1
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
www.na u e.com/scien i ic epo s
1000 Genomes-based me a-
analysis iden i ies 10 no el loci o
kidney unc ion
Ma hias Go ski1,2,*, Pe e J. an de Mos 3,*, Alexande Teume 4,*, Aud ey Y. Chu5,6,*, Man Li7,8,
Vladan Mija o ic9, Ilja M. Nol e3, Massimiliano Cocca10,11, Daniel Taliun12, Felicia Gomez13, Yong Li14,
Bamidele Tayo15, Ad ienne Tin7, Ma y F. Fei osa13, Tho Aspelund16,17, John A ia18,19, Reine Bi a 20,
Mu ielle Bochud21, E ic Boe winkle22, Ing id Bo ecki23, E win P. Bo inge 24, Ming-Huei Chen5,
Vincen Chou aki25, Ma ina Ciullo26,27, Jose Co esh7, Ma ilyn C. Co nelis28, Ga y C. Cu han29,30,
Adamo Pio d’Adamo31, Abbas Dehghan32, Lau a Dengle 2, Jingzhong Ding33, Gudny Ei iksdo i 16,
Ka lhans Endlich34, S e an En o h35, Tõnu Esko36, Osca H. F anco32, Paolo Gaspa ini37,38,
Ch is ian Giege 39,40,41, Gio gia Gi o o37,38, Om i Go esman24, Vilmundu Gudnason16,42, Ul Gyllens en35,
S ephen J. Hancock18,43, Tama a B. Ha is44, Ca he ine Helme 45,46, Simon Hölle e 1, Edi h Ho e 47,48,
Albe Ho man32, Elizabe h G. Holliday19, Geo g Homu h49, F ank B. Hu50, Co nelia Hu h41,51,
Nina Hu i-Kähönen52, Shih-Jen Hwang5, Medea Imboden53,54, Åsa Johansson35, Mika Kähönen55,56,
Wol gang König57,58,59, Holly K ame 15, Be nha d K. K äme 60, Ashish Kuma 53,54,61, Zol an Ku alik21,
Jean-Cha les Lambe 25, Leno e J. Laune 44, Te ho Leh imäki62,63, Ma in H. de Bo s 64, Ge jan Na is64,
Mo is Swe z64, Yongmei Liu33, Ku Lohman33, Ru h J. F. Loos24,65, Yingchang Lu24,
Leo-Pekka Lyy ikäinen62,63, Ma k A. McE oy18, Ch is a Meisinge 41, Thomas Mei inge 66,67,
And es Me spalu36, Ma ie Me zge 68, E elin Mihailo 36, Paul Mi chell69, Ma hias Nauck70,71,
Albe ine J. Oldehinkel72, Ma hias Olden1,5, B enda WJH Penninx73, Gio gio Pis is10, Pe e P. P ams alle 12,
Nicole P obs -Hensch53,54, Olli T. Rai aka i74,75, Raine Re ig76, Paul M. Ridke 6,77, Fe nando Ri adenei a78,
An onie a Robino38, Syl ia E. Rosas79, Douglas Rude e 24, Daniela Ruggie o26, Yasaman Saba80,
Cinzia Sala10, Helena Schmid 80, Reinhold Schmid 47, Rodney J. Sco 81,82, Sanaz Sedagha 32,
Albe V. Smi h16,42, Rossella So ice26,27, Benedic e S engel68, Syl ia S acke83, Kons an in S auch39,84,
Daniela Toniolo10, And e G. Ui e linden78, Sheila Uli i38, Jo ma S. Viika i85,86, Uwe Völke 49,71,
Pe e Vollenweide 87, Hen y Völzke4,71,88, D agana Vucko ic37,38, Melanie Waldenbe ge 40,41, Jie Jin Wang69,
Qiong Yang89, Daniel I. Chasman6,90,91, Ge a d T omp92, Ha old Sniede 3, I is M. Heid1, Ca oline S. Fox5,
Anna Kö gen14,93,†, C is ian Pa a o12,†, Ca s en A. Böge 2,† & Ch is ian Fuchsbe ge 12,†
1Depa men o Gene ic Epidemiology, Uni e si y Regensbu g, Regensbu g, Ge many. 2Depa men o
Neph ology, Uni e si y Hospi al Regensbu g, Regensbu g, Ge many. 3Depa men o Epidemiology, Uni e si y o
G oningen, Uni e si y Medical Cen e G oningen, P.O. box 30.001, 9700 RB G oningen, The Ne he lands. 4Ins i u e
o Communi y Medicine, Uni e si y Medicine G ei swald, Wal he -Ra henau-S . 48, 17475 G ei swald, Ge many.
5NHLBI’s F amingham Hea S udy, F amingham, MA 01702, USA. 6Di ision o P e en i e Medicine, B igham and
Women’s Hospi al and Ha a d Medical School, Bos on, MA, 02215, USA. 7Depa men o Epidemiology, Johns
Hopkins Bloombe g School o Public Heal h, 615N Wol e S , Bal imo e, MD, 21205, USA. 8Di ision o Neph ology
and Depa men o Human Gene ics, Uni e si y o U ah, USA. 9Depa men o Li e and Rep oduc ion Sciences,
Uni e si y o Ve ona, S ada Le G azie 8, 37134, Ve ona, I aly. 10Di ision o Gene ics and Cell Biology, San Ra aele
Scien i ic Ins i u e, 20132, Milano, I aly. 11Depa men o Medical, Su gical and Heal h Sciences, Uni e si y o
T ies e, 34100, T ies e, I aly. 12Cen e o Biomedicine, Eu opean Academy o Bozen/Bolzano (EURAC), a ilia ed o
he Uni e si y o Lübeck, Bolzano, I aly. 13Di ision o S a is ical Genomics, Depa men o Gene ics, Washing on
Uni e si y School o Medicine, S Louis, MO, 63108, USA. 14Di ision o Gene ic Epidemiology, Medical Cen e and
Facul y o Medicine - Uni e si y o F eibu g, F eibu g, Ge many. 15Loyola Uni e si y Chicago, 2160 Sou h Fi s
A enue, Bldg 105, Maywood, IL 60153, USA. 16Icelandic Hea Associa ion, Kopa ogu , Iceland. 17Uni e si y o
Iceland, Reykja ik, Iceland. 18School o Medicine and Public Heal h, Uni e si y o Newcas le, Aus alia. 19Public
Heal h P og am, Hun e Medical Resea ch Ins i u e, Newcas le, New Sou h Wales, Aus alia. 20Clinic o
P os hodon ic Den is y, Ge os oma ology and Ma e ial Science, Uni e si y Medicine G ei swald, Fe dinand-
Saue b uch-S ., 17475 G ei swald, Ge many. 21Ins i u e o Social and P e en i e Medicine, Lausanne Uni e si y
Hospi al (CHUV), Rou e de la Co niche 10, 1010, Lausanne, Swi ze land. 22Uni e si y o Texas Heal h Science
Recei ed: 14 Oc obe 2016
Accep ed: 20 Feb ua y 2017
Published: 28 Ap il 2017
OPEN
www.na u e.com/scien i ic epo s/
2
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
Cen e a Hous on, USA. 23Regene on Gene ics Cen e , Regene on Pha maceu icals, Ta y own, NY, USA. 24The
Cha les B on man Ins i u e o Pe sonalized Medicine, Ichan School o Medicine a Moun Sinai, USA. 25Inse m
U1167, Lille Uni e si y, Ins i u Pas eu de Lille, Lille, F ance. 26Ins i u e o Gene ics and Biophysics, “Ad iano
Buzza i-T a e so”-CNR, Via P. Cas ellino 111, 80131 Napoli, I aly. 27IRCCS Neu omed, ia dell’Ele onica, Pozzilli
(Is), I aly. 28Depa men o P e en i e Medicine, No hwes e n Uni e si y Feinbe g School o Medicine, 680 N Lake
Sho e D i e, Sui e 1400 Chicago, IL 60611, USA. 29Renal Di ision, B igham and Women’s Hospi al, USA. 30Channing
Di ision o Ne wo k Medicine, B igham and Women’s Hospi al, Bos on, MA, USA. 31Clinical Depa men o Medical,
Su gical and Heal h Science, Uni e si y o T ies e, I aly. 32Depa men o Epidemiology, E asmus Medical Cen e ,
Ro e dam, The Ne he lands. 33Wake Fo es School o Medicine, USA. 34Ins i u e o Ana omy and Cell Biology,
Uni e si y Medicine G ei swald, F ied ich-Loe le -S . 23c, 17475 G ei swald, Ge many. 35Depa men o
Immunology, Gene ics, and Pa hology, Biomedical Cen e , SciLi eLab Uppsala, Uppsala Uni e si y, SE-75108
Uppsala, Sweden. 36Es onian Genome Cen e , Uni e si y o Ta u, Ta u, Es onia. 37Depa men o Medical
Sciences, Chi u gical and Heal h Depa men , Uni e si y o T ies e, T ies e, I aly. 38Ins i u e o Ma e nal and Child
Heal h - IRCCS “Bu lo Ga o olo”, T ies e, I aly. 39Ins i u e o Gene ic Epidemiology, Helmhol z Zen um München,
Ge man Resea ch Cen e o En i onmen al Heal h, Ingols äd e Lands . 1, 85764 Neuhe be g, Ge many.
