Genome-wide association study of classical Hodgkin lymphoma identifies key regulators of disease susceptibility

ARTICLE Genome-wide associa ion s udy o classical Hodgkin lymphoma iden ifies key egula o s o disease suscep ibili y Ami Sud 1, Hauke Thomsen 2, Philip J. Law 1, As a Fö s i2,3, Miguel Inacio da Sil a Filho2, Amy Hol oyd1, Pe e B ode ick 1, Giulia O lando1, Oleg Leni e1, Lau en W igh 1, Rosie Cooke1, Douglas Eas on4,5, Paul Pha oah 4,5, Alison Dunning4, Julian Pe o6, Fede ico Canzian7, Rosalind Eeles 1,8, ZSofia Ko e-Ja ai1, Kenne h Mui 9,10, No a Pashayan 5,11, The PRACTICAL conso ium, Pe Ho mann12,13, Ma kus M. Nö hen13,14, Ka l-Heinz Jöckel15, Elke Pogge on S andmann16, T acy Ligh oo 17, Eleano Kane17, E e Roman17, Anne e Lake18, Do o hy Mon gome y18, Ru h F. Ja e 18, An hony J. Swe dlow1,19, And eas Enge 16, Nick O 20, Ka i Hemminki2,3 & Richa d S. Houls on1,20 Se e al suscep ibili y loci o classical Hodgkin lymphoma ha e been epo ed. Howe e , much o he he i able isk is unknown. He e, we pe o m a me a-analysis o wo exis ing genome-wide associa ion s udies, a new genome-wide associa ion s udy, and eplica ion o alling 5,314 cases and 16,749 con ols. We iden i y isk loci o all classical Hodgkin lymphoma a 6q22.33 ( s9482849,P=1.52 × 10−8) and o nodula scle osis Hodgkin lym- phoma a 3q28 ( s4459895, P=9.43 × 10−17), 6q23.3 ( s6928977, P=4.62 × 10−11), 10p14 ( s3781093, P=9.49 × 10−13), 13q34 ( s112998813, P=4.58 × 10−8) and 16p13.13 ( s34972832,P=2.12 × 10−8). Addi ionally, independen loci wi hin he HLA egion a e obse ed o nodula scle osis Hodgkin lymphoma ( s9269081, HLA-DPB1*03:01, Val86 in HLA-DRB1) and mixed cellula i y Hodgkin lymphoma ( s1633096, s13196329, Val86 in HLA- DRB1). The new and es ablished isk loci localise o a eas o ac i e ch oma in and show an o e - ep esen a ion o ansc ip ion ac o binding o de e minan s o B-cell de elopmen and immune esponse. DOI: 10.1038/s41467-017-00320-1 OPEN 1Di ision o Gene ics and Epidemiology, The Ins i u e o Cance Resea ch, London SW7 3RP, UK. 2Di ision o Molecula Gene ic Epidemiology, Ge man Cance Resea ch Cen e, Heidelbe g 69120, Ge many. 3Cen e o P ima y Heal h Ca e Resea ch, Lund Uni e si y, Malmö 221 00, Sweden. 4Cen e o Cance Gene ic Epidemiology, Depa men o Oncology, Uni e si y o Camb idge, Camb idge CB1 8RN, UK. 5Cen e o Cance Gene ic Epidemiology, Depa men o Public Heal h and P ima y Ca e, Uni e si y o Camb idge, Camb idge CB1 8RN, UK. 6Depa men o Non-Communicable Disease Epidemiology, London School o Hygiene and T opical Medicine, London WC1E 7HT, UK. 7Genomic Epidemiology G oup, Ge man Cance Resea ch Cen e (DKFZ), Heidelbe g 69120, Ge many. 8Royal Ma sden NHS Founda ion T us , London SM2 5NG, UK. 9Ins i u e o Popula ion Heal h, Uni e si y o Manches e , Manches e M1 3BB, UK. 10 Di ision o Heal h Sciences, Wa wick Medical School, Wa wick Uni e si y, Wa wick CV4 7AL, UK. 11 Depa men o Applied Heal h Resea ch, Uni e si y College London, London WC1E 7HB, UK. 12 Depa men o Biomedicine, Di ision o Medical Gene ics, Uni e si y o Basel, Basel 4031, Swi ze land. 13 Ins i u e o Human Gene ics, Uni e si y o Bonn, Bonn 53127, Ge many. 14 Depa men o Genomics, Li e & B ain Cen e , Uni e si y o Bonn, Bonn 53127, Ge many. 15 Uni e si y o Duisbu g–Essen, Essen 47057, Ge many. 16 Depa men o In e nal Medicine, Uni e si y Hospi al o Cologne, Cologne 50937, Ge many. 17 Depa men o Heal h Sciences, Uni e si y o Yo