Meta-analysis of Genome-Wide Association Studies for Extraversion: Findings from the Genetics of Personality Consortium

ORIGINAL RESEARCH Me a-analysis o Genome-Wide Associa ion S udies o Ex a e sion: Findings om he Gene ics o Pe sonali y Conso ium S e ´phanie M. an den Be g 1 •Ma leen H. M. de Moo 2,3,4 •Ka in J. H. Ve weij 5,6 • Robe F. K uege 7 •Michelle Luciano 8,9 •Alejand o A ias Vasquez 10,11,12,13 • Lindsay K. Ma eson 7 •Jaime De inge 14 •To ˜nu Esko 15 •Naja Amin 16 • Sco D. Go don 5 •Na elle K. Hansell 5 •Amy B. Ha 17 •Ilkka Seppa ¨la ¨ 18 • Jenni e E. Hu man 19 •Be ina Kon e 20 •Ja i Lah i 21,22 •Minyoung Lee 23 • Mike Mille 7 •Te esa Nu ile 24 •Toshiko Tanaka 25 •Alexande Teume 26 • Alexande Vik o in 27 •Juho Wedenoja 28 •Abdel Abdellaoui 2 •Goncalo R. Abecasis 29 • Daniel E. Adkins 30 •A pana Ag awal 31 •Ju ¨ i Allik 32,33 •Ka ja Appel 34 • Timo hy B. Bigdeli 23 •Fabio Busone o 35 •Ha y Campbell 36 •Paul T. Cos a 37 • Geo ge Da ey Smi h 38 •Gail Da ies 8,9 •Ha ie de Wi 39 •Jun Ding 67 • Ba ba a E. Engelha d 40 •Johan G. E iksson 22,41,42,43,44 •I yna O. Fedko 2 • Luigi Fe ucci 25 •Ba ba a F anke 10,11,12 •Ina Giegling 20 •Richa d G ucza 31 • Anne e M. Ha mann 20 •And ew C. Hea h 31 •Ka i Heinonen 21 •Anjali K. Hende s 5 • Geo g Homu h 45 •Jouke-Jan Ho enga 2 •William G. Iacono 7 •Joos Janzing 11 • Ma kus Jokela 21 •Robe Ka lsson 27 •John P. Kemp 38,46 •Ma hew G. Ki kpa ick 39 • An i La ala 41,28 •Te ho Leh ima ¨ki 18 •Da id C. Liewald 8,9 •Pamela A. F. Madden 31 • Chia a Mag i 47 •Pa ik K. E. Magnusson 27 •Jona han Ma en 19 •And ea Maschio 35 • Hamdi Mba ek 2 •Sa ah E. Medland 5 •E elin Mihailo 15,48 •Yu i Milaneschi 49 • Edi ed by S acey Che ny. S e ´phanie M. an den Be g and Ma leen H. M. de Moo sha ed i s au ho ship. Elec onic supplemen a y ma e ial The online e sion o his a icle (doi:10.1007/s10519-015-9735-5) con ains supplemen a y ma e ial, which is a ailable o au ho ized use s. &S e ´phanie M. an den Be g [email p o ec ed] 1 Depa men o Resea ch Me hodology, Measu emen and Da a-Analysis (OMD), Facul y o Beha iou al, Managemen , and Social Sciences, Uni e si y o Twen e, PO Box 217, 7500 AE Enschede, The Ne he lands 2 Depa men o Biological Psychology, VU Uni e si y Ams e dam, Ams e dam, The Ne he lands 3 Depa men o Clinical Child and Family S udies, VU Uni e si y Ams e dam, Ams e dam, The Ne he lands 4 Depa men o Me hods, VU Uni e si y Ams e dam, Ams e dam, The Ne he lands 5 QIMR Be gho e Medical Resea ch Ins i u e, B isbane, Aus alia 6 Depa men o De elopmen al Psychology and EMGO Ins i u e o Heal h and Ca e Resea ch, VU Uni e si y Ams e dam, Ams e dam, The Ne he lands 7 Depa men o Psychology, Uni e si y o Minneso a, Minneapolis, USA 8 Depa men o Psychology, Uni e si y o Edinbu gh, Edinbu gh, UK 9 Cen e o Cogni i e Ageing and Cogni i e Epidemiology, Uni e si y o Edinbu gh, Edinbu gh, UK 10 Donde s Ins i u e o Cogni i e Neu oscience, Radboud Uni e si y Nijmegen, Nijmegen, The Ne he lands 11 Depa men o Psychia y, Radboud Uni e si y Nijmegen Medical Cen e , Nijmegen, The Ne he lands 12 Depa men o Human Gene ics, Radboud Uni e si y Nijmegen Medical Cen e , Nijmegen, The Ne he lands 13 Depa men o Cogni i e Neu oscience, Radboud Uni e si y Nijmegen Medical Cen e , Nijmegen, The Ne he lands 14 Depa men o Psychology, Uni e si y o Illinois a U bana- Champaign, Champaign, IL, USA 15 Es onian Genome Cen e , Uni e si y o Ta u, Ta u, Es onia 123 Beha Gene (2016) 46:170–182 DOI 10.1007/s10519-015-9735-5 G an W. Mon gome y 5 •Ma hias Nauck 50 •Michel G. Ni a d 2 •Klaasjan G. Ouwens 2 • Aa no Palo ie 51,52 •E ik Pe e sson 27 •Oz en Polasek 53 •Yong Qian 67 • Lau a Pulkki-Ra ˚back 21 •Olli T. Rai aka i 54,55 •Anu Realo 32 •Richa d J. Rose 56 • Daniela Ruggie o 24 •Ca s en O. Schmid 26 •Wendy S. Slu ske 57 •Rossella So ice 24 • John M. S a 9,58 •Bea e S Pou cain 38,59,60 •Angelina R. Su in 25,61 • Nicholas J. Timpson 38 •Holly T oche 19 •Si a Ve meulen 12,62 •Ee o Vuoksimaa 28 • Elisabe h Widen 52 •Jaspe Wouda 1,2 •Ma ga e J. W igh 5 •Lina Zgaga 36,63 • Gene a ion Sco land 64 •Da id Po eous 65 •Alessand a Minelli 47 • Ab aham A. Palme 17,39 •Dan Rujescu 20 •Ma ina Ciullo 24 •Ca oline Haywa d 19 • Igo Rudan 36 •And es Me spalu 15,33 •Jaakko Kap io 41,28,52 •Ian J. Dea y 8,9 • Ka i Ra ¨ikko ¨nen 21 •James F. Wilson 19,36 •Liisa Kel ikangas-Ja ¨ inen 21 • Lau a J. Bie u 31 •John M. He ema 23 •Hans J. G abe 34,66 •B enda W. J. H. Penninx 49 • Co nelia M. an Duijn 16 •Da id M. E ans 38 •Da id Schlessinge 67 • Nancy L. Pede sen 27 •An onio Te acciano 22,25 •Ma McGue 7,68 • Nicholas G. Ma in 5 •Do e I. Boomsma 2 Recei ed: 30 Oc obe 2014 / Accep ed: 10 Augus 2015 / Published online: 11 Sep embe 2015 ÓThe Au ho (s) 2015. This a icle is published wi h open access a Sp inge link.com Abs ac Ex a e sion is a ela i ely s able and he i able pe sonali y ai associa ed wi h nume ous psychosocial, li es yle and heal h ou comes. Despi e i s subs an ial he i- abili y, no gene ic a ian s ha e been de ec ed in p e ious genome-wide