40Resea ch Uni o Molecula Epidemiology, Helmhol z Zen um München - Ge man Resea ch Cen e o
En i onmen al Heal h, Neuhe be g, Ge many. 41Ins i u e o Epidemiology II, Helmhol z Zen um München,
Ge man Resea ch Cen e o En i onmen al Heal h, Ingols äd e Lands . 1, 85764 Neuhe be g, Ge many.
42Facul y o Medicine, Uni e si y o Iceland, Reykja ik, Iceland. 43Heal h Se ices Resea ch G oup, Uni e si y o
Newcas le, Aus alia. 44In amu al Resea ch P og am, Labo a o y o Epidemiology and Popula ion S udies,
Na ional Ins i u e on Aging, USA. 45INSERM, Cen e INSERM Resea ch Cen e U1219, Bo deaux, F ance.
46Uni e si y Bo deaux, ISPED, Bo deaux, F ance. 47Clinical Di ision o Neu oge ia ics, Depa men o Neu ology,
Medical Uni e si y o G az, Aus ia. 48Ins i u e o Medical In o ma ics, S a is ics and Documen a ion, Medical
Uni e si y o G az, Aus ia. 49In e acul y Ins i u e o Gene ics and Func ional Genomics, Uni e si y Medicine
G ei swald, F ied ich-Ludwig-Jahn-S . 15a, 17475 G ei swald, Ge many. 50Depa men o Nu i ion, Ha a d
School o Public Heal h and Channing Di ision o Ne wo k Medicine, B igham and Women’s Hospi al, USA.
51Ge man Cen e o Diabe es Resea ch (DZD), Neuhe be g, Ge many. 52Depa men o Pedia ics, Facul y o
Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e 33014, Finland. 53Uni Ch onic Disease Epidemiology,
Swiss T opical and Public Heal h Ins i u e, Basel, Swi ze land. 54Uni e si y o Basel, Swi ze land. 55Depa men o
Clinical Physiology, Tampe e Uni e si y Hospi al, Tampe e 33521, Finland. 56Depa men o Clinical Physiology,
Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e 33014, Finland. 57Deu sches He zzen um
München, Technische Uni e si ä München, Munich, Ge many. 58DZHK (Ge man Cen e o Ca dio ascula
Resea ch), pa ne si e Munich Hea Alliance, Munich, Ge many. 59Depa men o In e nal Medicine II - Ca diology,
Uni e si y o Ulm Medical Cen e , Ulm, Ge many. 60Uni e si y Medical Cen e Mannheim, 5 h Depa men o
Medicine, Uni e si y o Heidelbe g, Theodo Ku ze U e 1–3, 68167 Mannheim, Ge many. 61Ins i u e o
En i onmen al Medicine, Ka olinska Ins i u e, S ockholm, Sweden. 62Depa men o Clinical Chemis y, Fimlab
Labo a o ies, Tampe e 33520, Finland. 63Depa men o Clinical Chemis y, Facul y o Medicine and Li e Sciences,
Uni e si y o Tampe e, Tampe e 33014, Finland. 64Uni e si y Medical Cen e G oningen, Uni e si y o G oningen,
The Ne he lands. 65The Mindich Child Heal h De elopmen Ins i u e, Icahn School o Medicine a Moun Sinai,
USA. 66Ins i u e o Human Gene ics, Helmhol z Zen um München, Ge man Resea ch Cen e o En i onmen al
Heal h, Neuhe be g, Ge many. 67Ins i u e o Human Gene ics, Technische Uni e si ä München, Munich, Ge many.
68Inse m U1018, Uni e si y Pa is-Sud, UVSQ, Uni e si y Pa is-Saclay, Villejui , F ance. 69Cen e o Vision Resea ch,
Depa men o Oph halmology and Wes mead Ins i u e o Medical Resea ch, Uni e si y o Sydney C24, NSW,
2145, Aus alia. 70Ins i u e o Clinical Chemis y and Labo a o y Medicine-Uni e si y Medicine G ei swald,
Fe dinand-Saue b uch-S ., 17475 G ei swald, Ge many. 71DZHK (Ge man Cen e o Ca dio ascula Resea ch),
pa ne si e G ei swald, G ei swald, Ge many. 72Depa men o Psychia y, Uni e si y o G oningen, Uni e si y
Medical Cen e G oningen, P.O. box 30.001, 9700 RB, G oningen, The Ne he lands. 73Depa men o Psychia y,
V ije Uni e si ei , VU Uni e si y Medical Cen e , NESDA, A.J. E ns s aa 1187, 1081HL Ams e dam, The
Ne he lands. 74Depa men o Clinical Physiology and Nuclea Medicine, Tu ku Uni e si y Hospi al, Tu ku 20521,
Finland. 75Resea ch Cen e o Applied and P e en i e Ca dio ascula Medicine, Uni e si y o Tu ku, Tu ku 20520,
Finland. 76Ins i u e o Physiology, Uni e si y Medicine G ei swald, 17475 G ei swald, Ge many. 77Di ision o
Ca dio ascula Medicine, B igham and Women’s Hospi al and Ha a d Medical School, Bos on MA 02115, USA.
78Depa men o In e nal Medicine, E asmus Medical Cen e , Ro e dam, The Ne he lands. 79Joslin Diabe es
Cen e . Ha a d Medical School, Bos on, MA, USA. 80Ins i u e o Molecula Biology and Biochemis y, Cen e o
Molecula Medicine, Medical Uni e si y o G az, Aus ia. 81School o Biomedical Sciences and Pha macy, Uni e si y
o Newcas le, Aus alia. 82Molecula Medicine, Pa hology No h Ph. 0409926764, Newcas le, Aus alia. 83Clinic o
In e nal Medicine A, Uni e si y Medicine G ei swald, Fe dinand-Saue b uch-S ., 17475 G ei swald, Ge many.
84Ins i u e o Medical In o ma ics, Biome y and Epidemiology, Chai o Gene ic Epidemiology, Ludwig-
Maximilians-Uni e si ä , Munich, Ge many. 85Di ision o Medicine, Tu ku Uni e si y Hospi al, Tu ku 20521, Finland.