k, Yo k YO10 5DD, UK. 18 MRC Uni e si y o Glasgow Cen e o Vi us Resea ch, Glasgow G61 1QH, UK. 19 Di ision o B eas Cance Resea ch, The Ins i u e o Cance Resea ch, London SW7 3RP, UK. 20 Di ision o Molecula Pa hology, The Ins i u e o Cance Resea ch, London SW7 3RP, UK. Ami Sud and Hauke Thomsen con ibu ed equally o his wo k. Ka i Hemminki and Richa d S. Houls on join ly supe ised his wo k. Co espondence and eques s o ma e ials should be add essed o R.S.H. (email: [email p o ec ed]). A ull lis o conso ium membe s appea s a he end o he pape . NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions 1 1234567890 Classical Hodgkin lymphoma (cHL) is a lymphoid malig- nancy o ge minal cen e (GC) B-cell o igin1,whichis cha ac e ised by Hodgkin and Reed–S e nbe g (HRS) cells wi h a dominan backg ound popula ion o eac i e inflamma o y cells1. O he ou majo sub ypes o cHL, nodula scle osis Hodgkin lymphoma (NSHL) and mixed cellula i y Hodgkin lym- phoma (MCHL) accoun o 65% and 20% o cHL, espec i ely2. While Eps ein–Ba i us (EBV) in ec ion is causally associa ed wi h a subse o cHL cases, p opo ionally highe in MCHL, no o he en i onmen al ac o has hus a been obus ly linked o cHL isk3. E idence o inhe i ed gene ic influence on suscep ibili y o cHL is p o ided by he amilial isk and he high conco dance be ween monozygo ic wins4,5. A s ong HLA associa ion o cHL isk is well es ablished; howe e , ou unde s anding o cHL he i abili y has been ans o med by ecen genome-wide associa ion s udies (GWAS), which ha e iden ified single- nucleo ide polymo phisms (SNPs) a se en non-HLA loci influencing isk6–9. Al hough p ojec ions indica e ha addi ional isk a ian s o cHL can be disco e ed by GWAS10, he s a is ical powe o published s udies is limi ed. To gain a mo e comp ehensi e insigh in o cHL p edisposi ion, we pe o med a me a-analysis o wo p e ious GWAS7,8and a new GWAS, he eby mo e han doubling s udy powe o disco e isk SNPs. Wi h eplica ion, ou s udy has allowed us o iden i y six new non-HLA isk loci. Addi ionally, by conduc ing egion- specific impu a ion we ha e defined he specific HLA associa ions unde lying NSHL and MCHL isk. Resul s Associa ion analysis. We analysed GWAS da a om h ee s udies o Eu opean ances y: a new GWAS om he UK Na ional S udy o Hodgkin Lymphoma Gene ics (NSHLG) and wo p e iously epo ed GWAS (Supplemen a y Table 1)7,8. A e quali y con ol he h ee s udies p o ided SNP geno ypes on 3,077 cases and 13,680 con ols (Supplemen a y Tables 2,3,4; Supplemen a y Fig. 1). To inc ease genomic esolu ion, we impu ed >10 million SNPs using he 1000 Genomes P ojec and he UK10K da a as e e ence11,12. Quan ile–quan ile (Q–Q) plo s o SNPs wi h mino allele equency (MAF) >0.05% pos impu a ion did no show e idence o subs an i e o e -dispe sion (λ=1.03–1.09; Supplemen a y Fig. 2). An o e iew o he ana- lysis s a egy is ou lined in Supplemen a y Fig. 3. Me a-analysing he associa ion es esul s om he h ee GWAS in o a join disco e y se , we calcula ed join odds a ios and 95% confidence in e als o each SNP and associa ed pe -allele P- alue o all cHL, NSHL and MCHL cases s. con ols (Supplemen a y Fig. 4). In his analysis, associa ions o he es ablished non-HLA isk loci a 2p16.1, 3p24.1, 5q31.1, 6q23.3, 8q24.21, 10p14 and 19p13.3 we e consis en in di ec ion and magni ude o e ec wi h p e- iously epo ed s udies (Supplemen a y Fig. 4; Supplemen a y Table 5)6–8. We sough alida ion o associa ion