associa ion (GWA) s udies, which may be due o ela i ely small sample sizes o hose s udies. He e, we epo on a la ge me a-analysis o GWA s udies o ex a e sion in 63,030 subjec s in 29 coho s. Ex a e sion i em da a om mul iple pe sonali y in en o ies we e ha - monized ac oss in en o ies and coho s. No genome-wide signi ican associa ions we e ound a he single nucleo ide polymo phism (SNP) le el bu he e was one signi ican hi a he gene le el o a long non-coding RNA si e (LOC101928162). Genome-wide complex ai analysis in wo la ge coho s showed ha he addi i e a iance explained by common SNPs was no signi ican ly di e en 16 Depa men o Epidemiology, E asmus Uni e si y Medical Cen e , Ro e dam, The Ne he lands 17 Depa men o Human Gene ics, Uni e si y o Chicago, Chicago, IL, USA 18 Depa men o Clinical Chemis y, Fimlab Labo a o ies and School o Medicine, Uni e si y o Tampe e, Tampe e, Finland 19 MRC Human Gene ics Uni , MRC IGMM, Wes e n Gene al Hospi al, Uni e si y o Edinbu gh, Edinbu gh, UK 20 Depa men o Psychia y, Uni e si y o Halle, Halle, Ge many 21 Ins i u e o Beha iou al Sciences, Uni e si y o Helsinki, Helsinki, Finland 22 Folkha ¨lsan Resea ch Cen e , Helsinki, Finland 23 Depa men o Psychia y, Vi ginia Ins i u e o Psychia ic and Beha io al Gene ics, Vi ginia Commonweal h Uni e si y, Richmond, VA, USA 24 Ins i u e o Gene ics and Biophysics ‘‘A. Buzza i-T a e so’’ – CNR, Naples, I aly 25 Na ional Ins i u e on Aging, NIH, Bal imo e, MD, USA 26 Ins i u e o Communi y Medicine, Uni e si y Medicine G ei swald, G ei swald, Ge many 27 Depa men o Medical Epidemiology and Bios a is ics, Ka olinska Ins i u e , S ockholm, Sweden 28 Depa men o Public Heal h, Uni e si y o Helsinki, Helsinki, Finland 29 Depa men o Bios a is ics, Cen e o S a is ical Gene ics, Uni e si y o Michigan School o Public Heal h, Ann A bo , MI, USA 30 Pha maco he apy & Ou comes Science, Vi ginia Commonweal h Uni e si y, Richmond, VA, USA 31 Depa men o Psychia y, Washing on Uni e si y School o Medicine, S . Louis, MO, USA 32 Depa men o Psychology, Uni e si y o Ta u, Ta u, Es onia 33 Es onian Academy o Sciences, Tallinn, Es onia 34 Depa men o Psychia y and Psycho he apy, Uni e si y Medicine G ei swald, G ei swald, Ge many 35 Is i u o di Rice ca Gene ica e Biomedica (IRGB), CNR, Monse a o, I aly 36 Ushe Ins i u e o Popula ion Heal h Sciences and In o ma ics, Uni e si y o Edinbu gh, Edinbu gh, UK 37 Beha io al Medicine Resea ch Cen e , Duke Uni e si y School o Medicine, Du ham, NC, USA 38 Medical Resea ch Council In eg a i e Epidemiology Uni , School o Social and Communi y Medicine, Uni e si y o B is ol, B is ol, UK Beha Gene (2016) 46:170–182 171 123 om ze o, bu polygenic isk sco es, weigh ed using link- age in o ma ion, signi ican ly p edic ed ex a e sion sco es in an independen coho . These esul s show ha ex a e sion is a highly polygenic pe sonali y ai , wi h an a chi ec u e possibly di e en om o he complex human ai s, including o he pe sonali y ai s. Fu u e s udies a e equi ed o u he de e mine which gene ic a ian s, by wha modes o gene ac ion, cons i u e he he i able na u e o ex a e sion. Keywo ds Pe sonali y Pheno ype ha moniza ion  Common gene ic a ian s Impu a ion Polygenic isk In oduc ion Ex a e sion is a pe sonali y ai cha ac e ized by he endency o expe ience posi i e emo ions, o be ac i e and eel ene ge ic, o be alka i e and o enjoy social in e ac- ions. Ex a e sion is associa ed wi h nume ous psy- chosocial, li es yle and heal h ou comes, such as academic and job pe o mance, well-being, obesi y, subs ance use, physical ac i i y, bipola diso de , bo de line pe sonali y diso de , Alzheime ’s disease, and longe i y (De Moo e al. 2006,2011; Dis el e al. 2009a; Fu nham e al. 2013; Judge e al. 2013; Middeldo p e al. 2011; Rhodes and Smi h 2006; Su in e al. 2011; Te acciano e al. 2008; Te acciano e al. 2014; Weiss e al. 2008). Ex a e sion can be measu ed wi h mul iple in en o ies ha ha e been de eloped as pa o di e en pe sonali y heo ies. Fo example, ex a e sion is one o he i e pe - sonali y domains as assessed wi h he Neu o icism–Ex- a e sion–Openness o Expe ience (NEO) pe sonali y in en o ies (Cos a and McC ae 1992). Ex a e sion is also included in Eysenck’s h ee-dimensional heo y o pe - sonali y (Eysenck and Eysenck 1964,1975; Eysenck e al. 1985). In Cloninge ’s heo y on empe amen s and cha - ac e s (Cloninge 1987; Cloninge e al. 1993), Ha m A oidance, No el y Seeking and Rewa d Dependence a e ela ed o ex a e sion (De F uy e al. 2000). Tellegen’s pe sonali y heo y posi s he highe o de domain o Posi- i e Emo ionali y (Pa ick e al. 2002), which esembles and is highly co ela ed wi h ex a e sion (Chu ch 1994). We showed ecen ly, by pe o ming an I em Response Theo y (IRT) analysis using es linking (Kolen and B ennan 2004), ha i em da a on Ex a e sion, Rewa d dependence and Posi i e Emo ionali y can be ha monized o b oadly assess he same unde lying ex a e sion con- s uc ( an den Be g e al. 2014). This ha moniza ion was pe o med in o e 160,000 indi