86Depa men o Medicine, Uni e si y o Tu ku, Tu ku 20520, Finland. 87Depa men o In e nal Medicine, Lausanne
Uni e si y Hospi al (CHUV), Lausanne, Swi ze land. 88DZD (Ge man Cen e o Diabe es Resea ch), Si e G ei swald,
G ei swald, Ge many. 89Depa men o Bios a is ics, Bos on Uni e si y School o Public Heal h, 715 Albany S ee ,
Bos on, MA 02118, USA. 90Di ision o Gene ics, B igham and Women’s Hospi al and Ha a d Medical School,
Bos on MA, USA. 91B oad Ins i u e o MIT and Ha a d, Camb idge MA 02142 USA. 92Weis Cen e o Resea ch,
Geisinge Clinic, Dan ille, Pennsyl ania, USA. 93Depa men o Epidemiology, Johns Hopkins Bloombe g School o
Public Heal h, Bal imo e, USA. *These au ho s con ibu ed equally o his wo k.†These au ho s join ly supe ised
his wo k. Co espondence and eques s o ma e ials should be add essed o M.G. (email: ma hias.go ski@klinik.
uni- egensbu g.de) o C.F. (email: [email p o ec ed]) o C.P. (email: [email p o ec ed])
www.na u e.com/scien i ic epo s/
3
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
HapMap impu ed genome-wide associa ion s udies (GWAS) ha e e ealed >50 loci a which common
a ian s wi h mino allele equency >5% a e associa ed wi h kidney unc ion. GWAS using mo e
comple e e e ence se s o impu a ion, such as hose om The 1000 Genomes p ojec , p omise o
iden i y no el loci ha ha e been missed by p e ious e o s. To in es iga e he alue o such a mo e
comple e a ian ca alog, we conduc ed a GWAS me a-analysis o kidney unc ion based on he
es ima ed glome ula il a ion a e (eGFR) in 110,517 Eu opean ances y pa icipan s using 1000
Genomes impu ed da a. We iden i ied 10 no el loci wi h p- alue < 5 × 10−8 p e iously missed by
HapMap-based GWAS. Six o hese loci (HOXD8, ARL15, PIK3R1, EYA4, ASTN2, and EPB41L3) a e
agged by common SNPs unique o he 1000 Genomes e e ence panel. Using pa hway analysis, we
iden i ied 39 signi ican (FDR < 0.05) genes and 127 signi ican ly (FDR < 0.05) en iched gene se s,
which we e missed by ou p e ious analyses. Among hose, he 10 iden i ied no el genes a e pa o
pa hways o kidney de elopmen , ca bohyd a e me abolism, ca diac sep um de elopmen and glucose
me abolism. These esul s highligh he u ili y o e-impu ing om dense e e ence panels, un il whole-
genome sequencing becomes easible in la ge samples.
Ch onic kidney disease (CKD) is a majo public heal h conce n a ec ing ~10% o he global adul popula ion1.
CKD is de ined based on he glome ula il a ion a e es ima ed om se um c ea inine (eGFRc ea), a quan i a-
i e pheno ype o which 53 loci ha e been iden i ied so a by me a-analyses o genome-wide associa ion s udies
(GWAS)2–7. These GWAS me a-analyses we e based on ~2.5 million a ian s impu ed om he HapMap P ojec
e e ence panel8. Simila o he gene ic a ian s iden i ied o o he pheno ypes, all a ian s associa ed wi h eGFR-
c ea had a mino allele equency (MAF) o > 5%. Howe e , hough he i abili y o eGFR has been es ima ed in
amily s udies o ange be ween 36–75%9,10, he iden i ied a ian s explain less han 4% o he a iance o eGFR-
c ea7 and a e loca ed in egions o ex ended linkage disequilib ium (LD). So a , causal genes o a ian s ha e only
been iden i ied o a ew o he associa ion signals11,12.
I has been shown ha a ian s poo ly agged by GWAS a ays and HapMap impu a ion, pa icula ly
low- equency a ian s (1% ≤ MAF ≤ 5%), can explain addi ional a iabili y13. Recen echnological ad ances
esul ed in la ge collec ions o whole-genome sequence da a, such as hose om The 1000 Genomes p ojec 14,15.
These da a p o ide be e co e age and inc eased impu a ion quali y compa ed o p e ious HapMap impu a-
ion16, pa icula ly o low- equency a ian s.
We unde ook a me a-analysis o GWAS om 33 s udies ha impu ed geno ypes om The 1000 Genomes
e e ence panel, hypo hesizing ha his would unco e no el common a ian s associa ed wi h eGFRc ea, ex end
o low- equency a ian s, e eal no el pa hways o eGFRc ea associa ed genes, and imp o e ine-mapping o
known eGFRc ea loci p e iously iden i ied by ou HapMap-based GWAS3–7.
Resul s
S udy cha ac e is ics. In o al, 110,517 adul indi iduals o Eu opean ances y om 33 s udies pa icipa ed
in GWAS me a-analysis o eGFRc ea using geno ypes impu ed wi h The 1000 Genomes Phase I e e ence panel14
(1000 Genomes me a-analysis). In addi ion, we pe o med a GWAS me a-analysis o eGFR de i ed om cys a in
C (eGFRcys), an al e na i e ma ke o kidney unc ion a ailable in 11 o he 33 s udies (n = 24,063). Pa icipa ing
s udies, pheno ypic cha ac e is ics, geno ype in o ma ion, and me hods o analysis a e epo ed in Supplemen a y
Tables1,2,3and4, espec i ely. The 1000 Genome me a-analysis esul s on eGFRc ea a e compa ed wi h ou
p e iously published HapMap impu ed da a7, which was a HapMap-based me a-analysis o 133,814 Eu opean
ances y indi iduals om 50 s udies.
Impu a ion quali y o a ian s impu ed wi h The 1000 Genomes e e ence panel. The
1000 Genomes me a-analysis consis ed o 10,971,307 gene ic a ian s (10,159,097 SNPs and 812,210
inse ion-dele ions) wi h impu a ion quali y IQ > 0.417 in each o he s udies and p esen in a leas 50% o
he subjec s. Depending on he impu a ion me hodology used, he IQ was epo ed as RSQ18 o in o-sco e19
(Supplemen a yTable3). Compa ed o he HapMap me a-analysis, he 1000 Genomes me a-analysis included a
highe numbe o well impu ed a ian s (8,103,124 e sus 2,249,027 a ian s wi h IQ > 0.8), pa icula ly among
he low- equency a ian s (1,585,176 e sus 191,580, Supplemen a yTable5). While a e a ian s (MAF ≤ 1%)
we e no a ailable in he p e ious HapMap me a-analysis, he e we e e en 632,526 well-impu ed a e a ian s in
he 1000 Genomes me a-analysis. When limi ing he compa ison o a ian s a ailable in bo h panels, he p opo -
ion o well-impu ed a ian s was highe in he 1000 Genomes compa ed o he HapMap me a-analysis (96.9%
e sus 93.3% o all; 88.3% e sus 78.4% o he less equen a ian s, Supplemen a yTable5).
1000 Genomes me a-analysis esul s. The 1000 Genomes me a-analysis iden i ied 49 genome-wide
signi ican loci o eGFRc ea including 10 no el loci (lead a ian p- alue < 5 × 10−8, Table1, Fig.1, and
Supplemen a yFigu e1). All iden i ied lead a ian s we e SNPs, and all we e common, excep s187355703
nea HOXD8 (MAF = 0.03). None o he no el loci con ained genes known o cause monogenic o ms o kid-
ney disease and o mos genes no connec ion o kidney unc ion o kidney disease has ye been desc ibed
(Supplemen a yTable6). Howe e , i should be acknowledged ha gene ic a ian s iden i ied in GWAS a e no
necessa ily associa ed wi h he unc ion o he physically closes gene. O he 53 known eGFRc ea loci iden-
i ied p e iously based on HapMap2–7, 39 we e also genome-wide signi ican in he cu en 1000 Genomes
me a-analysis (Supplemen a yTable7) and he emaining 14 showed di ec ions o associa ion consis en wi h
published epo s, bu did no each signi icance (p- alues 2.2 × 10−2 o 5.2 × 10−7; Supplemen a yTable8). These
www.na u e.com/scien i ic epo s/
4
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
esul s a e consis en wi h ou expec a ions om powe compu a ions (Fig.2). Among he 39 lead a ian s in
p e iously published loci ha we e genome-wide signi ican in he 1000 Genomes me a-analysis, 6 lead a ian s
we e ound o be he same as he p e iously published a ian s, 25 we e highly co ela ed ( 2 > 0.6), and 8 showed
mode a e o no co ela ion ( 2 ≤ 0.6).