SNPs wi h a P- alue om he me a-analysis unde a fixed-e ec s model a P<1.0 × 10−7 and P<1.0 × 10−6 o loci no p e iously associa ed wi h cHL and NSHL isk, espec i ely, by geno yping wo addi ional indepen- den se ies (Supplemen a y Table 1), o alling 2,237 cases and 3,069 con ols (Table 1; Supplemen a y Table 6). Whe e he s onges signal was p o ided by an impu ed SNP, we confi med he fideli y o impu a ion by geno yping (Supplemen a y Table 7). In he combined me a-analysis, we iden ified genome-wide significan associa ions o cHL (Table 1; Supplemen a y Tables 8 and 9), a 3q28 ( s4459895, P=4.45 × 10−18), 6q22.33 ( s9482849, P=1.52 × 10−8), 6q23.3 ( s6928977, P=1.24 × 10−10) and 10p14 Table 1 Summa y esul s o newly iden ified isk loci Nea es geneaRisk allele ( equency) Disco e y GWAS me a-analysis UK Replica ion 1 UK Replica ion 2 Me a-analysis Posi ion (hg19, bp) P- alue OR (95% CI) P- alue OR (95% CI) P- alue OR (95% CI) P- alue OR (95% CI) I2 (%) P he 3q28, s4459895 LPP A (0.20) 187954414 cHL 4.16 × 10−10 1.27 (1.18–1.36) 6.85 × 10−91.44 (1.27–1.63) 0.02 1.26 (1.04–1.52) 4.45 × 10−18 1.30 (1.23–1.38) 13 0.33 NSHL 9.16 × 10−91.37 (1.23–1.53) 1.37 × 10−81.43 (1.26–1.62) 0.04 1.30 (1.02–1.66) 9.43 × 10−17 1.39 (1.28–1.50) 0 0.93 MCHL 0.92 1.04 (0.91–1.19) 0.98 1.00 (0.68–1.47) 0.55 1.04 (0.92–1.19) 0 0.82 6q22.33, s9482849 PTPRK C (0.17) 128288536 cHL 5.02 × 10−81.24 (1.15–1.35) 0.13 1.11 (0.97–1.27) 0.13 1.17 (0.95–1.43) 1.52 × 10−81.20 (1.13–1.28) 3 0.39 NSHL 2.91 × 10−61.32 (1.17–1.48) 0.17 1.10 (0.96–1.25) 0.20 1.19 (0.91–1.54) 4.13 × 10−61.21 (1.12–1.33) 10 0.35 MCHL 0.17 1.11 (0.96–1.28) 0.78 1.06 (0.70–1.61) 0.16 1.10 (0.96–1.26) 0 0.97 6q23.3, s6928977 AHI1 G (0.57) 135626348 cHL 1.66 × 10−81.18 (1.12–1.26) 0.01 1.14 (1.03–1.26) 0.05 1.16 (1.00–1.34) 1.24 × 10−10 1.17 (1.12–1.23) 0 0.85 NSHL 9.34 × 10−10 1.31 (1.20–1.42) 0.03 1.12 (1.01–1.24) 0.01 1.30 (1.06–1.58) 4.62 × 10−11 1.23 (1.16–1.31) 26 0.25 MCHL 0.24 1.06 (0.96–1.18) 0.69 1.06 (0.79–1.42) 0.22 1.06 (0.96–1.17) 0 0.22 10p14, s3781093 GATA3 T (0.88) 8101927 cHL 4.89 × 10−12 1.35 (1.23–1.47) 4.00 × 10−41.32 (1.25–1.44) 0.11 1.21 (0.96–1.52) 4.91 × 10−12 1.28 (1.19–1.37) 64 0.01 NSHL 9.16 × 10−12 1.53 (1.36–1.75) 2.00 × 10−41.44 (1.31–1.61) 0.64 0.92 (0.68–1.26) 9.49 × 10−13 1.39 (1.28–1.53) 61 0.06 MCHL 0.03 1.18 (1.02–1.36) 0.05 1.56 (1.02–2.40) 0.16 0.91 (0.79–1.04) 73 0.03 13q34, s112998813 UPF3A C (0.08) 115059729 cHL 3.63 × 10−31.19 (1.06–1.33) 0.03 1.23 (1.03–1.47) 0.43 1.12 (0.84–1.50) 2.70 × 10−41.19 (1.08–1.30) 13 0.32 NSHL 8.43 × 10−81.58 (1.34–1.88) 0.03 1.22 (1.02–1.47) 0.28 1.23 (0.85–1.78) 4.58 × 10−81.39 (1.23–1.56) 27 0.24 MCHL 0.92 0.99 (0.80–1.22) 0.27 1.35 (0.80–2.23) 0.75 1.03 (0.85–1.25) 0 0.56 16p13.13, s34972832 CLEC16A A (0.18) 11198938 cHL 1.45 × 10−41.15 (1.07–1.23) 6.34 × 10−31.18 (1.05–1.34) 0.10 1.17 (0.97–1.42) 8.03 × 10−71.16 (1.09–1.23) 6 0.37 NSHL 7.47 × 10−71.24 (1.15–1.34) 6.53 × 10−31.30 (1.17–1.45) 0.28 1.15 (0.89–1.50) 2.12 × 10−81.24 (1.15–1.34) 37 0.18 MCHL 0.65 0.97 (0.85–1.11) 0.91 1.02 (0.69–1.52) 0.70 0.98 (0.86–1.10) 0 0.94 The isk allele is he allele co esponding o he es ima ed odds a io. F equency o he isk allele is om he CEU popula ion om 1000 Genomes P ojec cHL classical Hodgkin lymphoma, NSHL nodula scle osis Hodgkin lymphoma, MCHL mixed cellula i y Hodgkin lymphoma, bp base pai , OR odds a io, CI confidence in e al, I2p opo ion o he o al a ia ion