iduals om 23 coho s pa icipa ing in he Gene ics o Pe sonali y Conso ium (GPC). B ie ly, ha moniza ion was ca ied ou in each coho sepa a ely by i s i ing an IRT model o da a om 39 Depa men o Psychia y and Beha io al Neu oscience, Uni e si y o Chicago, Chicago, USA 40 Depa men o Bios a is ics and Bioin o ma ics, Duke Uni e si y, Du ham, NC, USA 41 Na ional Ins i u e o Heal h and Wel a e (THL), Helsinki, Finland 42 Depa men o Gene al P ac ice and P ima y Heal h Ca e, Uni e si y o Helsinki, Helsinki, Finland 43 Uni o Gene al P ac ice and P ima y Heal h Ca e, Uni e si y o Helsinki, Helsinki, Finland 44 Vasa Cen al Hospi al, Vaasa, Finland 45 In e acul y Ins i u e o Gene ics and Func ional Genomics, Uni e si y o G ei swald, G ei swald, Ge many 46 T ansla ional Resea ch Ins i u e, Uni e si y o Queensland Diaman ina Ins i u e, B isbane, Aus alia 47 Depa men o Molecula and T ansla ional Medicine, Uni e si y o B escia, B escia, I aly 48 Depa men o Bio echnology, Uni e si y o Ta u, Ta u, Es onia 49 Depa men o Psychia y, EMGO?Ins i u e, Neu oscience Campus Ams e dam, VU Uni e si y Medical Cen e , Ams e dam, The Ne he lands 50 Ins i u e o Clinical Chemis y and Labo a o y Medicine, Uni e si y Medicine G ei swald, G ei swald, Ge many 51 Wellcome T us Sange Ins i u e, Wellcome T us Genome Campus, Hinx on, Camb idge, UK 52 Ins i u e o Molecula Medicine Finland (FIMM), Uni e si y o Helsinki, Helsinki, Finland 53 Depa men o Public Heal h, Facul y o Medicine, Uni e si y o Spli , Spli , C oa ia 54 Depa men o Clinical Physiology and Nuclea Medicine, Tu ku Uni e si y Hospi al, Tu ku, Finland 55 Resea ch Cen e o Applied and P e en i e Ca dio ascula Medicine, Uni e si y o Tu ku, Tu ku, Finland 56 Depa men o Psychological & B ain Sciences, Indiana Uni e si y, Blooming on, IN, USA 57 Depa men o Psychological Sciences and Missou i Alcoholism Resea ch Cen e , Uni e si y o Missou i, Columbia, MO, USA 58 Alzheime Sco land Demen ia Resea ch Cen e, Uni e si y o Edinbu gh, Edinbu gh, UK 59 School o O al and Den al Sciences, Uni e si y o B is ol, B is ol, UK 60 School o Expe imen al Psychology, Uni e si y o B is ol, B is ol, UK 61 College o Medicine, Flo ida S a e Uni e si y, Tallahassee, FL, USA 172 Beha Gene (2016) 46:170–182 123 indi iduals who had comple ed a leas wo di e en pe - sonali y ques ionnai es. Nex , based on calib a ed i em pa ame e s, pe sonali y sco es we e es ima ed based on all a ailable da a o each indi idual, i espec i e o wha pe sonali y ques ionnai e was used. The ha monized ex a e sion pheno ype was he i able. A b oad-sense he i- abili y o 49 % was es ima ed, based on a me a-analysis in six win coho s ha a e included in he GPC (29,501 win pai s), o which 24 % was due o addi i e gene ic a iance and 25 % due o non-addi i e gene ic a iance. The b oad- sense he i abili y es ima e is simila o he i abili y es i- ma es ob ained o ex a e sion as assessed wi h single measu emen ins umen s (Boucha d and Loehlin 2001; Dis el e al. 2009b; Finkel and McGue 1997; Kelle e al. 2005; Re ew e al. 2008; Yamaga a e al. 2006). Some e idence o quali a i e sex di e ences in he gene ic in luences on ex a e sion was sugges ed by a gene ic co ela ion in opposi e-sex win pai s o 0.38 ( an den Be g e al. 2014). Ex a e sion becomes mo e gene ically s able du ing adolescence un il i is almos pe ec ly gene ically s able in adul hood (B iley and Tucke -D ob 2014; Kandle 2012), ha is, he same genes a e esponsible o ex a e sion measu ed a di e en ages. A hand ul o genome-wide associa ion (GWA) s udies o ex a e sion ha e been published, aimed a de ec ing speci ic single nucleo ide polymo phisms (SNPs) ha explain pa o he he i abili y. The i s GWA s udy o pe sonali y, which ocused on he i e NEO pe sonali y ai s, was conduc ed in 3972 adul s (Te acciano e al. 2010). No genome-wide signi ican SNP associa ions we e ound o ex a e sion, al hough some in e es ing associa- ions wi h P- alues 10 -5 we e seen wi h SNPs in wo cadhe in genes and he b ain-de i ed neu o ophic ac o (BDNF) gene. A subsequen me a-analysis o GWA esul s o he NEO pe sonali y ai s, conduc ed in 17,375 subjec s, also did no yield any genome-wide signi ican associa ions o ex a e sion (De Moo e al. 2012). Two o he GWA s udies epo ed a simila lack o genome-wide signi icance o Cloninge ’s empe amen scales (Se ice e al. 2012; Ve weij e al. 2010). In e es ingly, a s udy ha pe o med a gene ic complex ai analysis (GCTA; Yang e al. 2010) o neu o icism and ex a e sion in a ound 12,000 un ela ed indi iduals epo ed ha 12 % (SE =3 %) o he a iance in ex a e sion was explained by common SNPs o addi i e e ec (Vinkhuyzen e al. 2012). Taken oge he , he esul s om win and genome-wide s udies sugges ha common SNPs o addi i e e ec a e impo an , ha gene ic non-ad- di i i y may play a ole, and ha la ge sample sizes a e likely o be equi ed o iden i y speci ic a ian s. In his pape , we epo he esul s o he la ges me a- analysis o GWA esul s o ex a e sion so a , ca ied ou in 29 coho s ha pa icipa e in he GPC. A o al