The 1000 Genomes me a-analysis o eGFRcys con i med p e iously iden i ied loci in o nea CST3/CST9
(p- alue = 4.1 × 10−153), UMOD (p- alue = 2.9 × 10−10), and ATXN2 (p- alue = 1.6 × 10−8), bu did no e eal
any no el signal.
The en no el eGFRc ea loci in he con ex o he di e en e e ence panels. Fo six o he
en no el loci (HOXD8, ARL15, PIK3R1, EYA4, ASTN2, and EPB41L3), he lead a ian iden i ied in he 1000
Genomes me a-analysis was no obse ed in any p e ious HapMap me a-analysis and in ac was no geno-
yped as pa o he HapMap e e ence panel. Mo eo e , no a ian in LD wi h any o hese six lead a ian s
( 2 o D’ ≥ 0.4) was a ailable in he HapMap panel. These loci ha e been missed due o he limi ed co e age o he
HapMap panel.
Fo one u he locus, RHOC, he 1000 Genome me a-analysis lead a ian was p esen also in ou p e ious
HapMap me a-analysis, bu wi h a lowe impu a ion quali y (1000 Genomes median IQ ac oss all s udies o
0.96 e sus HapMap median IQ o 0.86). The e ec size was sligh ly highe in he 1000 Genomes compa ed o
he HapMap me a-analysis (0.0061 e sus 0.0051 ln ml/min/1.73 m2, Supplemen a yTable9). This locus migh
ha e been missed in he HapMap me a-analysis due o he highe unce ain y in he impu ed geno ypes, which is
known o diminish powe and o a enua e e ec size in linea eg ession20.
Fo he emaining h ee loci (LPHN2, SLC7A6 and RNF152), he lead a ian s o he 1000 Genomes
me a-analysis we e obse ed in he HapMap me a-analysis and simila ly well impu ed (IQ nea 1.0 o bo h panels).
The e ec sizes we e simila o all h ee SNPs in bo h 1000 Genomes and HapMap me a-analyses (0.0057 e sus
0.0041, 0.0061 e sus 0.0049, 0.0064 e sus 0.0050 ln ml/min/1.73 m2 espec i ely) and he HapMap es ima es
lie well wi hin he 98.5% con idence in e al o he 1000 Genomes es ima es. No subs an ial be ween-s udy he -
e ogenei y was obse ed (I2 = 19%, 0%, o 21%, espec i ely, Supplemen a yTable9). Since he p- alues in he
HapMap analysis we e jus sho o genome-wide signi icance (p- alues 8.38 × 10−6 o 2.33 × 10−7; ype II e o
o 14–29%), i is concei able ha hese a ian s ha e been missed p e iously by chance.
Pa hway analyses. Da a-d i en Exp ession P io i ized In eg a ion o complex T ai s (DEPICT)21 anal-
ysis o eGFRc ea iden i ied 39 signi ican (FDR < 0.05) genes and 127 signi ican ly (FDR < 0.05) en iched
gene se s ha we e no iden i ied p e iously7. Among hose, 23 gene se s con ained a leas one o he 10 no el
index genes as a op 10 hi , unde pinning he in luence o u e e ic bud mo phogenesis on kidney de elopmen
and he in luence o abno mal glucose homeos asis and glucan me abolic p ocess on ca bohyd a e me abo-
lism (Supplemen a yTable10). All 127 signi ican gene se s we e u he g ouped in o me a gene se s, co e-
sponding o hei co ela ion o gene exp ession. The wo mos signi ican me a gene se s we e Ca diac Sep um
De elopmen (p- alue = 4.48 * 10−5) and Glucose Me abolism (p- alue = 6.11 * 10−5), con aining one o he 10
no el index genes (Supplemen a yFigu e2). We epea ed he analysis wi h a ying pa ame e s (50, 200, and
500 epe i ions and 500, 2000, and 5000 pe mu a ions, espec i ely), con i ming ou p ima y op gene se s a an
FDR o < 0.05. P- alues anged om 1.32 × 10−3 o 4.48 × 10−5 and om 8.27 × 10−4 o 4.98 × 10−5 o Ca diac
Sep um De elopmen and Glucose Me abolism, espec i ely. We eplica ed also he s ong in luence o emb yonic
de elopmen , kidney ansmemb ane anspo e ac i i y, and kidney and u ogeni al sys em mo phology in he
genesis o CKD om ou p e ious indings7: en ichmen o all 148 p e iously iden i ied gene se s was nominally
signi ican (p- alue < 0.05).
Va ian ID Ch Posi ion
Index
Gene
E ec allele/
non-e ec allele
E ec allele
equency E ec (SE) p- alue I2 (%) IQ
Numbe o subjec s
in analysis
s10874312 1 82,944,571 LPHN2 A/G 0.67 − 0.0057 (0.0011) 2.20 × 10−08 19 1.00 107,335
s12144044 1 113,248,791 RHOC A/C 0.28 − 0.0061 (0.0011) 2.87 × 10−08 0 0.96 110,517
s187355703 2 176,993,583 HOXD8 C/G 0.97 0.0182 (0.0030) 5.15 × 10−10 5 0.89 109,257
s111366116 5 53,295,546 ARL15 T/C 0.11 0.0094 (0.0015) 6.27 × 10−10 22 0.97 110,517
s113246091 5 67,739,274 PIK3R1 A/G 0.10 − 0.0095 (0.0016) 1.98 × 10−09 43 0.98 110,105
s7764488 6 133,812,872 EYA4 A/G 0.32 0.0061 (0.0011) 4.08 × 10−09 1 0.98 110,516
s13298297 9 119,264,108 ASTN2 A/G 0.20 − 0.0075 (0.0014) 1.53 × 10−08 0 0.81 110,514
s1111571 16 68,363,181 SLC7A6 A/G 0.71 0.0061 (0.0011) 6.20 × 10−09 0 1.00 109,275
s9962915 18 5,593,171 EPB41L3 T/C 0.48 − 0.0055 (0.0010) 7.19 × 10−09 0 0.98 110,516
s12458009 18 59,350,507 RNF152 T/G 0.78 − 0.0064 (0.0012) 2.90 × 10−08 21 1.00 107,325
Table 1. The 10 no el genome-wide signi ican loci (p < 5 × 10−8) associa ed wi h eGFRc ea in up o
110,517 subjec s om up o 33 s udies. Posi ions a e gi en on GRCh build 37. The gene closes o he a ian
is lis ed (index gene). E ec sizes a e gi en on he log scale. IQ = Impu a ion quali y me ic compu ed as
median o in o sco e (Impu eV2) o RSQ (minimac) ac oss s udies. SE = s anda d e o . I2 = be ween-s udy
he e ogenei y s a is ic.
www.na u e.com/scien i ic epo s/
5
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
Figu e 1. Manha an Plo o he esul s o he 1000 Genome me a-analysis o eGFRc ea. Shown a e he
(− log10) p- alues by genomic posi ion (GRCh build 37). Highligh ed a e he 10 no el loci iden i ied wi h
genome-wide signi icance (blue, anno a ed by nea es gene), he 39 p e iously published2–7 and con i med
(genome-wide signi ican ) loci (g een) and he 14 p e iously published loci ha we e no genome-wide
signi ican in his analysis (o ange).
Figu e 2. E ec s o he 1000 Genomes lead a ian s o all no el and known loci. Shown a e he e ec sizes
and mino allele equencies (MAF) o he 1000 Genomes lead a ian s ( a ian s wi h smalles p- alue) in each
o he 10 no el (blue), he 39 known genome-wide signi ican loci (g een), and he 14 known loci ha we e no
genome-wide signi ican in his analysis (o ange). Addi ionally, he 80% powe o de ec such e ec s in a sample
size o 110,000 subjec s (as in his 1000 Genomes me a-analysis) is shown as a ed line. A known locus is de ined
by he published lead a ian ± 1 Mb; a no el locus is de ined by he 1000 Genome lead a ian ± 1 Mb.