due o he e ogenei y I2 alue ≥75% is conside ed o be cha ac e is ic o la ge he e ogenei y aNea es gene may no be he unc ional gene ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 2NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions ( s3781093, P=4.91 × 10−12), which we e p edominan ly d i en by an associa ion wi h NSHL. The s6928977 associa ion is independen o he p e iously iden ified associa ion a 6q23.3 ma ked by s9402684 (Supplemen a y Table 5); espec i e condi ional P- alues, P=1.28 × 10−8and P=9.80 × 10−6(pai - wise LD me ics 2=0.002, D’=0.007)8. Fu he mo e, he s3781093 associa ion is independen o he p e iously iden ified 10p14 associa ion ma ked by s2388486 (Supplemen a y Table 5); espec i e condi ional P- alues a e P=3.38 × 10−8and P= 1.32 × 10−12 (pai wise LD me ics 2=0.002, D’=0.27)7. Fo NSHL we iden ified wo new associa ions a 13q34 ( s112998813, P=4.58 × 10−8) and 16p13.13 ( s34972832, P=2.12 × 10−8, Table 1). Rela ionship be ween he new isk SNPs and pheno ype.A hallma k o cHL epidemiology is he bimodal age-specific inci- dence and i has been a gued ha he disease in young adul s and olde adul s is ae iologically di e en ; in pa icula he e is a low p e alence o EBV-posi i e disease in NSHL pa ien s aged 16–353. Case-only analysis did no p o ide e idence o sex di - e ences a newly iden ified isk SNPs (Supplemen a y Table 10) o a ela ionship be ween age in he NSHL subg oup. Albei no significan a e co ec ion o mul iple es ing, we obse ed an associa ion be ween EBV-posi i e disease and cHL a 6q23.3 in 796 cases analysed ( s6928977, P=0.03, Supplemen a y Table 10). Biological in e ence. Fi e o he six new isk SNPs localise in o nea genes which ha e ei he been p e iously implica ed in he de elopmen o cHL o ha e es ablished oles in B-cell de elop- men and a e he e o e s ong candida es o cHL suscep ibili y. Specifically, he 6q22.33 associa ion ma ked by s9482849 maps in e genically o THEMIS ( hymocy e-exp essed molecule in ol ed in selec ion) and PTPRK ( ecep o - ype y osine p o ein phospha ase kappa) (Fig. 1). Down egula ion o PTPRK by he EBV-encoded EBNA1 con ibu es o he g ow h and su i al o HRS cells13. The 6q23.3 associa ion defined by s6928977 loca- lises o in on 3 o AHI1 (abelson helpe in eg a ion si e-1) (Fig. 1) which has been implica ed in he de elopmen o bo h B- and T-cell lymphoma14,15. The 13q34 associa ion ma ked by s112998813 is loca ed in in on 5 o UPF3A (Fig. 1), a egula o o nonsense ansc ip s16. The LD egion o associa ion also ha bou s CDC16 (cell di ision cycle p o ein 16). CDC16, a sub- uni o he anaphase-p omo ing complex17, a ge s cell cycle egula o y p o eins o p o easome deg ada ion, he eby allowing cell cycle p og ession, and is down egula ed in HRS cells18.A 16p13.13, he s34972832 associa ion o NSHL maps o in on 18 o CLEC16A (C- ype lec in domain amily 1, Fig. 1) whose loss o unc ion a ec s bo h B-cell numbe and unc ion19. The 10p14 associa ion ma ked by s3781093 maps in onic o GATA3 (Fig. 1). T ansc ip ional ep ession o GATA3 is essen ial o ea ly B-cell commi men , and abe an GATA3 exp ession has been obse ed in HRS cells20,21. In iguingly, he s3781093 isk allele o NSHL has p e iously been demons a ed o be p o ec i e o paedia ic B-cell acu e lymphoblas ic leukaemia (ALL)22. To he ex en ha hey ha e been deciphe ed, many GWAS isk loci map o non-coding egions o he genome and influence gene egula ion. Hence, o gain insigh