs and chromosomal positions are based on NCBI Build 36 of the genome. CHR, chromosome
Lipids Change Incident Hyperlipidemia
SNP Gene CHR Position Code allele other allele Beta(S.E.) P OR(S.E.) P
TC change and incident hypercholesterolemia
rs1800588 LIPC 15 56510967 T C 1.985(0.569) 4.84 × 10-04 1.222(0.075) 7.76 × 10-03
rs3764261 CETP 16 55550825 A C 2.596(0.730) 3.83 × 10-04 1.281(0.092) 7.15 × 10-03
rs7258950 LDLR 19 11111139 G A 2.849(0.655) 1.39 × 10-05 1.240(0.093) 2.08 × 10-02
LDL-C change and incident high levels of low-density lipoprotein cholesterol
rs7258950 LDLR 19 11111139 G A 2.464(0.552) 8.27 × 10-06 1.317(0.127) 3.02 × 10-02
HDL-C change and incident low levels of high-density lipoprotein cholesterol
rs2575876 ABCA1 9 106705560 A G -0.847(0.271) 1.77 × 10-03 1.310(0.082) 9.14 × 10-04
rs662799 APOA1-C3-A4-A5 11 116168917 G A -1.256(0.252) 6.24 × 10-07 1.282(0.077) 1.20 × 10-03
rs1800588 LIPC 15 56510967 C T -1.246(0.225) 3.35 × 10-08 1.266(0.074) 1.38 × 10-03
rs3764261 CETP 16 55550825 C A -1.959(0.284) 5.62 × 10-12 1.452(0.100) 1.80 × 10-04
TG change and incident hypertriglyceridemia
rs12678919 LPL 8 19888502 A G 0.030(0.006) 3.31 × 10-06 1.453(0.110) 6.40 × 10-04
rs2954029 TRIB1 8 126560154 A T 0.017(0.004) 6.72 × 10-06 1.303(0.056) 2.33 × 10-06
rs662799 APOA1-C3-A4-A5 11 116168917 G A 0.036(0.004) 4.62 × 10-18 1.629(0.060) 4.67 × 10-16
rs157582 APOE 19 50088059 T C 0.021(0.009) 1.76 × 10-02 1.576(0.118) 1.21 × 10-04
0588 LIPC 15 56510967 T C 1.985(0.569) 4.84 × 10 1.222((0.0.0.0700 5) 7.76 × 10
4261 CETP 16 55550825 A C 2.596(0.730) 3.83 × 10-04 1.282881(1((0.00.090992)2)2) 7.7.7.15151 × 10
8950 LDLR 19 11111139 G A 2.849(0.655) 1.39 × 10-05 1.240(0(0(000.0093)3)3) 222.080808 ××× 111000
C chaangngngeee anananddd inincicicided ntntt hihihigh levels of low-density lipoprotein cholesttererrololol
89595950 0 0 LDLR 19 11111139 G AA 2.464(0.552)2)2) 8.27 × 10-06 1.317(0.127) 3.02 × 10
CCC chhhange and incidentntt looow lel vvvelsss ofofof hhhigighhh--ddedensittty y y llilipoppproootein ccchooolesterrroll
555878 666 ABCA1 9 1000677705560 AAA G -0-0-0.884447(((0.27711) 1.77 ××× 1110-033 1.1 310(000.080882) 9.14 × 10
79999 APAPAPOAOAOA111--CC3C3-AA4A4- -AA5A5- 11111 1161616168688919191777 G GG AA -11-1.2.2.25556(0(0(0.2.252522) )) 6.6.6.2424 ××× 11000-07 1.1.1.2828282(2(2(0.0.0.00077)7)7) 1.1.20200 × 11100
0588 LIPC 15 56665151510900 67 C T T -11.2.2.2464646(0.225) 3.353535 ××× 10-08 1.266(0.074) 1.38 × 10
4261 CETP 161616 55555555555 080808252525 C C C A AA -1-11.9.99595959(0(0(0.2.2.28444) 5.5.5.626262 × 11000-12122 1.452(0.100) 1.80 × 10
hhahangngee ananddd iinin iciciddede tntnt hhyhypepe trtrt iriri lglglycycereridididememiiaia
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
23
Table 4: Association of genetic risk scores for lipid levels with their corresponding lipid levels change and incident hyperlipidemia
Continuous GRS (per SD)*
Quintiles of GRS*
Q2 vs Q1 Q3 vs Q1 Q4 vs Q1
Lipid levels change Beta(S.E.) P value Beta(S.E.) Beta(S.E.) Beta(S.E.) P for trend
TC GRS vs TC change 3.724(0.386) 6.59 × 10-22 4.488(1.082) 7.04(1.085) 9.930(1.092) 9.72 × 10-21
LDL-C GRS vs LDL-C change 2.252(0.325) 4.72 × 10-12 2.012(0.918) 4.53(0.922) 5.529(0.924) 6.55 × 10-11
HDL-C GRS vs HDL-C change -1.699(0.153) 2.06 × 10-28 -1.174(0.445) -2.129(0.440) -4.178(0.433) 4.77 × 10-23
TG GRS vs TG change 0.033(0.003) 5.71 × 10-36 0.021(0.007) 0.038(0.007) 0.081(0.007) 1.42 × 10-28