Independen associa ion signals a no el and known loci. To iden i y independen associa ion
signals wi hin a known o no el locus, we pe o med join condi ional analysis o eGFRc ea based on agg e-
ga ed s udy-speci ic s a is ics using he GCTA so wa e22. Among he combined 49 loci (39 known and 10 no el)
a aining genome-wide signi icance, we unco e ed eigh independen signals, all among he p e iously epo ed
loci, wi h p- alues anging om 2.39 × 10−8 o 2.78 × 10−17 a e condi ioning on he lead a ian s a each locus
(Supplemen a yTable11 and Supplemen a yFigu e3). We ound ha in all bu one locus (DDX1), he p e iously
epo ed lead a ian was also genome-wide signi ican in ou 1000 Genomes me a-analysis. A mo e de ailed

www.na u e.com/scien i ic epo s/
6
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
easoning o he independen associa ion signals is p oposed in Supplemen a yTable12. In o ma ion abou
biological knowledge o he highligh ed genes is p esen ed in Supplemen a yTable13.
P opo ion o pheno ypic a iance explained and polygenic isk sco e (PRS) analysis. The
o e all p opo ion o pheno ypic a iance o eGFRc ea explained by he lead a ian s o he 1000 Genomes
me a-analysis in all no el and known loci was 3.99%: 0.46% by he 10 lead a ian s in he no el loci, 3.12% by he
39 lead a ian s in he known loci, and 0.41% by he 1000 Genomes lead a ian s in he 14 known loci ha we e
no genome-wide signi ican in his analysis.
Nex , we es ed he p opo ion o eGFRc ea a iance ha could be explained by common gene ic a ian s in
1,071 independen adolescen s pa icipa ing in he TRAILS s udy. Gi en p io e idence ha eGFRc ea-associa ed
genes a e p e e en ially exp essed in he kidney and en iched o genes impo an in kidney de elopmen 23, ex e -
nal in luences on eGFRc ea such as hose o he wo main d i e s o CKD, diabe es and hype ension, may
be less impo an in his se ing. In TRAILS, he maximum p opo ion o a iance explained by SNPs asso-
cia ed a p e-de ined p- alue h esholds was 2.2% o a PRS composed o SNPs associa ed wi h eGFRc ea a
p- alue < 1 × 10−5 (Supplemen a yTable14).
SNP-based he i abili y analysis. The he i abili y es ima e using a ian s o MAF > 0.01 o eGFRc ea
in he ARIC s udy was 0.21 (95% CI 0.14–0.28) and 0.31 (95% CI 0.20–0.41) o all a ian s. This is in line wi h
es ima es in he li e a u e om popula ion-based amily s udies such as he F amingham Hea S udy (adjus ed
h2 0.33, 95% CI 0.19–0.47)24.
Exp ession quan i a i e ai loci (eQTL) lookup. To explo e po en ial unc ional implica ions o he
no el loci, we in e oga ed published da abases o cis eQTL in whole blood25 o he signi ican SNPs o hei
p oxy a ian s ( 2 > 0.8 wi hin a 1 MB window). A 2 no el loci, signi ican associa ion (p- alue < 0.004) wi h
gene exp ession we e ound: s1111571 wi h SLC7A6, ZFP90, LYPLA3 and NFATC3, and o s12144044 wi h
RHOC and ST7L (Supplemen a yTable15).
We expanded ou downs eam analysis by anno a ing he signi ican a ian s wi h known and p edic ed eg-
ula o y elemen s using Regulome DB26: We con i med s1111571 and s12144044 as signi ican associa ions wi h
gene exp ession and ound suppo ing e idence ha hese wo a ian s show also e idence o ansc ip ion
ac o binding si es and DNase peaks. Fo he locus iden i ied by s187355703 no p oxy was ound o lookup.
Gene ic co ela ion. To in es iga e he gene ic co ela ion o se um c ea inine wi h ela ed pheno ypes, we
que ied LD Hub27 and iden i ied modes gene ic co ela ion wi h me abolic synd ome ai s such as HDL, LDL,
Type 2 diabe es, as ing glucose, BMI, and wais (LD sco e eg ession gene ic co ela ion be ween − 0.07 and
0.05). Li le e idence o kidney damage is epo ed o a isk sco e o SNPs which a e signi ican p edic o s o
blood p essu e28.
Discussion
The main inding o ou s udy is ha impu ing om dense and la ge e e ence panels is a alid s a egy o
ad ance gene mapping e en when he sample size canno be inc eased. Using geno ype impu a ion based on The
1000 Genomes panel led o he iden i ica ion o 10 no el genome-wide signi ican loci o kidney unc ion ha
we e missed by ea lie HapMap-impu ed GWAS o la ge sample size, pa ly due o he enhanced co e age o
genomic a ia ion. This phenomenon was obse ed in simila analyses o o he pheno ypes29. S ill, i needs o be
acknowledged ha he addi ional p opo ion o ai a iance explained by hese new loci is mode a e, which is
also in line wi h indings om GWAS o o he pheno ypes30.
The e a e se e al me hodological insigh s ha can be gained om ou analyses. Fi s , his 1000
Genomes-based me a-analysis o 110,517 indi iduals has iden i ied 10 no el loci and 8 independen associa ion
signals in known loci ha we e missed by ou la es HapMap based analysis7. Ou de ailed dissec ion shows
ha 1000 Genomes impu a ion (i) p o ides a ian s missed o poo ly agged by HapMap based analysis and
(ii) achie es a highe e ec i e sample size h ough inc eased impu a ion quali y.
Second, al hough he 1000 Genomes impu a ion enables he analysis o low- equency a ian s, inse ions
and dele ions, all iden i ied op a ian s we e SNPs, and all bu one (nea HOXD8) we e common. Mo eo e , we
did no iden i y any low- equency a ian o la ge e ec . Ou esul s a e highly conco dan wi h hose o o he
ecen complex diseases s udies31 showing ha low- equency a ian s a e also con ibu ing o complex disease
isk, bu ha mos obse ed e ec sizes a e small o modes , and hund eds o housands o subjec s a e equi ed
o de ec ion. To iden i y he con ibu ion o a e a ian s (MAF < 1%) o eGFRc ea, la ge-scale sequencing da a
in addi ion o genomic chip da a ha e been shown o be a p omising app oach31.
Thi d, hese no el loci, missed by ou p e ious analysis7, ex end ou knowledge o pa hways unde lying kidney
unc ion, which depic s he in luence o kidney de elopmen , kidney s uc u e, and me abolic ac i i y on he
de elopmen o CKD.
The compa ison o ou 1000 Genomes me a-analysis wi h ou p e ious HapMap me a-analysis is limi ed by
se e al ac o s: he cu en analysis consis s o a educed numbe o samples and a sligh ly di e en s udy com-
posi ion. Fu he mo e, di e en 1000 Genomes e e ence panels we e used o impu e geno ypes and ad ances
in impu a ion so wa e and me hodology mus be acknowledged32,33. Ne e heless, six o he en lead a ian s
in he no el loci a e only co e ed by The 1000 Genomes e e ence panels, which demons a es he ad an age o
me a-analyses on 1000 Genomes o e HapMap impu ed geno ypes.
In conclusion, we iden i ied 10 no el loci and 8 addi ional independen associa ion a ian s wi hin known loci
associa ed wi h kidney unc ion and iden i ied 127 no el pa hways o kidney unc ion. These esul s highligh
www.na u e.com/scien i ic epo s/
7
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
he u ili y o e-impu ing s udies om imp o ed e e ence panels as an in e media e cos -e icien app oach o
scan he ull allelic equency ange o kidney unc ion associa ed a ian s, un il whole genome sequencing is
easible in la ge samples.