in o he biological mechanisms o he associa ions o he newly iden ified isk SNPs, we in e oga ed publicly accessible exp ession da a on lymphoblas oid cell lines (LCLs)23,24. We used he summa y da a-based Mendelian andomisa ion (SMR) analysis o es o pleio opy be ween GWAS signal and cis-exp ession quan i a i e ai (eQTL) o genes wi hin 1 Mb o he sen inel SNP a each locus o iden i y a causal ela ionship25. A 6q23.3 and 10p14, significan eQTLs we e obse ed wi h AHI1 (P SMR =8.63 × 10−6; Supplemen a y Table 11 and Supplemen a y Fig. 5) and GATA3 (P SMR =4.70 × 10−8; Supplemen a y Table 11 and Supplemen a y Fig. 5). Since spa ial p oximi y be ween specific genomic egions and ch oma in looping in e ac ions a e cen al o he egula ion o gene exp ession, we iden ified pa e ns o ch oma in in e ac ions a candida e causal SNPs by analysing p omo e cap u e Hi-C da a on GM12878 as a sou ce o B-cell in o ma ion26. Looping ch oma in in e ac ions we e shown a 3q28 ( s4459895), 6q23.3 ( s6928977), 10p14 ( s3781093) and 16p13.13 ( s34972832). While no significan eQTL was shown o hese ch oma in looping in e ac ions hey in ol ed a numbe o genes wi h biological ele ance o cHL de elopmen (Fig. 1). A 3q28, he looping in e ac ion implica es BCL6 and mi -28, which ha e well documen ed oles in B-cell umou biology and GC B-cell de elopmen 27,28. A 6q23.3, we obse ed in e ac ions wi h p omo e sequences ups eam in MYB and ALDH8A1. A 10p14, bo h isk SNPs encompass a egion ha in e ac s wi h TAF3, which encodes ansc ip ion ini ia ion ac o TFIID subuni 3. TAF3 o ms pa o he ansc ip ion ini ia ion ac o TFIID and is necessa y o haema opoiesis29. Finally, we obse ed in e ac- ions a he 16p13.13 isk locus wi h RMI2 (encoding RecQ media ed genome ins abili y 2) (Fig. 1). RMI2 is an essen ial componen o he Bloom helicase-double Holliday junc ion dissol asome and is esponsible o genomic s abili y30. Ac oss he new and es ablished isk loci o cHL we confi med a significan en ichmen o DNase hype sensi i i y in GM12878 cells ( alse disco e y a e (FDR) adjus ed P- alue =0.0035), as well as enhance elemen s in p ima y B-cells (FDR adjus ed P- alue =0.00064) and GM12878 cells (FDR adjus ed P- alue =0.015)31. Analysis o ChIP-seq da a on 82 ansc ip ion ac o s (TFs) showed an o e - ep esen a ion o he binding o TFs ha play a cen al ole in B-cell signalling-ne wo ks such as RELA (nuclea ac o NF-kappa-B p65), EBF1 (ea ly B-cell ac o 1), RUNX3 ( un - ela ed ansc ip ion ac o 3) and BATF (basic leucine zippe ansc ip ion ac o , ATF-like) (Fig. 2). Collec- i ely, hese obse a ions suppo he asse ion ha isk loci o cHL media e hei e ec s h ough B-cell de elopmen al ne wo ks, and a e s ongly in ol ed in ansc ip ional ini ia ion and enhancemen . The HLA egion. To ob ain addi ional insigh in o plausible unc ional a ian s wi hin he majo his ocompa ibili y egion a 6p21, we impu ed he classical HLA alleles, amino-acid esidues and SNPs using SNP2HLA32. To isola e independen associa ions o NSHL and MCHL, we pe o med condi ional s ep-wise logis ic eg ession on he s onges associa ed SNP, amino acid o allele, un il no u he a ian s a ained genome-wide significance (Fig. 3; Supplemen a y Table 12). Fo NSHL, we iden ified he s onges associa ion a s9269081 (P=1.74 × 10−39), which maps wi hin he class II HLA egion, 3’o HLA-DRA. Addi ional class II associa ions we e shown by HLA-DPB1*03:01 (P=3.35 × 10−17) and Val86 in HLA-DRB1 (P=3.52 × 