o 63,030 subjec s wi h ha monized ex a e sion and genome-wide geno ype da a we e included in he me a-analysis. A 30 h coho was used o eplica ion. In his conso ium we epo ed ea lie on a genome-wide signi ican hi o neu- o icism (De Moo e al. 2015), indica ing ha we may begin o analyze da a om su icien ly la ge samples, o ob ain he i s signi ican indings om GWA s udies o pe sonali y. In addi ion o me a-analysis o GWA esul s, we compu ed weigh ed polygenic sco es in an independen coho and associa ed hem wi h ex a e sion, and es i- ma ed a iance explained by SNPs in wo la ge coho s. Ma e ials and me hods Coho s The ull me a-analysis was pe o med on 63,030 subjec s om 29 disco e y coho s. All samples we e o Eu opean o igin. Twen y-one coho s we e om Eu ope, six om he Uni ed S a es and wo om Aus alia. Sample sizes o he indi idual coho s anged om 177 o 7210 subjec s. Please no e ha some coho s we e also pa o p e iously published GWA s udies on ex a e sion. The Gene a ion Sco land: Sco ish Family Heal h S udy (GS:SFHS) coho was included as a eplica ion sample (9,783 subjec s). A b ie o e iew o all coho s is p o ided in Table 1. A desc ip ion o each indi idual coho is ound in he Supplemen a y ma e ials and me hods (see also De Moo e al. 2015). Pheno yping A ha monized la en ex a e sion sco e was es ima ed o all pa icipan s in all 29 coho s ha we e included in he GWA me a-analysis. This sco e was based on all a ailable ex a e sion i em da a o each indi idual ( o a de ailed 62 Depa men o Heal h E idence, Radboud Uni e si y Medical Cen e , Nijmegen, The Ne he lands 63 Depa men o Public Heal h and P ima y Ca e, T ini y College Dublin, Dublin, I eland 64 Sco ish Family Heal h S udy, A Collabo a ion Be ween he Uni e si y Medical Schools and NHS, Abe deen, Dundee, Edinbu gh and Glasgow, UK 65 Medical Gene ics Sec ion, Cen e o Genomics and Expe imen al Medicine, Ins i u e o Gene ics and Molecula Medicine, Wes e n Gene al Hospi al, The Uni e si y o Edinbu gh, Edinbu gh, UK 66 Depa men o Psychia y and Psycho he apy, HELIOS Hospi al S alsund, S alsund, Ge many 67 Labo a o y o Gene ics, Na ional Ins i u e on Aging, Na ional Ins i u es o Heal h, Bal imo e, MD, USA 68 Ins i u e o Public Heal h, Uni e si y o Sou he n Denma k, Odense, Denma k Beha Gene (2016) 46:170–182 173 123 desc ip ion see an den Be g e al. 2014). Ex a e sion i em da a came om he ex a e sion scales o he NEO Pe sonali y In en o y, he NEO Fi e Fac o In en o y, he 50-i em Big-Fi e e sion o he In e na ional Pe sonali y I em Pool in en o y, he Eysenck Pe sonali y Ques ion- nai e and he Eysenck Pe sonali y In en o y, om he Rewa d Dependence scale o he Cloninge ’s T idimen- sional Pe sonali y Ques ionnai e, and om he Posi i e Emo ionali y scale o he Mul idimensional Pe sonali y Ques ionnai e (see an den Be g e al. 2014 and Supple- men a y ma e ials and me hods). In he GS:SFHS coho ha was included o eplica ion o op signals, ex a e - sion was based on he summed sco e o he ex a e sion scale o he EPQ Re ised Sho Fo m. Geno yping and impu a ion Geno yping in all coho s was ca ied ou on Illumina o A yme ix pla o ms, a e which quali y con ol (QC) was pe o med, ollowed by impu a ion o geno ypes. QC o geno ype da a was pe o med in each coho sepa a ely, wi h compa able bu coho speci ic c i e ia. S anda d QC checks included es s o Eu opean ances y, sex inconsis- encies, Mendelian e o s, and high genome-wide homozygosi y. Checks o ela edness we e conduc ed in hose coho s ha aimed o include un ela ed indi iduals only. O he checks o geno ype da a we e based on mino allele equencies (MAF), SNP call a e (% o subjec s wi h missing geno ypes pe SNP), sample call a e (% o miss- ing SNPs pe subjec ) and Ha dy–Weinbe g Equilib ium (HWE). Geno ype da a we e impu ed using he 1000Gen- omes phase 1 e sion 3 (build37, hg19) e e ence panel wi h s anda d so wa e packages such as IMPUTE, MACH, o Minimac, see Supplemen a y Table 1. S a is ical analyses GWA analysis pe coho GWA analyses we e conduc ed independen ly in each coho . Since he coho s used di e en esea ch designs (case–con ol, popula ion win s udies, ex ended pedig ees, e c.), GWA me hods we e op imized o each coho . Ex a e sion sco es we e eg essed on each SNP unde an addi i e model, wi h sex and age included as co a ia es. Co a ia es such as ances y P incipal Componen s (PCs) we e added i deemed necessa y o a pa icula coho . In all analyses, he unce ain y o he impu ed geno ypes was aken in o accoun , ei he using dosage sco es o mix u es o dis ibu ions. In hose coho s ha included ela ed indi iduals, he dependency among pa icipan s was accoun ed o using coho -speci ic me hods. S anda d so wa e packages o GWA analyses we e used (see Supplemen a y Table 1). Me a-analysis o GWA esul s ac oss coho s A me a-analysis o he GWA esul s was conduc ed wi h he weigh ed in e se a iance me hod in METAL (h p://www.sph.umich.edu/csg/abecasis/me al/index.h ml). Excluded om me a-analysis we e poo ly impu ed SNPs ( 2 0.30 o p ope _in o 0.40) and SNPs