Incident hyperlipidemia OR(S.E.) P value OR(S.E.) OR(S.E.) OR(S.E.) P for trend
TC GRS vs HyperTC 1.312(0.055) 6.49 × 10-7 1.292(0.175) 1.645(0.168) 1.935(0.162) 1.09 × 10-5
LDL-C GRS vs HyperLDL-C 1.286(0.075) 8.46 × 10-4 0.794(0.238) 1.392(0.210) 1.701(0.203) 5.33 × 10-4
HDL-C GRS vs LowHDL-C 1.351(0.052) 6.17 × 10-9 1.200(0.129) 1.582(0.135) 2.146(0.145) 1.72 × 10-8
TG GRS vs HyperTG 1.482(0.041) 2.00 × 10-21 1.319(0.126) 1.563(0.123) 2.661(0.118) 5.33 × 10-18
GRS, genetic risk score; OR, odds ratio; SD, standard deviation. HyperTC: hypercholesterolemia, HyperTG: hypertriglyceridemia, LowHDL-C:low levels of high-density lipoprotein cholesterol, HyperLDL-C: high levels of low-density lipoprotein cholesterol*The covariates were sex, age, BMI, and the baseline lipid levels corresponding to the genetic risk scores. The OR and Beta for continuous GRS are for 1 standard deviation unit increase in GRS.
RS vs TC change 3.724(0.386) 6.59 × 10-22 4.488(1.082) 7.04(1.085) 9.930(0((1.1.1.09090 2)2)2) 9.9.9.727272 × 1
C GRS vs LDL-C change 2.252(0.325) 4.72 × 10-12 2.012(0.918) 4.53(0.922) 5.52229(9(9(0.00 9292924)4)4) 6.6.6.555555 ××× 111
C GRS vs HDL-C change -1.699(0.153) 2.06 × 10-28 -1.174(0.445) -2.129(0.440) ) -4.178(0.433) 4.77 × 1
RSRSRS vsvsvs TG channngegege 0.033(0.003) 5.71 × 10-36 0..021(0.007) 0..0303038(0.007)7) 0.081(0.007) 1.42 × 1
eeent hhyperlipidemia OR(S..EEE.)) P vvvalluue OR(S..E.E.E.))) ORRR(S.E.))) OR(S..E.E.))) P fofofor tr
RSRSS vs ss HyHyH perTTCCC 1..313 2(0.0050 5)5)5) 6.66 499 × 11100-7 1...2929222(0.0.0.171755)5) 11..64445(5((0.00 1668) 1.1.9399 5((00.0.1616162) 1.0099 ×××
C GRS vs HyperLDL-C 1.1 286(0.0707075)5)5) 8.46 × 10-4 0.794(0.232338)8)8) 1.392((0.0..2121210)00 1.701(0.203) 5.33 ×
C GRS vs LowHDL-C 1.1.1.3535351(1(1(00.0.0505052)2)2) 6.6.6.171717 ×× 111000-9-9-9 1.1.1.202020000(0.1212129)9)9) 1.1..585882(2(2(0.0.0.1313135)5)5) 222.146(0.145) 1.72 ×
RS vs Hyypep rTG 1.482((0.041)) 2.00 × 10-2121 1.319((0.126)) 1.563((0.123)) 2.661((0.118)) 5.33 × 1
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
DOI: 10.1161/CIRCGENETICS.115.001096
24
Figure Legend:
Figure 1: Distribution of genetic risk scores by their corresponding incident hyperlipidemia in
follow up. (a) HyperTC: hypercholesterolemia, (b) HyperLDL-C: high levels of low-density
lipoprotein cholesterol, (c) LowHDL-C: low levels of high-density lipoprotein cholesterol, (d)
HyperTG: hypertriglyceridemia; The y-axis is the proportion of the group (either with or without
incident hyperlipidemia) with a given GRS. Plots were generated using R package PredictABEL.
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
Peng and Dongfeng GuDongshuang Guo, Zhengyuan Zhou, Zili Yang, Renping Wang, Jun Yang, Weili Yan, XiaozhongLiu, Chen Huang, Chong Shen, Jinjin Shen, Ling Yu, Lihua Xu, Jianjun Mu, Xianping Wu, Xu Ji, Yongchen Hao, Jianxin Li, James E. Hixson, Yunzhi Li, Min Cheng, Xiaoli Liu, Jie Cao, Fangcao
Shufeng Chen, Jing Chen, C. Charles Gu, Jichun Chen, Ying Li, Liancheng Zhao, Hongfan Li, Xiangfeng Lu, Jianfeng Huang, Zengnan Mo, Jiang He, Laiyuan Wang, Xueli Yang, Aihua Tan,
Hyperlipidemia in Chinese PopulationsGenetic Susceptibility to Lipid Levels and Lipid Change Over Time and Risk of Incident
Print ISSN: 1942-325X. Online ISSN: 1942-3268 Copyright © 2015 American Heart Association, Inc. All rights reserved.