Me hods
Pheno ype de ini ion. Each s udy measu ed se um c ea inine as desc ibed in Supplemen a yTable1.
Be ween-labo a o y a ia ion has been accoun ed o by calib a ing c ea inine o he US na ionally ep esen a i e
Na ional Heal h and Nu i ion Examina ion S udy (NHANES) da a in all s udies4,34,35. GFR based on se um c e-
a inine (eGFRc ea) was es ima ed using he ou - a iable Modi ica ion o Die in Renal Disease (MDRD) S udy
Equa ion36,37. In a subse o s udies, se um cys a in C was also ob ained and eGFRcys es ima ed as 76.7*(se um
cys a in C)−1 19 (see also e . 38). The eGFRc ea and eGFRcys alues < 15 ml/min/1.73 m2 we e se o 15, and al-
ues > 200 we e se o 200 ml/min/1.73 m2. I no s a ed o he wise, ou p esen ed da a and esul s a e o eGFRc ea,
which was ou main analysis.
Geno yping. Geno yping was conduc ed in each s udy as speci ied in Supplemen a yTable3. A e applying
app op ia e quali y il e s, pa icipa ing s udies pe o med geno ype impu a ion wi h s anda d impu ing p o-
cedu es32,33,39 using any e sion o he 1000 Genome Phase 1 e e ence panels. The ob ained impu ed gene ic
a ian s we e coded as allelic dosages. De ails o s udy speci ic impu a ion p ocedu e and speci ic e e ence panel
a e gi en in Supplemen a yTable3.
Genome-wide associa ion analysis. Each s udy pe o med GWAS acco ding o a uni o m analysis plan
by eg essing sex- and age-adjus ed esiduals o he na u al loga i hm o eGFRc ea and eGFRcys on he allelic
dosage le els. When app op ia e, adjus men o s udy-speci ic ea u es such as s udy si e o gene ic p incipal
componen s was included in he model. Family-based s udies accoun ed o ela edness using mixed e ec mod-
els. De ails on he s udy-speci ic me hods a e epo ed in Supplemen a yTable4.
GWAS me a-analysis. All GWAS iles unde wen quali y con ol using he GWA oolbox package40. GWAS
me a-analyses o eGFRc ea and eGFRcys we e pe o med using he so wa e METAL41 assuming ixed e ec s
ac oss s udies and using in e se- a iance weigh ing, excluding a ian s wi h impu a ion quali y IQ ≤ 0.4 o a i-
an s p esen in less han 50% o he 110,517 subjec s (yielding 10,971,307 a ian s). The genomic in la ion ac o
λ was es ima ed o each s udy as he a io be ween he median o all obse ed es s a is ics (b/SE)2 and he
expec ed median o a chi-squa ed wi h 1 deg ee o eedom, wi h b and SE ep esen ing he e ec o each SNP on
ln eGFRc ea o ln eGFRcys and i s s anda d e o , espec i ely. Genomic-con ol (GC) co ec ion42 was applied o
p- alues and SEs in case o λ > 1 (1s GC co ec ion). To limi he possibili y o alse posi i es, a second GC co -
ec ion on he agg ega ed esul s was applied a e he me a-analysis. Be ween-s udy he e ogenei y was assessed
wi h he I2 s a is ic43.
De ini ion o known and no el loci. Known loci we e de ined by a p e iously published lead a i-
an ha had shown genome-wide signi ican associa ion wi h eGFRc ea (p- alue < 5 × 10−8) and he gene ic
segmen a ound i (lead SNP ± 1 Mb)2–7. Va ian s ou side such segmen s and associa ed wi h eGFRc ea a a
p- alue < 5 × 10−8 in he 1000 Genomes me a-analysis de ined he no el loci. Each no el locus was pinpoin ed by
he lead a ian wi h he smalles p- alue ± 1 Mb.
Compa ison o 1000 Genomes and HapMap esul s. Fo he a ian s a ailable in bo h he 1000
Genomes and HapMap me a-analyses, we compa ed lead a ian s, e ec sizes, impu a ion quali y as well as he
powe ha we had in he da a o de ec he espec i e e ec s. Fo his compa ison, we also u ilized he associa ion
esul s o ou p e ious HapMap me a-analysis7 in 50 s udies including a maximum o 133,814 subjec s. Powe
was calcula ed in R (www. -p ojec .o g) o he app oxima e maximum numbe o subjec s in he 1000 Genomes
me a-analyses (n = 110,000) o iden i y he lead a ian s wi h an alpha o 5 × 10−8. Fu he , e ec i e powe , which
akes in o accoun he impu a ion quali y o he a ian , was calcula ed based on he e ec i e numbe o subjec s,
which is he numbe o subjec s pe a ian mul iplied by he median o he impu a ion quali y ac oss s udies.
Pa hway Analyses. Pa hway analyses, comp ised o pa hway/gene se en ichmen and issue/cell ype
analyses, we e pe o med by applying a so wa e package called Da a-D i en Exp ession P io i ized In eg a ion
o Complex T ai s (DEPICT)21. DEPICT pe o ms gene se en ichmen analyses by es ing whe he genes in
GWAS-associa ed loci a e en iched o econs i u ed e sions o known molecula pa hways (join ly e e ed o
as econs i u ed gene se s). The econs i u ion is accomplished by iden i ying genes ha a e co- egula ed wi h
o he genes in a gi en gene se based on a panel o 77,840 gene exp ession mic oa ays44. Genes ha a e ound o
be ansc ip ionally co- egula ed wi h genes om he o iginal gene se a e added o he gene se , which esul s in
he econs i u ion. Se e al ypes o gene se s we e econs i u ed in DEPICT: 5,984 p o ein molecula pa hways
de i ed om 169,810 high-con idence expe imen ally de i ed p o ein-p o ein in e ac ions45, 2,473 pheno ypic
gene se s de i ed om 211,882 gene-pheno ype pai s om he Mouse Gene ics Ini ia i e46, 737 Reac ome da a-
base pa hways47, 184 Kyo o Encyclopedia o Genes and Genomes (KEGG) da abase pa hways48 and 5,083 Gene
On ology da abase e ms49. In o al, 14,461 gene se s we e assessed o en ichmen in genes in associa ed egions.
DEPICT also acili a es issue and cell ype en ichmen analyses by es ing whe he he genes in associa ed egions
a e highly exp essed in any o he 209 MeSH anno a ions o 37,427 mic oa ays on he A yme ix U133 Plus 2.0
A ay pla o m.
In ou analysis, we used DEPICT e sion 1 el194 and o be compa able o ou p e ious analysis, included
all a ian s eaching eGFRc ea associa ion p- alues < 1 × 10−5 om HapMap and 1000 Genomes impu ed da a
wi h genomic coo dina es de ined by genome build GRCh38 (h ps://genome.ucsc.edu/cgi-bin/hgLi O e ).
www.na u e.com/scien i ic epo s/
8
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
Since 1000 Genomes impu ed loci in he DEPICT analysis di e ed sligh ly om he HapMap impu ed loci, ou
HapMap and 1000 Genomes inpu was c ea ed by adding all signi ican 1000 Genomes a ian s o all signi ican
HapMap a ian s. This p ocess esul ed in a o al o 3,659 a ian s o HapMap, 7,894 a ian s o 1000 Genomes,
and 9,270 a ian s o HapMap and 1000 Genomes analyses. Nex , independen lead a ian s we e iden i ied
wi h Plink50 using ± 500 kb lanking egions and 2 > 0.01 wi h he 1000 Genomes da a14 as e e ence. Genomic
in e als a e gene a ed consis ing o all a ian s wi hin 2 > 0.5 o each lead a ian . I any o he 19,987 genes in
he analysis o e laps o esides wi hin a genomic in e al, i is mapped o ha in e al. A e me ging o o e lap-
ping egions and excluding egions wi hin he majo his ocompa ibili y complex on ch omosome 6, base pai s
25,000,000–35,000,000, DEPICT analyses we e conduc ed using he ollowing pa ame e s: 200 epe i ions o
compu e FDR and 2,000 pe mu a ions o compu e p- alues adjus ed o gene leng h by using 500 null GWAS. Fo
he en ichmen analysis we used 10,968 econs i u ed gene se s. Fo isualiza ion, all no el signi ican gene se s
we e u he me ged in o me a gene se s by unning an a ini y p opaga ion51 om Py hons sciki -lea n package
(h p://sciki -lea n.o g/). The ne wo k was isualized wi h Cy oscape (h p://cy oscape.o g/).