wi h low Table 1 O e iew o 29 disco e y coho s and 1 eplica ion coho o he Gene ics o Pe sonali y Conso ium Coho # Subjec s a # SNPs b 1 ALSPAC 4705 6,454,153 2 BLSA 820 4,989,411 3 BRESCIA 177 3,549,919 4 CHICAGO 311 3,755,416 5 CILENTO 627 1,123,089 6 COGA 647 5,127,101 7 COGEND 1279 5,932,838 8 EGCUT 1184 5,574,695 9 ERF 2300 5,142,865 10 FTC EPI 567 4,870,096 11 FTC NEO 813 5,092,018 12 HBCS 1456 5,612,790 13 CROATIA-Ko cula 808 5,094,034 14 LBC1921 437 4,363,611 15 LBC1936 952 5,168,754 16 MCTFR 7099 6,569,999 17 MGS 2101 5,900,898 18 NBS 1832 5,603,447 19 NESDA 2227 4,707,569 20 NTR 6416 5,339,798 21 ORCADES 1650 4,265,590 22 PAGES 476 4,547,293 23 QIMR adolescen s 2842 5,957,064 24 QIMR adul s 7210 6,343,920 25 Sa diNIA 4332 6,291,135 26 SHIP 2213 5,913,428 27 STR 4903 6,519,094 28 CROATIA-Vis 909 5,327,671 29 YFS 1737 5,914,679 To al 63,030 7,460,147 30 GS:SFHS 9783 74 NA No Applicable o eplica ion coho because only op hi s we e sough o eplica e a Numbe o subjec s wi h alid la en sco e o Ex a e sion and SNP da a (a e impu a ion and cleaning) b Numbe o SNPs (a e impu a ion and cleaning) wi h alid asso- cia ion esul s ha en e ed he me a-analysis 174 Beha Gene (2016) 46:170–182 123 MAF (MAF H(5/N), which co esponds o less han 5 es ima ed indi iduals in he leas equen geno ype g oup, unde he assump ion o HWE). This esul ed in a o al numbe o 7,460,147 unique SNPs in he inal me a-anal- ysis (wi h 1.1–6.6 M SNPs ac oss coho s). Fo 2182 SNPs, SNP loca ions could no be ma ched wi h s names. Fo an addi ional 516,362 SNPS, esul s we e based on one coho only and he e o e le ou o he analysis, so ha he esul s a e based on 6,941,603 SNPs. Genomic con ol in la ion ac o s (lambda), Manha an plo s and quan ile–quan ile plo s pe coho a e p o ided in Supplemen a y Table 2 and Supplemen a y Figs. 1, 2. A P alue o 5 910 -8 was used as he h eshold o genome-wide signi icance. The me a-analysis esul s (P- alues pe SNP) we e used as he inpu o compu e P- alues a he gene le el. We pe o med hese analyses in KGG (Li e al. 2012). A P- alue o 2.87 910 -6 was used as he h eshold o gen- ome-wide signi icance in hese gene-wide analyses, based on con olling o he alse-disco e y a e (Benjamini and Hochbe g 1995). All GWAS SNP op hi s wi h a P- alue smalle han 1910 -5 we e selec ed o eplica ion in he GS:SFHS coho . Polygenic isk sco e analysis Addi ional analyses we e conduc ed o es whe he ex a e sion could be p edic ed in an independen a ge coho based on he GWA me a-analysis esul s. The a ge coho was he Ne he lands Twin Regis e (NTR) coho (8648 subjec s). Polygenic isk sco es o his coho we e es ima ed using LDp ed (Vilhjalmsson e al. 2015) ha akes in o accoun linkage disequilib ium among he SNPs. The es ima ion was based on a GWA me a-analysis in which he NTR and NESDA coho s we e excluded ( u - he e e ed o as he disco e y se ). Wi h he LD-co - ec ed polygenic isk sco es, gene alized es ima ing equa ion (GEE) modeling was applied o es whe he he polygenic isk sco es p edic ed ex a e sion in he a ge coho . The co a ia es age, sex and en PCs we e included as ixed e ec s in he model. The model also included a andom in e cep wi h amily numbe as he clus e a i- able, o accoun o dependency among amily membe s. Ou lie s on he PCs, including e hnic ou lie s, we e excluded om he analysis. Va iance explained by SNPs In he NTR coho and he QIMR Be gho e Medical Resea ch Ins i u e (QIMR) adul coho (see also Supple- men a y ma e ials and me hods), GCTA so wa e (Vissche e al. 2010; Yang e al. 2010) was used o es ima e he p opo ion o a iance in ex a e sion ha can be explained by common SNPs o addi i e e ec . In he NTR, his analysis was ca ied ou in a se o 3597 un ela ed indi- iduals and in he QIMR adul coho his was done in 3369 un ela ed indi iduals (in each coho one membe pe amily was selec ed wi h ha monized ex a e sion and genome-wide SNP da a). GCTA analysis was based on bes guess geno ypes ob ained in PLINK using a h eshold o a maximum geno ype p obabili y [0.70, and addi ionally il e ing on -squa ed [0.80. Nex , in es ima ing he GRM ma ix in he GCTA so wa e, SNPs wi h MAF 0.05 we e excluded. The addi i e gene ic ela ionship ma ices (GRM) es ima ed based on SNPs o all indi iduals o med he basis o es ima e he p opo ion o pheno ypic a iance explained by SNPs in he NTR and QIMR coho s. In o he wo ds, i was de e mined o wha ex en pheno ypic simi- la i y be ween indi iduals co esponds o gene ic simila i y (a he SNP le el). Fo bo h NTR and QIMR, sex, age and a se o popula ion-speci ic PCs we e included as co a ia es. Resul s Me a-analysis o GWA esul s Me a-analysis o GWA esul s ac oss he 29 disco e y coho s did no yield genome-wide signi ican SNPs asso- cia ed wi h ex a e sion. The lowes P- alue obse ed was 2.9 910 -7 o a SNP loca ed on ch omosome 2. The e we e 74 SNPs wi h P- alues 1910 -5 . The Manha an and quan ile–quan ile plo s a e p o ided in Figs. 1and 2.A lis wi h he op i e SNPs is gi en in Table 2. A lis wi h all SNPs ha eached he le el o sugges