TX 75231is published by the American Heart Association, 7272 Greenville Avenue, Dallas,Circulation: Cardiovascular Genetics
published online November 18, 2015;Circ Cardiovasc Genet.
http://circgenetics.ahajournals.org/content/early/2015/11/18/CIRCGENETICS.115.001096World Wide Web at:
The online version of this article, along with updated information and services, is located on the
http://circgenetics.ahajournals.org/content/suppl/2015/11/18/CIRCGENETICS.115.001096.DC1Data Supplement (unedited) at:
http://circgenetics.ahajournals.org//subscriptions/
is online at: Circulation: Cardiovascular Genetics Information about subscribing to Subscriptions:
http://www.lww.com/reprints Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer this process is available in the
located, click Request Permissions in the middle column of the Web page under Services. Further information aboutnot the Editorial Office. Once the online version of the published article for which permission is being requested is
can be obtained via RightsLink, a service of the Copyright Clearance Center,Circulation: Cardiovascular Genetics Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:
by guest on May 29, 2018
http://circgenetics.ahajournals.org/D
ownloaded from
1
SUPPLEMENTAL MATERIAL
Supplementary Methods
CAS
The China Atherosclerosis Study (CAS), including 1,010 cases of CAD and 3,998 controls,
was designed to identify the genetic susceptibility genes of coronary artery disease1. In the
present study, 3,998 participants in control were used to perform GWAS meta-analysis of
lipid levels. Study samples who were recruited from the International Collaborative Study
of Cardiovascular Disease in Asia. All individual were judged to be free of coronary,
cerebrovascular, renal diseases and other major chronic diseases by medical history.
Standard questionnaire were used by trained interviewers to obtain information on
demographic characteristics including age, sex, ethnicity, details of medical history,
smoking and alcohol consumption. Overnight fasting blood samples were drawn by
venipuncture to measure serum biochemical measurements including total cholesterol,
high-density lipoprotein cholesterol (HDL-C), and TG. Blood specimens were processed in
the central clinical laboratory at the Department of Population Genetics at Fuwai Hospital
of the Chinese Academy of Medical Sciences in Beijing. This laboratory participates in the
Lipid Standardization Program of the US Centers for Disease Control and Prevention. Total
cholesterol, HDL-C, and TG were analyzed enzymatically on a Hitachi 7060 Clinical
Analyzer (Hitachi High-Technologies Corp). The LDL-C concentrations were calculated by
use of the Friedewald equation.
BAS
The Beijing Atherosclerosis Study (BAS) comprised 466 individuals who were randomly
selected from subjects participating in a community-based survey of cardiovascular risk
factors in Beijing2. Details of medical history, cigarette smoking, and alcohol consumption
were obtained from all participants by standardized questionnaire. Overnight fasting blood
samples were drawn by venipuncture to measure serum biochemical measurements
including TC, HDL-C, and TG. The LDL-C concentrations were calculated by use of the
Friedewald equation.
The GenSalt Study
The Genetic Epidemiology Network of Salt-Sensitivity (GenSalt) study was conducted at
2
six study sites located in rural areas of northern China from October 2003 to July 2005. It
was a family feeding-study designed to examine the interaction between genes and dietary
sodium and potassium intake on BP. Detailed eligibility criteria for the participants have
been reported previously3. A community-based BP screening was conducted among
persons aged 18-60 years in the study villages to identify potential probands and their
families for the study. A total of 1,906 GenSalt probands and their siblings, spouses, and
offspring were eligible and 1,881 of them took part in the dietary intervention and GWAS
genotyping. Overnight (≥8 h) fasting blood specimens were obtained for measurement of
glucose and lipid concentrations. Plasma glucose concentration was measured with a
modified hexokinase enzymatic method (Hitachi automatic clinical analyser, model 7060,
Tokyo, Japan). Concentrations of total cholesterol, HDL cholesterol, and triglycerides were
assessed enzymatically with commercially available reagents. Concentration of LDL
cholesterol was calculated with the Friedewald equation.
FAMHES
The FAMHES project was initiated in Fangchenggang city, Guangxi, southwestern China
in 2009, including 4,303 Chinese men age 17 to 88 years old4. For this study, we included
only those aged 20 to 69 years old who reported Han ethnicity (n = 2,012). After stringent
quality control, 1,999 subjects were included in the analysis. Blood specimens were
obtained after participants had fasted overnight (≥8 h). The plasma TC, TG, LDL and HDL
levels were measured with electrochemiluminescence immunoassay on the COBAS 6000
system E601 immunoassay analyzer (Roche Diagnostics, GmbH, Mannheim, Germany),
with the same batch of reagents according to the manufacturer’s instructions.