10−13) (Fig. 3). In con as , he s onges associa ion o MCHL was p o ided by s1633096, a class I HLA associa ion 3’o HLA-F (P=2.72 × 10−23). Addi ional class II associa ions o MCHL we e obse ed a s13196329, loca ed in onic o C6o 10 (P=2.58 × 10−14) and Val86 in HLA-DRB1 (P=7.10 × 10−9) (Fig. 3). He i abili y o cHL.Byfi ing all SNPs om GWAS simul a- neously using Genome-wide Complex T ai Analysis33, he es i- ma ed he i abili y o cHL, NSHL and MCHL a ibu able o all common a ia ion is 24.0% (±2.3%), 25.2% (±3.4%) and 21.9% (±2.4%), espec i ely. This es ima e ep esen s he addi i e NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 ARTICLE NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions 3 0 2 4 6 8 –Log10P –Log10P –Log10P –Log10P –Log10P –Log10P 0 2 4 6 8 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 Recombina ion a e (cM/Mb) s4459895 1.0 0.5 2 1.0 0.5 2 1.0 0.5 2 1.0 0.5 2 1.0 0.5 2 2 LPP BCL6 MIR28 ch 3 posi ion (Mb) 187.40Mb 187.60Mb 187.80Mb 188.00Mb 188.20Mb 188.40Mb 0 2 4 6 0 2 4 6 8 0 5 10 15 20 25 30 35 40 Recombina ion a e (cM/Mb) PTPRK THEMIS 0 5 15 0 4 8 12 16 0 5 10 15 20 25 30 Recombina ion a e (cM/Mb) s3781093 GATA3 TAF3 ch 10 posi ion (Mb) 7.85Mb 7.90Mb 7.95Mb 8.00Mb 8.05Mb 8.10Mb 0 2 4 6 8 0 2 4 6 8 0 5 10 15 20 25 30 Recombina ion a e (cM/Mb) s112998813 TMCO3 TFDP1 GRK1 FAM70B FLJ41484 LOC100506394 FLJ44054 CDC16 UPF3A CHAMP1 ADPRHL1 DCUN1D2 ATP4B LOC100130386 GAS6 RASA3 3q28 a b c d e 6q22.33 10p14 13q34 s9482849 Ch omHMM T acks Ac i e p omo e Weak p omo e Poised p omo e S ong enhance T ansc ip ional ansi ion/elonga ion Weak/poised enhance Insula o Weakly ansc ibed Polycomb− ep essed He e och oma in; low signal ch 6 posi ion (Mb) 128.00Mb 128.10Mb 128.20Mb 128.30Mb 128.40Mb 128.50Mb 128.60Mb 128.70Mb ch 13 posi ion (Mb) 114.20Mb 114.40Mb 114.60Mb 114.80Mb 115.00Mb 0 2 4 6 8 0 2 4 6 0 5 10 15 20 25 30 35 40 45 Recombina ion a e (cM/Mb) s34972832 1.0 0.5 CIITA CLEC16A RMI2 DEXI SOCS1 TNP2 PRM3 PRM2 PRM1 ch 16 posi ion (Mb) 11.00Mb 11.10Mb 11.20Mb 11.30Mb 11.40Mb 6q23.3 16p13.13 0 2 4 6 8 0 2 4 6 8 10 0 5 10 15 20 25 30 35 40 45 50 Recombina ion a e (cM/Mb) s6928977 MYB HBS1L ALDH8A1 AHI1 ch 6 posi ion (Mb) 135.20Mb 135.40Mb 135.60Mb 135.80Mb 136.00Mb Fig. 1 Regional plo s o associa ion esul s and ecombina ion a es o he newly iden ified classical Hodgkin lymphoma (NSHL) isk loci. Resul s o a3q28 ( s4459895) and nodula scle osis Hodgkin lymphoma (NSHL) isk, b6q22.33 ( s9482849) and classical Hodgkin lymphoma (cHL) isk, c6q23.3 ( s6928977) and NSHL isk, d10p14 ( s3781093) and NSHL isk, e13q34 ( s112998813) and NSHL isk, and 16p13.13 ( s34972832) and NSHL isk. Plo s show associa ion esul s o bo h geno yped ( iangles) and impu ed (ci cles) single-nucleo ide polymo phisms (SNPs) in he genome-wide associa ion s udy samples and ecombina ion a es. −log 10 P- alues (y-axes) o he SNPs a e shown acco ding o hei ch omosomal posi ions (x-axes). The sen inel SNP in each combined analysis is shown as a la ge ci cle o iangle and is labelled by i s sID. The colou in ensi y o each symbol eflec s he ex en o LD wi h he op SNP, whi e ( 2=0) h ough o da k ed ( 2=1.0). Gene ic ecombina ion a es, es ima ed using 1000 Genomes P ojec samples, a e shown wi h a ligh blue line. Physical posi ions a e based on NCBI build 37 o he human genome. Also shown a e he ela i e posi ions o genes and ansc ip s mapping o he egion o associa ion. Genes ha e been ed awn o show