i e genome-wide signi icance (P 1910 -5 ) is ound in Supplemen a y Table 3. The esul s o all SNPs can be downloaded om www. weelingen egis e .o g/GPC. A gene-based es showed one signi ican hi o LOC101928162, a long non- coding RNA si e, P=2.87 910 -6 . A lis wi h he op i e genes om he gene-based analysis is p o ided in Table 3. Supplemen a y Table 4 p o ides he op 30 genes. Among he op 30 genes was B ain-De i ed Neu o ophic Fac o (BDNF,P=0.0003), a gene also implica ed, hough no genome-wide signi ican , in Te acciano e al. (2010), as was he BDNF an i-sense RNA gene (P=0.0001). Resul s o he ollow-up analysis o he op i e SNPs in he GS:SFHS coho can be ound in Table 2. O he op i e SNPs, none showed a signi ican e ec . Fo an o e - iew o he eplica ion esul s o all op SNPs wi h P- alue 1910 -5 see Supplemen a y Table 3. O he 74 SNPs es ed in he eplica ion coho , h ee SNPs showed nomi- nal e idence o associa ion (P 0.05), which is less han he numbe expec ed based on chance alone (0.05 974 =3.7). Beha Gene (2016) 46:170–182 175 123 Polygenic isk sco e analysis The e we e 8201 pe sons indi iduals wi h polygenic sco es o p edic ion o ex a e sion. The LDp ed-based gene ic isk sco es signi ican ly p edic ed ex a e sion in he a ge coho , B=0.059, X 2 (1) =27.30, P 0.001. Va iance explained by SNPs In he NTR coho , an es ima ed 5.0 % (SE =7.2) o he a iance in ex a e sion was explained by all SNPs, bu his es ima e was no signi ican ly di e en om ze o (P=0.24). In he QIMR coho , 0.0001 % (SE =15) o he a iance was explained by SNPs (P=0.46). Discussion This s udy assessed he in luence o common gene ic a ian s on ex a e sion in 63,030 indi iduals om 29 coho s in he GPC. Fi s , a me a-analysis o GWA anal- yses ac oss 29 disco e y coho s showed no genome-wide signi ican SNPs. Top SNPs de ec ed in he me a-analysis o GWA esul s in he disco e y phase we e no eplica ed in he GS:SFHS coho . The SNPs wi h lowes P- alues ha e no p e iously epo ed ela ionship wi h pe sonali y, psychopa hology o b ain unc ioning. Polygenic isk sco es based on he me a-analysis esul s p edic ed ex a e sion in an independen da a se . SNP-based Fig. 1 Manha an plo o me a-analysis esul s o 29 disco e y coho s o ex a e sion in he Gene ics o Pe sonali y Conso ium Fig. 2 Quan ile-Quan ile plo s o me a-analysis esul s o 29 disco e y coho s o ex a e sion in he Gene ics o Pe sonali y Conso ium Table 2 Top SNPs om he me a-analysis o GWA esul s in 29 disco e y coho s o ex a e sion, and hei eplica ion in he GS:SFHS coho , in he Gene ics o Pe sonali y Conso ium SNP Ch _BP Alleles Closes gene Disco e y esul s Replica ion esul s E ec SE P- alue E ec SE P- alue s2024488 2_217662968 A/G LOC101928250 -0.0303 0.0059 2.939 x 10 -7 0.0285 0.0164 0.08244 s2712162 2_217661788 T/C LOC101928250 -0.0300 0.0059 3.872 x 10 -7 0.0278 0.0164 0.08947 s797182 12_10900487 A/G NA[-0.0277 0.0056 6.673 x 10 -7 -0.0135 0.0153 0.37721 s8010306 14_37150160 A/G SLC25A21 0.0629 0.0128 8.730 x 10 -7 0.0180 0.0314 0.56650 s117292860 19_2227621 A/C DOT1L 0.0553 0.0113 9.191 x 10 -7 -0.0350 0.0239 0.14368 176 Beha Gene (2016) 46:170–182 123 he i abili ies o ex a e sion we e no signi ican ly di - e en om ze o in wo la ge coho s o he GPC. Al hough he e we e no genome-wide signi ican esul s o indi idual SNPs, in he gene-based analysis, he e was a signi ican hi o one locus, LOC101928162. This is long noncoding RNA si e whose unc ion emains elusi e. In e es ingly, among he op 30 genes we e genes p e i- ously implica ed in ex a e sion o in psychia ic diso de s associa ed wi h ex a e sion. The low P- alue o CRTAC1 (P=2.97 x 10 -5 ), ha ks back o an in e es ing ex a e sion SNP ( s7088779) in a p e ious GWAS on pe sonali y (Amin e al. 2013) ha is loca ed be ween CRTAC1 and C10o 28.RELN (P=5.69 x 10 -5 ) has been epo ed o inc ease he isk o schizoph enia and bipola diso de (Kuang e al. 2011; O adia and Shi man 2011), while ADAM12 (7.65 x 10 -5 ) was p e iously ound o be in ol ed in schizoph enia (Fa kas e al. 2010), and bipola diso de ea men (Nad i e al. 2007). The BDNF gene was also implica ed in a p e ious ex a e sion GWAS (Te acciano e al. 2010), hough no genome-wide signi ican . Liu e al. (2005) epo ed a end owa ds associa ion o BDNF a ian s wi h subs ance abuse, Jiao e al. (2011) epo ed an associa ion wi h obe- si y, and Lang e al. (2007) and Beu en e al. (2005) e- po ed associa ions wi h smoking beha io . As ex a e sion is known o be associa ed wi h li es yle, obesi y and subs ance abuse, we deem BDNF o be an in e es ing candida e gene o ex a e sion in u u e s ud- ies, along wi h CRTAC,ADAM12 and RELN. Wi h he cu en me a-analysis we mo e han ipled he sample size as compa ed o he la ges p e iously published me a-analysis o ex a e sion (De Moo e al. 2012). In con as o neu o icism, no genome-wide signi ican SNPs we e ound. Some ha e a gued (Tu kheime e al. 2014) ha he he i abili y o pe sonali y ai s ep esen s nonspeci ic gene ic backg ound, which is composed o so many gene ic a ian s wi h ex emely small