CCHS
China Cardiovascular Health Study (CCHS) project has been a population-based
investigation of risk factors for cardiovascular diseases in China since 2006. Data collected
through in-person interviews and examinations include risk factors (based on questionnaire
and blood samples), anthropometry and BP. Overnight fasting blood samples were drawn
by venipuncture to measure serum biochemical measurements including total cholesterol,
High-density lipoprotein cholesterol (HDL-C), TG and glucose (GLU). Blood specimens
were processed in the central clinical laboratory at the Department of Population Genetics
at Fuwai Hospital of the Chinese Academy of Medical Sciences in Beijing. This laboratory
participates in the Lipid Standardization Program of the US Centers for Disease Control
3
and Prevention. Total cholesterol, HDL-C, TG and GLU were analyzed enzymatically on a
Hitachi 7060 Clinical Analyzer (Hitachi High-Technologies Corp). The LDL-C
concentrations were calculated by use of the Friedewald equation.
1. Lu X, Wang L, Chen S, He L, Yang X, Shi Y, et al. Genome-wide association study
in Han Chinese identifies four new susceptibility loci for coronary artery disease.
Nat Genet. 2012;44:890-894
2. Hou L, Chen S, Yu H, Lu X, Chen J, Wang L, et al. Associations of PLA2G7 gene
polymorphisms with plasma lipoprotein-associated phospholipase A2 activity and
coronary heart disease in a Chinese Han population: the Beijing atherosclerosis
study. Hum Genet. 2009;125:11-20.
3. The GenSalt Collaborative Research Group. GenSalt: rationale, design, methods
and baseline characteristics of study participants. J Hum Hypertens.
2007;21:639-646.
4. Tan A, Sun J, Xia N, Qin X, Hu Y, Zhang S, et al. A genome-wide association and
gene-environment interaction study for serum triglycerides levels in a healthy
Chinese male population. Hum Mol Genet. 2012;21:1658-1664.
4
Supplementary Figures
Figure S1. Q-Q plots of the genome-wide association results
a. TC, b. LDL-C, c. HDL-C, d. TG
The observed P values (y-axis) were compared with the expected P values under the null
distribution (x-axis). The Q-Q plot in black includes all SNPs that passed the inclusion
criteria.
c. HDL-C λ=1.030
b. LDL-C λ=1.033 a. TC λ=1.038
d. TG λ=1.033
5
Figure S2. Manhattan plot of genome-wide association results
The genes that are genome wide significant in the discovery analysis are indicated with its
gene names.
6
Figure S3. Effect sizes for single SNPs for lipids change over time are correlated with
the original estimated effects in GWAS
7
Supplementary Tables
Table S1. Stage 1 genome-wide genotyping, imputation and genotype-phenotype analysis by study
Study
Platform
Calling
algorithm
Filtering of genotypes before imputation Number of
SNPs for
Imputation
Imputation
software
NCBI;
HapMap
CHB+JPT
Filtering of
imputed
Genotype-phenotype
software Individuals
call rate
SNP
call
rate
SNP
HWE
SNP
MAF
CAS Axiom
CHB1 BRLMM ≥0.97 ≥0.95 >1.0E-5 ≥0.01 621,779
MACH
v1.0.16 36 r22
Rsq≥0.3;