Iden i ica ion o independen associa ion signals wi h GCTA. We sea ched o independen associ-
a ion signals in he known and no el loci wi h a join condi ional analysis on he agg ega ed me a-analysis esul s
using he GCTA-COJO me hod (condi ional and join genome-wide associa ion analysis)22,52. The KORA-F4
GWAS da a53 we e used o es ima e he LD ( 2) in he join condi ional analysis, and o quan i y he ex en o
coinhe i ance (D’)50. A po en ial independen associa ion signal wi hin a gi en locus was epo ed i he a ian
wi h he smalles condi ional p- alue was genome-wide signi ican (p- alue < 5 × 10−8) a e condi ioning on he
p e iously epo ed a ian in a locus.
SNP-based he i abili y analysis. The he i abili y o eGFRc ea was es ima ed using GCTA GREML-LDMS
me hods54 ( e sion 1.25) wi h impu ed geno ype accoun ing o linkage disequilib ium. The impu ed geno ype
was based on dosage (p obabili y > 0.9) impu ed using he 1000 Genomes Phase I e e ence panel and il e ed by
he ollowing c i e ia: HWE < 1 × 10−6, indi idual missingness > 5%, SNP missingness > 5%, and MAF < 0.0005
(~3 copies).
P opo ion o pheno ypic a iance explained. To quan i y he impac o he iden i ied gene ic loci on
enal unc ion, he pe cen o pheno ypic a iance explained by all lead a ian s in he no el and known loci was
es ima ed as ∗
Be a a a ian
a eGFRc ea
()
( esid(ln( )))
2
, whe e a ( a ian ) = 2 * MAF * (1− MAF)and be a is he es ima ed e ec o he
a ian in he 1000 Genomes me a-analysis55. The a iance o he esiduals o ln (eGFRc ea) is compu ed in he
ARIC s udy (n = 9,038). All a ian s we e assumed o ha e independen e ec s on he pheno ype.
Polygenic isk sco e analysis. P io i yP une (h p://p io i yp une .sou ce o ge.ne ) was used o selec
independen SNPs om The 1000 Genomes e e ence panel using an algo i hm ha p e e en ially selec s
SNPs ha a e mo e signi ican in he cu en 1000 Genomes me a-analysis compa ed o he p e ious HapMap
me a-analysis. Polygenic isk sco es (PRSs), using a ious h esholds o signi icance, as ob ained om he 1000
Genomes me a-analysis esul s and weigh ed o he e ec s sizes wi hin s udy we e gene a ed in TRAILS56
(n = 1,071), an independen s udy o adolescen s, which was no pa o he me a-analysis. These PRSs we e es ed
o associa ion wi h eGFRc ea using linea eg ession in R and he a iance explained by he PRSs was calcula ed.
Re e ences
1. Ecka d , K. U. e al. E ol ing impo ance o kidney disease: om subspecial y o global heal h bu den. Lance 382, 158–69 (2013).
2. Chambe s, J. C. e al. Gene ic loci in luencing kidney unc ion and ch onic kidney disease. Na Gene 42, 373–5 (2010).
3. Chasman, D. I. e al. In eg a ion o genome-wide associa ion s udies wi h biological knowledge iden i ies six no el genes ela ed o
kidney unc ion. Hum Mol Gene 21, 5329–43 (2012).
4. Ko gen, A. e al. Mul iple loci associa ed wi h indices o enal unc ion and ch onic kidney disease. Na Gene 41, 712–7 (2009).
5. Ko gen, A. e al. New loci associa ed wi h kidney unc ion and ch onic kidney disease. Na Gene 42, 376–84 (2010).
6. Pa a o, C. e al. Genome-wide associa ion and unc ional ollow-up e eals new loci o kidney unc ion. PLoS Gene 8, e1002584
(2012).
7. Pa a o, C. e al. Gene ic associa ions a 53 loci highligh cell ypes and biological pa hways ele an o kidney unc ion. Na
Commun 7, 10023 (2016).
8. In e na ional HapMap, C. e al. In eg a ing common and a e gene ic a ia ion in di e se human popula ions. Na u e 467, 52–8
(2010).
9. Boge , C. A. & Heid, I. M. Ch onic kidney disease: no el insigh s om genome-wide associa ion s udies. Kidney Blood P ess Res 34,
225–34 (2011).
10. Pa a o, C. e al. Genome-wide linkage analysis o se um c ea inine in h ee isola ed Eu opean popula ions. Kidney In 76, 297–306
(2009).
11. T udu, M. e al. Common noncoding UMOD gene a ian s induce sal -sensi i e hype ension and kidney damage by inc easing
u omodulin exp ession. Na Med 19, 1655–60 (2013).
12. Yeo, N. C. e al. Sh oom3 con ibu es o he main enance o he glome ula il a ion ba ie in eg i y. Genome Res 25, 57–65 (2015).
13. S einbjo nsson, G. e al. Ra e mu a ions associa ing wi h se um c ea inine and ch onic kidney disease. Hum Mol Gene 23, 6935–43
(2014).
14. Genomes P ojec , C. e al. An in eg a ed map o gene ic a ia ion om 1,092 human genomes. Na u e 491, 56–65 (2012).
15. Genomes P ojec , C. e al. A global e e ence o human gene ic a ia ion. Na u e 526, 68–74 (2015).
16. Wood, A. R. e al. Impu a ion o a ian s om he 1000 Genomes P ojec modes ly imp o es known associa ions and can iden i y
low- equency a ian -pheno ype associa ions unde ec ed by HapMap based impu a ion. PLoS One 8, e64343 (2013).
17. Nikpay, M. e al. A comp ehensi e 1,000 Genomes-based genome-wide associa ion me a-analysis o co ona y a e y disease. Na
Gene 47, 1121–30 (2015).
18. Li, Y., Wille , C. J., Ding, J., Schee , P. & Abecasis, G. R. MaCH: using sequence and geno ype da a o es ima e haplo ypes and
unobse ed geno ypes. Gene Epidemiol 34, 816–34 (2010).
www.na u e.com/scien i ic epo s/
9
Scien i ic RepoR s | 7:45040 | DOI: 10.1038/s ep45040
19. Ma chini, J., Howie, B., Mye s, S., McVean, G. & Donnelly, P. A new mul ipoin me hod o genome-wide associa ion s udies by
impu a ion o geno ypes. Na Gene 39, 906–13 (2007).
20. Ca oll, R. J. Measu emen e o in nonlinea models: a mode n pe spec i e. xx iii, 455 p. (Chapman & Hall/CRC, Boca Ra on, FL,
2006).
21. Pe s, T. H. e al. Biological in e p e a ion o genome-wide associa ion s udies using p edic ed gene unc ions. Na Commun 6, 5890
(2015).
22. Yang, J. e al. Condi ional and join mul iple-SNP analysis o GWAS summa y s a is ics iden i ies addi ional a ian s in luencing
complex ai s. Na Gene 44, 369–75, S1–3 (2012).
23. Pa a o, C. e al. Gene ic associa ions a 53 loci highligh cell ypes and biological pa hways ele an o kidney unc ion. Na
Commun 7, 10023 (2016).
24. Fox, C. S. e al. Genomewide linkage analysis o se um c ea inine, GFR, and c ea inine clea ance in a communi y-based popula ion:
he F amingham Hea S udy. J Am Soc Neph ol 15, 2457–61 (2004).