e ec sizes ha indi idually hese ha e no causal biological in e p e a ion. I may be ha ex a e sion di e s in his espec om neu o icism. One o he di e ence was indica ed om he analyses o he IRT- based ex a e sion and neu o icism sco es: whe eas o neu o icism no e idence o geno ype x sex in e ac ion was seen ( an den Be g e al. 2014), o ex a e sion he e was signi ican e idence o sex limi a ion. I also is in e es ing o no e ha despi e he ac ha o ex a e sion no genome- wide signi ican indings eme ged o single SNPs, we we e able o p edic ex a e sion in an independen da ase , based on he polygenic isk coho s om he disco e y se . This indica es ha some ue signal is en ailed in he me a-anal- ysis esul s. The esul s o he polygenic isk sco e analysis a e in con as wi h he esul s om he GCTA analysis, in which no signi ican p opo ion o a iance explained by SNPs was de ec ed in wo la ge coho s o he GPC. Ou s udy on neu o icism epo ed a SNP-based he i abili y o 15 % (De Moo e al. 2015). The cu en ex a e sion GCTA indings a e also somewha a odds wi h wo p e ious GCTA s udies o pe sonali y ai s. One s udy ocused on neu o icism and ex a e sion as measu ed wi h di e en ins umen s in ou coho s, and ound on a e age 12 % explained a i- ance o ex a e sion, al hough ac oss coho s hese es i- ma es a ied widely (0–27 %) (Vinkhuyzen e al. 2012). Es ima es o neu o icism also a ied, bu we e gene ally lowe han o ex a e sion in his s udy, wi h an a e age o 6 % explained a iance. In ano he s udy, be ween 4.2 and 9.9 % o explained a iances we e ound o he ou Cloninge empe amen s in a combined sample o ou coho s (Ve weij e al. 2012). The p opo ions o a iances o Ha m A oidance, No el y Seeking and Pe sis ence we e signi ican a P 0.05, whe eas in e es ingly he p opo ion o a iance o Rewa d Dependence was no . I should be no ed ha bo h hese s udies included he QIMR coho in hei analyses, so he e is some o e lap in sub- jec s ac oss s udies. The di e ence is ha in he ea lie s udies ex a e sion and ewa d dependence we e based on single pe sonali y in en o ies, while in ou s udy ex a e sion sco es ha monized among di e en pe sonal- i y in en o ies we e analyzed. Wha ou esul s and he esul s in he p e ious s udies ha e in common hough, is ha he es ima es a e conside ably smalle han he he i- abili y es ima es based on win s udies. Gi en ha abou hal o he he i abili y o ex a e sion consis s o non-ad- di i e gene ic a iance ( an den Be g e al. 2014), i is no unlikely ha his disc epancy is caused by he in luence o Table 3 Top genes om he me a-analysis o GWA esul s in 29 disco e y coho s o Ex a e sion in he Gene ics o Pe sonali y Conso ium Gene Full gene name Pa hways P- alue LOC101928162 [Long non-coding RNA] Unknown 0.00000287 LOC729506 [Long non-coding RNA] Unknown 0.00000893 PLEKHJ1 Plecks in Homology Domain Con aining, Family J Membe 1 Phospholipid binding, ci cadian clock unc ioning 0.0000132 POU2F3 POU Class 2 Homeobox 3 In luenza A 0.0000179 CRTAC1 Ca ilage Acidic P o ein 1 Unknown 0.0000297 Beha Gene (2016) 46:170–182 177 123 common a ian s ha in e ac wi hin loci (dominance) o ac oss loci (epis asis). In addi ion, he in luence o a e a ian s may be implica ed. The ela i ely limi ed in lu- ence o common addi i e gene ic a ia ion, as well as a p e iously epo ed inding ha highe le els o inb eeding a e associa ed wi h less socially desi able pe sonali y ai le els, has led o he idea ha he gene ic a ia ion in pe sonali y ai s may ha e been main ained by mu a ion– selec ion balance (Ve weij e al. 2012), and ou esul s a e consis en wi h his idea. This s udy comes wi h some limi a ions. Geno yping, QC, and impu a ion we e ca ied ou sepa a ely in each coho . Any di e ence in p ocedu es may ha e caused some loss o s a is ical powe o de ec SNPs in he me a- analysis. Simila ly, ex a e sion i em da a we e ha mo- nized as much as possible ( an den Be g e al. 2014), bu he Rewa d Dependence i em da a om he TCI we e leas success ully linked o he ex a e sion da a om he o he in en o ies. This may also ha e caused some loss in powe . Impo an ly howe e , i should be no ed ha by combining geno ype and pheno ype da a ac oss coho s as pe o med in his s udy, a subs an ial inc ease in sample size was ob ained. I is non i ial ha he gain in powe associa ed wi h his inc ease in sample size la gely ou weighs any po en ial loss in powe due o any emaining geno yping o pheno yping di e ences ac oss coho s. In conclusion, ex a e sion is a he i able, highly poly- genic pe sonali y ai wi h a gene ic backg ound ha may be quali a i ely di e en om ha o o he complex beha io al ai s. Fu u e s udies a e equi ed o inc ease ou knowledge o which ypes o gene ic a ian s, by which modes o gene ac ion, cons i u e he he i able na u e o ex a e sion. Ul ima ely, his knowledge can be used o inc ease ou unde s anding o how ex a e sion is ela ed o a ious impo an psychosocial and heal h ou comes. Acknowledgmen s We would like o hank all pa icipa ing sub- jec s. Analyses we e ca ied ou on he Gene ic Clus