MAF≥0.01; PLINK v1.07
BAS Affymetrix
500K BRLMM ≥ 0.97 ≥0.95 > 1.0E-4 ≥0.01 367,129
MACH
v1.0.16 36 r22
Rsq≥0.3;
MAF≥0.01; PLINK v1.07
GenSalt Affymetrix
6.0 Birdseed v2 None ≥0.95 >1.0E-5 ≥0.01 662,049
MACH
v1.0.16 36 r22
Rsq≥0.3;
MAF≥0.01; GWAF
FAMHES Illumina
Omin
Beadstudio/
Genomestudio ≥ 0.95 ≥0.95 >1.0E-3 ≥0.01 709,211
IMPUTE
V2.1.2 36 r22
info≥0.9;
MAF≥0.01; SNPTEST
CAS, China atherosclerosis study; BAS, Beijing atherosclerosis study; GenSalt, the Genetic Epidemiology Network of Salt-Sensitivity study; FAMHES, Guangxi Fangchenggang Area
Male Health and Examination Survey; MAF, minor allele frequency
8
Table S2. Association results for follow-up genotyping
SNP CHR BP Code/
other allele
Code allele
Frequency Trait P_discovery P_replication
Meta-analysis
Beta(SE) P N
rs7525649 1 55271744 T/C 0.72 LDL-C 1.19E-05 4.65E-07 1.954(0.29) 2.69E-11 22376
TC 4.25E-05 8.66E-07 2.212(0.34) 1.62E-10 22612
rs12042319 1 62822407 G/A 0.81 TC 4.74E-07 2.36E-03 2.229(0.404) 4.85E-08 23028
TG 1.49E-04 1.22E-08 0.02(0.003) 8.33E-12 23027
rs312949 2 21187788 C/G 0.28 TC 1.50E-04 5.53E-03 1.594(0.355) 6.92E-06 23031
LDL-C 1.25E-04 2.97E-03 1.415(0.303) 2.87E-06 22788
rs1260333 2 27602128 A/G 0.52 TC 2.18E-04 3.22E-05 1.773(0.32) 2.94E-08 22873
TG 2.52E-08 1.44E-14 0.022(0.002) 2.16E-21 22873
rs6871667 5 74640498 A/G 0.55 TC 1.33E-08 4.07E-06 2.214(0.324) 7.99E-12 22668
LDL-C 2.15E-07 3.65E-07 2.014(0.278) 7.41E-13 22427
rs1042391 6 16398740 A/T 0.89 TC 8.85E-04 8.14E-01 1.088(0.508) 2.82E-02 22824
LDL-C 7.07E-04 6.26E-01 1.018(0.435) 1.49E-02 22587
rs9357121 6 31348458 T/G 0.85 TC 2.17E-06 8.25E-05 2.062(0.330) 1.50E-10 23001
LDL-C 4.01E-04 3.40E-07 1.842(0.301) 3.28E-10 22759
rs3127569 6 160837549 T/C 0.14 TC 1.05E-04 1.22E-01 2.148(0.545) 9.90E-05 15530
TG 3.15E-04 8.13E-01 0.009(0.004) 1.33E-02 15530
rs13231516 7 72501185 T/G 0.87 TG 4.49E-05 2.35E-13 0.03(0.004) 9.65E-17 23039
rs12678919 8 19888502 A/G 0.91 HDL-C 1.76E-08 1.80E-14 -1.954(0.204) 1.86E-21 22648
TG 1.91E-10 3.03E-16 0.044(0.004) 3.86E-25 22649
rs2954029 8 126560154 A/T 0.41 TC 1.84E-05 1.37E-07 2.218(0.324) 1.16E-11 23008
9
TG 2.45E-06 7.90E-16 0.023(0.002) 1.86E-20 23008
rs2575876 9 106705560 G/A 0.79 TC 1.12E-06 2.00E-11 3.224(0.392) 1.17E-16 23027
HDL-C 1.41E-04 5.14E-10 1.006(0.14) 4.03E-13 23027
rs10980596 9 112659578 A/C 0.29 HDL-C 1.30E-04 2.11E-01 -0.447(0.126) 9.71E-04 22955
rs579459 9 135143989 C/T 0.8 TC 1.38E-07 1.74E-04 3.038(0.48) 2.72E-10 18474
LDL-C 8.77E-09 6.26E-06 2.853(0.392) 5.70E-13 18259
rs662799 11 116168917 G/A 0.28 TG 9.81E-76 2.11E-139 0.081(0.003) 4.18E-213 23021
HDL-C 2.63E-29 4.79E-58 -2.516(0.126) 1.84E-85 23021
rs671134 11 121558270 A/G 0.08 TC 2.25E-05 7.72E-01 1.418(0.598) 2.06E-02 23036
LDL-C 1.13E-04 2.96E-01 0.79(0.509) 1.37E-01 22792
rs11068274 12 108708557 A/G 0.46 TG 5.08E-03 3.28E-01 0(0.002) 3.66E-01 22984
rs11066280 12 111302166 T/A 0.22 TG 3.25E-05 5.14E-05 0.017(0.003) 9.79E-09 22589
rs4924334 15 37154191 A/C 0.63 HDL-C 8.48E-03 6.03E-01 0.255(0.117) 4.58E-02 22973