25. Wes a, H. J. e al. Sys ema ic iden i ica ion o ans eQTLs as pu a i e d i e s o known disease associa ions. Na Gene 45, 1238–43
(2013).
26. Boyle, A. P. e al. Anno a ion o unc ional a ia ion in pe sonal genomes using RegulomeDB. Genome Res 22, 1790–7 (2012).
27. Zheng, J. e al. LD Hub: a cen alized da abase and web in e ace o pe o m LD sco e eg ession ha maximizes he po en ial o
summa y le el GWAS da a o SNP he i abili y and gene ic co ela ion analysis. Bioin o ma ics 33, 272–279 (2017).
28. Eh e , G. B. e al. The gene ics o blood p essu e egula ion and i s a ge o gans om associa ion s udies in 342,415 indi iduals. Na
Gene 48, 1171–84 (2016).
29. Ho ikoshi, M. e al. Disco e y and Fine-Mapping o Glycaemic and Obesi y-Rela ed T ai Loci Using High-Densi y Impu a ion.
PLoS Gene 11, e1005230 (2015).
30. Vissche , P. M., B own, M. A., McCa hy, M. I. & Yang, J. Fi e yea s o GWAS disco e y. Am J Hum Gene 90, 7–24 (2012).
31. F i sche, L. G. e al. A la ge genome-wide associa ion s udy o age- ela ed macula degene a ion highligh s con ibu ions o a e and
common a ian s. Na Gene 48, 134–43 (2016).
32. Howie, B., Fuchsbe ge , C., S ephens, M., Ma chini, J. & Abecasis, G. R. Fas and accu a e geno ype impu a ion in genome-wide
associa ion s udies h ough p e-phasing. Na Gene 44, 955–9 (2012).
33. Fuchsbe ge , C., Abecasis, G. R. & Hinds, D. A. minimac2: as e geno ype impu a ion. Bioin o ma ics 31, 782–4 (2015).
34. Co esh, J. e al. Calib a ion and andom a ia ion o he se um c ea inine assay as c i ical elemen s o using equa ions o es ima e
glome ula il a ion a e. Am J Kidney Dis 39, 920–9 (2002).
35. Fox, C. S. e al. P edic o s o new-onse kidney disease in a communi y-based popula ion. JAMA 291, 844–50 (2004).
36. Le ey, A. S. e al. A mo e accu a e me hod o es ima e glome ula il a ion a e om se um c ea inine: a new p edic ion equa ion.
Modi ica ion o Die in Renal Disease S udy G oup. Ann In e n Med 130, 461–70 (1999).
37. Le ey, A. S. e al. Using s anda dized se um c ea inine alues in he modi ica ion o die in enal disease s udy equa ion o
es ima ing glome ula il a ion a e. Ann In e n Med 145, 247–54 (2006).
38. S e ens, L. A. e al. Es ima ing GFR using se um cys a in C alone and in combina ion wi h se um c ea inine: a pooled analysis o
3,418 indi iduals wi h CKD. Am J Kidney Dis 51, 395–406 (2008).
39. Po cu, E., Sanna, S., Fuchsbe ge , C. & F i sche, L. G. Geno ype impu a ion in genome-wide associa ion s udies. Cu P o oc Hum
Gene Chap e 1, Uni 1 25 (2013).
40. Fuchsbe ge , C., Taliun, D., P ams alle , P. P., Pa a o, C. & conso ium, C. K. GWA oolbox: an R package o as quali y con ol and
handling o genome-wide associa ion s udies me a-analysis da a. Bioin o ma ics 28, 444–5 (2012).
41. Wille , C. J., Li, Y. & Abecasis, G. R. METAL: as and e icien me a-analysis o genomewide associa ion scans. Bioin o ma ics 26,
2190–1 (2010).
42. De lin, B. & Roede , K. Genomic con ol o associa ion s udies. Biome ics 55, 997–1004 (1999).
43. Higgins, J. P., Thompson, S. G., Deeks, J. J. & Al man, D. G. Measu ing inconsis ency in me a-analyses. BMJ 327, 557–60 (2003).
44. Feh mann, R. S. e al. Gene exp ession analysis iden i ies global gene dosage sensi i i y in cance . Na Gene 47, 115–25 (2015).
45. Lage, K. e al. A human phenome-in e ac ome ne wo k o p o ein complexes implica ed in gene ic diso de s. Na Bio echnol 25,
309–16 (2007).
46. Blake, J. A. e al. The Mouse Genome Da abase: in eg a ion o and access o knowledge abou he labo a o y mouse. Nucleic Acids
Res 42, D810–7 (2014).
47. C o , D. e al. Reac ome: a da abase o eac ions, pa hways and biological p ocesses. Nucleic Acids Res 39, D691–7 (2011).
48. Kanehisa, M., Go o, S., Sa o, Y., Fu umichi, M. & Tanabe, M. KEGG o in eg a ion and in e p e a ion o la ge-scale molecula da a
se s. Nucleic Acids Res 40, D109–14 (2012).
49. Ashbu ne , M. e al. Gene on ology: ool o he uni ica ion o biology. The Gene On ology Conso ium. Na Gene 25, 25–9 (2000).
50. Pu cell, S. e al. PLINK: a ool se o whole-genome associa ion and popula ion-based linkage analyses. Am J Hum Gene 81, 559–75
(2007).
51. F ey, B. J. & Dueck, D. Clus e ing by passing messages be ween da a poin s. Science 315, 972–6 (2007).
52. W igh , A. K. & Thompson, M. R. Hyd odynamic s uc u e o bo ine se um albumin de e mined by ansien elec ic bi e ingence.
Biophys J 15, 137–41 (1975).
53. Wichmann, H. E., Giege , C., Illig, T. & G oup, M. K. S. KORA-gen– esou ce o popula ion gene ics, con ols and a b oad spec um
o disease pheno ypes. Gesundhei swesen 67 Suppl 1, S26–30 (2005).
54. Yang, J. e al. Gene ic a iance es ima ion wi h impu ed a ian s inds negligible missing he i abili y o human heigh and body
mass index. Na Gene 47, 1114–20 (2015).
55. Rosne , B. Fundamen als o bios a is ics. x ii, 859 p. (B ooks/Cole, Cengage Lea ning, Bos on, 2011).
56. Huisman, M. e al. Coho p o ile: he Du ch ‘TRacking Adolescen s’ Indi idual Li es’ Su ey’; TRAILS. In J Epidemiol 37, 1227–35
(2008).
Acknowledgemen s
S udy speci ic acknowledgemen s and unding sou ces o pa icipa ing s udies a e epo ed in he supplemen .
Au ho Con ibu ions
S udy Design E. Bo inge , J. Co esh, G.C. Cu han, J. Ding, V. Gudnason, C. Helme , A. Ho man, M. Kähönen,
B.K. K äme , T. Leh imäki, Y. Liu, A. Me spalu, P.P. P ams alle , N. P obs -Hensch, O.T. Rai aka i, H. Schmid ,
R. Schmid , B. S engel, D. Toniolo, J.S. Viika i, P. Vollenweide , H. Völzke, A. Kö gen. S udy Managemen R.
Bi a , E. Boe winkle, E. Bo inge , J. Co esh, M.C. Co nelis, A. Dehghan, L. Dengle , G. Ei iksdo i , T. Esko,
O. F anco, V. Gudnason, S.J. Hancock, A. Ho man, N. Hu i-Kähönen, M. Imboden, M. Kähönen, W. König, H.
K ame , B.K. K äme , T. Leh imäki, R.J.F. Loos, M. Nauck, P.P. P ams alle , N. P obs -Hensch, O.T. Rai aka i, R.
Re ig, P.M. Ridke , H. Schmid , R. Schmid , D. Toniolo, J.S. Viika i, P. Vollenweide , D. Chasman, C. Fox. Subjec
Rec ui men E. Bo inge , M. Ciullo, J. Co esh, A.P. d’Adamo, G. Ei iksdo i , K. Endlich, P. Gaspa ini, G. Gi o o,