e Compu e (h p://www.gene icclus e .o g), which is inancially suppo ed by he Ne he lands O ganiza ion o Scien i ic Resea ch (NWO 480-05-003). ALSPAC We a e ex emely g a e ul o all he amilies who ook pa in his s udy, he midwi es o hei help in ec ui ing hem and he whole ALSPAC eam, which includes in e iewe s, compu e and labo a o y echnicians, cle ical wo ke s, esea ch scien is s, olun- ee s, manage s, ecep ionis s and nu ses. The UK Medical Resea ch Council (G an 74882), he Wellcome T us (G an 076467) and he Uni e si y o B is ol p o ide co e suppo o ALSPAC. We hank 23andMe o unding he geno yping o he ALSPAC child en’s sample. This publica ion is he wo k o he au ho s, and hey will se e as gua an o s o he con en s o his pape . BLSA We acknowledge suppo om he In amu al Resea ch P og am o he NIH, Na ional Ins i u e on Aging. We hank Robe McC ae. BRESCIA We acknowledge suppo om he I alian Minis y o Heal h (RC and RF2007 Con . 42) and Regione Lomba dia (ID: 17387 SAL-13). We hank Ila ia Gandin o impu a ion analysis suppo . CHICAGO This wo k was suppo ed by NIH G an s, DA007255 (ABH), HG006265 ( o BEE), DA02812 ( o HdW), and DA021336 and DA024845 ( o AAP). BEE was also unded h ough he Bioin- o ma ics Resea ch De elopmen Fund, suppo ed by Ka h yn and Geo ge Gould. We wish o hank And ew D. Skol o p o iding ad ice abou geno ype calling. CILENTO We acknowledge D Ma ia Enza Amendola o he es adminis a ion and hank he pe sonnel wo king in he o ganiza ion o he s udy in he illages. MC ecei ed unding suppo om he I alian Minis y o Uni e si ies (FIRB - RBNE08NKH7, INTERO- MICS Flaghip P ojec ), he Assesso a o Rice ca Regione Campania, he Fondazione con il SUD (2011-PDR-13), and he Fondazione Banco di Napoli. SAGE –COGA/CONGEND Funding suppo o he S udy o Addic ion: Gene ics and En i onmen (SAGE) was p o ided h ough he NIH Genes, En i onmen and Heal h Ini ia i e [GEI] (U01 HG004422). SAGE is one o he genome-wide associa ion s udies unded as pa o he Gene En i onmen Associa ion S udies (GEN- EVA) unde GEI. Assis ance wi h pheno ype ha moniza ion and geno ype cleaning, as well as wi h gene al s udy coo dina ion, was p o ided by he GENEVA Coo dina ing Cen e (U01 HG004446). Assis ance wi h da a cleaning was p o ided by he Na ional Cen e o Bio echnology In o ma ion. Suppo o collec ion o da ase s and samples was p o ided by he Collabo a i e S udy on he Gene ics o Alcoholism (COGA; U10 AA008401) and he Collabo a i e Gene ic S udy o Nico ine Dependence (COGEND; P01 CA089392). Funding suppo o geno yping, which was pe o med a he Johns Hopkins Uni e si y Cen e o Inhe i ed Disease Resea ch, was p o ided by he NIH GEI (U01HG004438), he Na ional Ins i u e on Alcohol Abuse and Alcoholism, he Na ional Ins i u e on D ug Abuse, and he NIH con ac ‘‘High h oughpu geno yping o s udying he gene ic con ibu ions o human disease’’(HHSN268200782096C). The Col- labo a i e S udy on he Gene ics o Alcoholism (COGA), P incipal In es iga o s B. Po jesz, V. Hesselb ock, H. Edenbe g, L. Bie u , includes en di e en cen e s: Uni e si y o Connec icu (V. Hessel- b ock); Indiana Uni e si y (H.J. Edenbe g, J. Nu nbe ge J ., T. Fo oud); Uni e si y o Iowa (S. Kupe man, J. K ame ); SUNY Downs a e (B. Po jesz); Washing on Uni e si y in S . Louis (L. Bie u , A. Goa e, J. Rice, K. Bucholz); Uni e si y o Cali o nia a San Diego (M. Schucki ); Ru ge s Uni e si y (J. Tisch ield); Texas Biomedical Resea ch Ins i u e (L. Almasy), Howa d Uni e si y (R. Taylo ) and Vi ginia Commonweal h Uni e si y (D. Dick). O he COGA collabo a o s include: L. Baue (Uni e si y o Connec icu ); D. Kolle , S. O’Conno , L. We he ill, X. Xuei (Indiana Uni e si y); G ace Chan (Uni e si y o Iowa); S. Kang, N. Manz, M. Rangaswamy (SUNY Downs a e); J. Roh baugh, J-C Wang (Washing on Uni e si y in S . Louis); A. B ooks (Ru ge s Uni e si y); and F. Alie (Vi ginia Commonweal h Uni e si y). A. Pa sian and M. Reilly a e he NIAAA S a Collabo a o s. This na ional collabo a i e s udy is suppo ed by NIH G an U10AA008401 om he Na ional Ins i u e on Alcohol Abuse and Alcoholism (NIAAA) and he Na ional Ins i u e on D ug Abuse (NIDA). The Collabo a i e Gene ic S udy o Nico ine Dependence (COGEND) p ojec is a collabo a i e esea ch g oup and pa o he NIDA Gene ics Conso ium. Subjec collec ion was sup- po ed by NIH G an P01 CA089392 (L.J. Bie u ) om he Na ional Cance Ins i u e. Pheno ypic and geno ypic da a a e s o ed in he NIDA Cen e o Gene ic S udies (NCGS) a h p://zo k.wus l.edu/ unde NIDA Con ac HHSN271200477451C (J. Tisch ield and J. Rice). Jaime De inge was suppo ed by NIH T32 MH016880. EGCUT AM and TE ecei ed suppo om FP7 G an s (201413 ENGAGE, 212111 BBMRI, ECOGENE (No. 205419, EBC)) and OpenGENE. AM and TE also ecei ed a ge ed inancing om Es onian Go e nmen SF0180142s08 and by EU ia he Eu opean Regional De elopmen Fund, in he ame o Cen e o Excellence in Genomics. The geno yping o he Es onian Genome P ojec samples we e pe o med in Es onian Biocen e Geno yping Co e Facili y, AM 178 Beha Gene (2016) 46:170–182 123