rs1800588 15 56510967 T/C 0.38 TC 3.40E-06 4.87E-11 2.719(0.33) 8.45E-16 22845
HDL-C 8.54E-16 1.88E-20 1.462(0.118) 1.90E-34 22844
TG 5.63E-05 7.89E-11 0.018(0.002) 2.52E-14 22845
rs11634431 15 60977088 T/G 0.45 HDL-C 8.24E-04 7.68E-02 -0.084(0.115) 5.61E-01 22875
rs3764261 16 55550825 A/C 0.17 TC 9.46E-03 1.64E-23 4.439(0.447) 9.69E-23 20576
HDL-C 5.60E-23 7.74E-77 3.353(0.159) 2.00E-97 20575
rs12927205 16 70582578 A/G 0.74 TC 1.03E-05 5.70E-04 1.784(0.362) 6.94E-07 23005
LDL-C 4.35E-05 9.06E-04 1.462(0.309) 8.02E-07 22763
rs12453914 17 64650473 A/C 0.42 TC 2.23E-05 1.59E-02 1.237(0.279) 1.67E-06 23032
10
LDL-C 5.57E-04 1.76E-02 0.993(0.250) 3.66E-05 22788
rs12970066 18 45361150 G/C 0.28 HDL-C 5.23E-09 1.67E-03 0.784(0.128) 2.35E-09 22850
TC 4.02E-05 8.42E-02 1.4(0.359) 1.23E-04 22850
rs7258950 19 11111139 G/A 0.78 TC 1.03E-08 2.28E-09 3.168(0.383) 2.39E-16 22495
LDL-C 4.21E-12 2.72E-14 3.387(0.326) 1.44E-24 22252
rs157582 19 50088059 T/C 0.18 TG 2.58E-07 1.73E-16 0.078(0.008) 3.92E-22 12320
11
Table S3. Single variants and lipid levels change and incident hyperlipidemia
Lipids Change Incident Hyperlipidemia
SNP Gene CHR Position Code allele other allele Beta(SE) P OR(SE) P
TC change and incident hypercholesterolemia
rs7525649 PCSK9 1 55271744 T C 0.889(0.591) 1.32 × 10-01 1.131(0.082) 1.34 × 10-01
rs12042319 ANGPTL3 1 62822407 G A 1.13(0.698) 1.06 × 10-01 1.052(0.095) 5.93 × 10-01
rs312949 APOB 2 21187788 C G 1.73(0.614) 4.87 × 10-03 1.231(0.080) 9.16 × 10-03
rs1260333 GCKR 2 27602128 A G 1.052(0.555) 5.83 × 10-02 1.196(0.075) 1.71 × 10-02
rs6871667 HMGCR 5 74640498 A G 1.151(0.553) 3.76 × 10-02 1.133(0.075) 9.49 × 10-02
rs9357121 HLA-C 6 31348458 T G 1.248(0.789) 1.14 × 10-01 1.064(0.109) 5.68 × 10-01
rs2954029 TRIB1 8 126560154 A T 1.344(0.561) 1.67 × 10-02 1.193(0.075) 1.85 × 10-02
rs2575876 ABCA1 9 106705560 G A 2.061(0.676) 2.31 × 10-03 1.143(0.094) 1.53 × 10-01
rs579459 ABO 9 135143989 C T 1.304(0.963) 1.76 × 10-01 1.020(0.131) 8.80 × 10-01
rs1800588 LIPC 15 56510967 T C 1.985(0.569) 4.84 × 10-04 1.222(0.075) 7.76 × 10-03
rs3764261 CETP 16 55550825 A C 2.596(0.730) 3.83 × 10-04 1.281(0.092) 7.15 × 10-03
rs12927205 HPR 16 70582578 A G 0.760(0.627) 2.25 × 10-01 1.188(0.088) 4.92 × 10-02
rs12453914 ABCA8 17 64650473 A C 0.938(0.561) 9.48 × 10-02 1.178(0.075) 2.85 × 10-02
rs7258950 LDLR 19 11111139 G A 2.849(0.655) 1.39 × 10-05 1.240(0.093) 2.08 × 10-02
LDL-C change and incident high levels of low-density lipoprotein cholesterol
rs7525649 PCSK9 1 55271744 T C 0.990(0.497) 4.65 × 10-02 1.139(0.110) 2.38 × 10-01
rs312949 APOB 2 21187788 C G 1.331(0.520) 1.05 × 10-02 1.214(0.107) 6.97 × 10-02
rs6871667 HMGCR 5 74640498 A G 1.322(0.470) 4.90 × 10-03 1.252(0.102) 2.68 × 10-02
rs9357121 HLA-C 6 31348458 T G 1.129(0.665) 8.94 × 10-02 1.093(0.147) 5.45 × 10-01
rs579459 ABO 9 135143989 C T 1.675(0.807) 3.81 × 10-02 1.278(0.164) 1.34 × 10-01
12
SNP IDs and chromosomal positions are based on NCBI Build 36 of the genome. CHR, chromosome.