hei ela i e posi ions; he e o e maps a e no o physical scale. The middle ack ep esen s he ch oma in s a e segmen a ion ack (Ch omHMM) o lymphoblas oid cells using da a om he HapMap ENCODE P ojec . The op ack ep esen s cap u e Hi-C p omo e con ac s in GM12878 cells. The colou in ensi y o each con ac eflec s he in e ac ion sco e ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 4NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions a iance, and he e o e does no include he po en ial impac o dominance e ec s o gene–en i onmen in e ac ions ha ing an impac on cHL isk. The cu en ly iden ified non-HLA isk SNPs hus only accoun o a ound 12% o he addi i e he i able isk. Co-he i abili y wi h au oimmune disease. Al hough no uni- e sal, some epidemiological s udies ha e epo ed associa ions be ween cHL and a ious au oimmune diseases, aising he pos- sibili y o common gene ic suscep ibili y and hence common biological pa hways34. Va ia ion a a numbe o he cHL isk loci, including 3p24.1, 5q31.1 and 6q23.3 has p e iously been impli- ca ed as de e minan s o au oimmune disease isk suppo ing such an asse ion (Supplemen a y Da a 1). To in es iga e co-he i abili y globally be ween cHL and au oimmune disease, we implemen ed c oss- ai LD sco e eg ession35. Using summa y-le el GWAS da a we es ima ed gene ic co ela ions be ween cHL and six au oimmune diseases cu a ed by ImmunoBase; specifically heuma oid a h i is36, sys emic lupus e y hema osus37, mul iple scle osis (MS)38, p ima y bilia y ci hosis39, ulce a i e coli is (UC)40 and coeliac disease41 GWAS da a (Supplemen a y Table 13). We obse ed a posi i e gene ic co ela ion be ween cHL and MS ( g =0.35, P=0.04) and a nega i e co ela ion be ween cHL and UC ( g =−0.23, P=0.01). Discussion To ou knowledge, we ha e pe o med he la ges GWAS o cHL o da e, iden i ying six new non-HLA isk loci. The a ailabili y o comp ehensi e e e ence panels o he HLA egion has allowed us o delinea e class I and class II associa ions o NSHL and MCHL, subs an ia ing ecen documen ed di e ences be ween hese cHL his ologies9. Al hough unc ional analyses a e equi ed o de e mine he biological basis o cHL associa ion signals, we ha e demons a ed ha hese isk loci a e en iched o egula o y elemen s in B-cells. Mo eo e , hey ea u e an o e - ep esen a ion o key B-cell TF binding, no ably RELA, RUNX3, EBF1 and BATF. RELA is a TF in ol ed in NF-κB he e odime o ma ion. HRS cells show high cons i u i e ac i i y o NF-κB (bo h canonical and non-canonical pa hways)42, which p omo es cell su i al and g ow h h ough inducing an i-apop o ic and p o-p oli e a i e gene p og ams43,44. Inhibi ion o NF-κB in HRS cells leads o caspase-independen apop osis43. EBF1 coope a es wi h E2A and PAX5 o egula e B-cell ma u a ion45. I s exp ession in HRS cells is low46, which is hough o con ibu e o he loss o no mal B-cell pheno ype47. RUNX3 has impo an oles in B-cell ma u a ion48 and down- egula ion o RUNX1 by RUNX3 is equi ed o EBV-d i en LCL 3 2 1 30,000,000 HLA-A HLA-C HLA-B HLA-A HLA-C HLA-B HLA-DRB1 HLA-DQB1 HLA-DQA1 HLA-DPA1 HLA-DPB1HLA-DRA HLA-DQA1 HLA-DPB1 HLA-DPAHLA-DQB1HLA-DRB1 HLA-DRA 31,000,000 32,000,000 33,000,000 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 Posi ion −Log10P −Log10P 3 2 1 30,000,000 31,000,000 32,000,000 33,000,000 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 Posi ion s9269081 HLA-DPB1*03:01 Val86 HLA-DRB1 s1633096 s13196329 Val86 HLA-DRB1 