rs12927205 HPR 16 70582578 A G 1.001(0.531) 5.95 × 10-02 1.340(0.122) 1.63 × 10-02
rs12453914 ABCA8 17 64650473 A C 0.664(0.474) 1.61 × 10-01 1.101(0.101) 3.39 × 10-01
rs7258950 LDLR 19 11111139 G A 2.464(0.552) 8.27 × 10-06 1.317(0.127) 3.02 × 10-02
HDL-C change and incident low levels of high-density lipoprotein cholesterol
rs12678919 LPL 8 19888502 A G -0.884(0.391) 2.37 × 10-02 1.294(0.134) 5.45 × 10-02
rs2575876 ABCA1 9 106705560 A G -0.847(0.271) 1.77 × 10-03 1.310(0.082) 9.14 × 10-04
rs662799 APOA1-C3-A4-A5 11 116168917 G A -1.256(0.252) 6.24 × 10-07 1.282(0.077) 1.20 × 10-03
rs1800588 LIPC 15 56510967 C T -1.246(0.225) 3.35 × 10-08 1.266(0.074) 1.38 × 10-03
rs3764261 CETP 16 55550825 C A -1.959(0.284) 5.62 × 10-12 1.452(0.100) 1.80 × 10-04
rs12970066 LIPG 18 45361150 C G -0.460(0.248) 6.36 × 10-02 1.039(0.079) 6.29 × 10-01
TG change and incident hypertriglyceridemia
rs12042319 ANGPTL3 1 62822407 G A 0.014(0.005) 2.49 × 10-03 1.118(0.071) 1.17 × 10-01
rs1260333 GCKR 2 27602128 A G 0.011(0.004) 1.91 × 10-03 1.114(0.055) 5.11 × 10-02
rs13231516 MLXIPL 7 72501185 T G 0.013(0.005) 1.23 × 10-02 1.297(0.085) 2.33 × 10-03
rs12678919 LPL 8 19888502 A G 0.030(0.006) 3.31 × 10-06 1.453(0.110) 6.40 × 10-04
rs2954029 TRIB1 8 126560154 A T 0.017(0.004) 6.72 × 10-06 1.303(0.056) 2.33 × 10-06
rs662799 APOA1-C3-A4-A5 11 116168917 G A 0.036(0.004) 4.62 × 10-18 1.629(0.060) 4.67 × 10-16
rs11066280 ALDH2 12 111302166 T A 0.008(0.005) 9.85 × 10-02 1.184(0.076) 2.64 × 10-02
rs1800588 LIPC 15 56510967 T C 0.005(0.004) 1.67 × 10-01 1.104(0.057) 8.14 × 10-02
rs157582 APOE 19 50088059 T C 0.021(0.009) 1.76 × 10-02 1.576(0.118) 1.21 × 10-04
13
Table S4. Discrimination after addition of GRS to traditional risk factors
Discrimination
Model without GRS Model with GRS C index (95%CI) C index (95%CI) P
TC GRS vs HyperTC 0.75 (0.73- 0.78) 0.76 (0.74- 0.78) 0.035
LDL-C GRS vs HyperLDL-C 0.76 (0.732- 0.795) 0.77 (0.74- 0.80) 0.122
HDL-C GRS vs LowHDL-C 0.67 (0.65- 0.70) 0.69 (0.67- 0.71) 0.005
TG GRS vs HyperTG 0.68 (0.66- 0.70) 0.70 (0.68- 0.72) 0.0002
GRS, Genetic Risk Score. HyperTC: hypercholesterolemia, HyperTG:hypertriglyceridemia, LowHDL-C: low levels of high-density lipoprotein cholesterol, HyperLDL-C: high levels of
low-density lipoprotein cholesterol
*The covariates were sex, age, BMI, and the baseline lipid levels corresponding to the genetic risk scores.
14
Table S5. The associations of novel Chinese-specific variants and lead SNPs in Europeans in GLGC datasets
Gene Note SNP LD(r2) CHR BP A1/A2 Freq.A1
in Europeans
Freq.A1
in Chinese N_TC P _TC N_LDL P_LDL N_HDL P_HDL N_TG P_TG
HLA Chinese specific SNPs rs9357121 ref 6 31348458 G/T 0.02 0.19 89534 7.29E-01 84967 6.05E-01 89252 6.54E-01 85952 5.03E-01
lead SNPs in Europeans rs3177928 0.02 6 32520413 A/G 0.18 0.05 179996 9.78E-22 165751 3.10E-17 179805 2.93E-03 170516 3.18E-03
LIPG Chinese specific SNPs rs12970066 ref 18 45361150 G/C 0.28 0.31 93067 5.79E-02 88433 4.19E-01 92820 4.77E-07 89485 2.12E-02
lead SNPs in Europeans rs7241918 0.017 18 45414951 G/T 0.19 0.09 93067 3.62E-18 88433 2.62E-04 92820 1.11E-44 89485 4.37E-01
LDLR Chinese specific SNPs rs7258950 ref 19 11111139 A/G 0.11 0.23 8867 2.02E-01 8762 2.33E-02 8868 2.33E-02 8871 6.97E-01
lead SNPs in Europeans rs6511720 0 19 11063306 T/G 0.11 0 184764 5.43E-202 170608 3.85E-262 184617 6.32E-05 175280 1.04E-01
GLGC, Global Lipids Genetics Consortium
All the LD data with r2 are obtained from 1000 Genomes data in JPT+CHB.
15
Table S6. Effects of lipid loci on established cardiovascular risk factors in replication samples
SBP DBP Hypertension BMI FPG
Gene SNP CHR BP
Code
allele
Other
allele Beta(SE) P Beta(SE) P OR(SE) P Beta(SE) P Beta(SE) P
PCSK9 rs7525649 1 55271744 T C -0.11(0.26) 6.67E-01 -0.07(0.15) 6.55E-01 1.01(0.03) 7.26E-01 0(0.04) 9.79E-01 -0.08(0.37) 8.21E-01
ANGPTL3 rs12042319 1 62822407 G A 0.37(0.31) 2.27E-01 0.04(0.17) 8.16E-01 1.02(0.03) 5.77E-01 0.03(0.05) 5.11E-01 -0.3(0.44) 4.99E-01
APOB rs312949 2 21187788 C G -0.37(0.27) 1.79E-01 -0.24(0.15) 1.20E-01 0.97(0.03) 2.77E-01 -0.05(0.05) 2.42E-01 0.49(0.39) 2.05E-01
GCKR rs1260333 2 27602128 A G -0.15(0.25) 5.48E-01 0.05(0.14) 7.26E-01 0.99(0.03) 7.42E-01 -0.05(0.04) 2.67E-01 -1.38(0.35) 7.93E-05
HMGCR rs6871667 5 74640498 A G -0.05(0.25) 8.47E-01 0.14(0.14) 2.92E-01 1(0.03) 9.46E-01 -0.07(0.04) 9.83E-02 0.24(0.35) 4.85E-01
HLA rs9357121 6 31348458 T G -0.55(0.36) 1.20E-01 -0.06(0.2) 7.72E-01 0.96(0.04) 2.12E-01 0.11(0.06) 6.43E-02 0.16(0.51) 7.46E-01
MLXIPL rs13231516 7 72501185 T G 0.2(0.36) 5.85E-01 0.24(0.2) 2.25E-01 1.05(0.04) 2.19E-01 -0.08(0.06) 2.06E-01 0(0.51) 9.98E-01
LPL rs12678919 8 19888502 A G 1.05(0.45) 1.83E-02 0.79(0.25) 1.27E-03 1.15(0.05) 2.42E-03 -0.07(0.08) 3.26E-01 -0.08(0.62) 8.96E-01
TRIB1 rs2954029 8 126560154 A T 0.13(0.25) 5.98E-01 0.03(0.14) 8.49E-01 1.01(0.03) 7.68E-01 -0.06(0.04) 1.34E-01 -0.02(0.36) 9.50E-01
ABCA1 rs2575876 9 106705560 G A 0.52(0.3) 8.75E-02 0.17(0.17) 3.07E-01 1.03(0.03) 3.41E-01 -0.05(0.05) 3.71E-01 0.11(0.43) 8.06E-01
ABO rs579459 9 135143989 C T -0.28(0.78) 7.20E-01 -0.53(0.47) 2.62E-01 0.95(0.09) 5.68E-01 -0.03(0.13) 8.37E-01 -0.64(1.17) 5.85E-01
APOA1-C3-A4-A5 rs662799 11 116168917 G A 0.2(0.27) 4.63E-01 0.11(0.15) 4.53E-01 1.03(0.03) 3.71E-01 0.05(0.05) 2.64E-01 0.12(0.39) 7.48E-01
ALDH2 rs11066280 12 111302166 T A 0.01(0.33) 9.78E-01 0.2(0.18) 2.68E-01 1.03(0.03) 4.31E-01 0.13(0.06) 2.44E-02 -0.06(0.47) 8.92E-01
LIPC rs1800588 15 56510967 T C 0.15(0.25) 5.53E-01 -0.09(0.14) 5.19E-01 0.99(0.03) 8.03E-01 0.09(0.04) 3.98E-02 0.05(0.36) 8.88E-01
CETP rs3764261 16 55550825 A C 0.04(0.33) 9.13E-01 -0.23(0.18) 2.03E-01 0.98(0.03) 5.91E-01 -0.03(0.06) 6.43E-01 0.38(0.46) 4.11E-01
HPR rs12927205 16 70582578 A G -0.29(0.28) 2.97E-01 -0.07(0.15) 6.38E-01 0.98(0.03) 5.33E-01 -0.04(0.05) 3.71E-01 0.81(0.4) 3.99E-02
ABCA8 rs12453914 17 64650473 A C 0.33(0.25) 1.88E-01 0.24(0.14) 7.75E-02 1.02(0.03) 3.91E-01 0.04(0.04) 3.83E-01 -0.75(0.35) 3.42E-02
LIPG rs12970066 18 45361150 G C -0.2(0.28) 4.65E-01 -0.09(0.15) 5.37E-01 0.96(0.03) 1.58E-01 -0.01(0.05) 8.77E-01 -0.16(0.39) 6.76E-01
LDLR rs7258950 19 11111139 G A -0.65(0.29) 2.49E-02 -0.43(0.16) 8.14E-03 0.94(0.03) 4.86E-02 0.06(0.05) 2.46E-01 0.52(0.41) 2.08E-01
APOE rs157582 19 50088059 T C -0.48(0.8) 5.51E-01 -0.46(0.48) 3.34E-01 0.96(0.09) 6.78E-01 -0.12(0.13) 3.68E-01 1.11(1.21) 3.57E-01
Information for SNP ID and chromosomal position is based on NCBI genome build 36. CHR, chromosome; FPG, fasting plasma glucose; BMI, body mass index; SBP, systolic blood
pressure; DBP, diastolic blood pressure. Hypertension was defined by the presence of SBP ≥ 140 mm Hg or DBP ≥ 90 mm Hg or self-reported of taking a medication for the treatment of
hypertension. Variables adjusted in the regression models are sex and age.