ab Fig. 3 Manha an plo ep esen a ion o he s ep-wise condi ional logis ic eg ession o isk o anodula scle osis Hodgkin lymphoma and bmixed cellula i y Hodgkin lymphoma wi hin he human leukocy e an igen (HLA) egion. (a1) Uncondi ioned es o he HLA egion. (a2) Resul s o he HLA egion a e condi ioning on s9269081. (a3) Resul s o he HLA egion a e condi ioning on s9269081 and HLA-DPB1*03:01. (b1) Uncondi ioned es o he HLA egion. (b2) Resul s o he HLA egion a e condi ioning on s1633096. (b3) Resul s o he HLA egion a e condi ioning on s1633096 and s13196329. Physical posi ions a e based on NCBI build 36 o he human genome. The −log 10 o he combined logis ic eg ession es P- alues a e plo ed agains hei physical ch omosomal posi ion. The b oken ed line ep esen s he genome-wide le el o significance (P<5×10 −8) ATF2 BATF BCL11A BCL3 BHLHE40 CEBPB CTCF EBF1 ELF1 EZH2 FOXM1 IKZF1 IRF4 MEF2A MEF2C MTA3 NFATC1 NFIC RELA EP300 PAX5 PAX5 PML POLR2A POLR2A POU2F2 RAD21 RUNX3 SP1 STAT5A TAF1 TCF3 WRNIP1 YY1 0 1 2 3 4 0.0 2.5 5.0 7.5 En ichmen –Log10P Fig. 2 En ichmen o ansc ip ion ac o s binding a classical Hodgkin lymphoma isk loci. The en ichmen and o e - ep esen a ion o ansc ip ion ac o s binding a all cHL isk loci. The ed line ep esen s he Bon e oni co ec ed P- alue h eshold NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 ARTICLE NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions 5 g ow h49.BATF also appea s o co-o dina e B-cell ma u a ion50, and is highly exp essed in HRS cells51. The s ong HLA associa ions we iden ified o NSHL and MCHL suppo ecen obse a ions o dis inc class I and class II ela ionships o hese cHL sub ypes9. Specifically, he class II NSHL associa ion ma ked by s9269081 is in s ong LD wi h he p e iously iden ified isk SNP s6903608 ( 2=0.92, D’=1.0) o EBV-nega i e NSHL9. Fo MCHL he class I associa ion s1633096 shows co ela ion wi h he p e iously iden ified ma ke SNP s2734986 (pai wise 2=0.41, D’=0.97) o EBV-posi i e cHL9. A class I associa ion o MCHL is consis en wi h a high EBV posi i i y and suppo s he no ion o de ec i e cy o oxic T-cell lymphocy e esponses in EBV-in ec ed HRS cells52. Va ia ion wi hin he class II HLA egion al e s he isk o au oimmune diseases53, bu he unde lying biological mechanism o hese associa ions has ye o be ully defined. The class II HLA associa ion o NSHL and MCHL isk, comp ising bo h coding a ian s and non-coding SNPs, may explain he impo ance o CD4+ T ollicula helpe (T FH ) cells in cHL pa hogenesis. In he GC, he e is a equi emen o CD4+ T FH cells o in e ac wi h GC B-cells h ough he T-cell an igen ecep o (TCR) and HLA class II p o eins o no mal plasma and memo y cells o ma- ion54. I is he e o e plausible ha a ia ion in pep ide binding and exp ession o he HLA class II p o eins con ibu es o cHL pa hogenesis h ough in e ac ion wi h CD4+ T FH cells. Such a model is suppo ed by he obse a ion o a ia ion a posi ion 86 o HLA-DRB1 influencing TCR Vαgene exp ession55, he p edominance o CD4+ T-cells in cHL umou s56, he eliance o HRS cells on he mic o-en i onmen o su i al1, and he loss o MHC class II exp ession on HRS cells57, he las o which is associa ed wi h ad e se p ognosis. An al e na i e explana ion o he class II HLA associa ion in cHL is he in ol emen o an uniden ified pa hogen playing a causa i e ole in cHL. Amino- acid a ian s and SNPs wi hin HLA-DRB1 modula e humo al immune esponses o common i uses, such as influenza A and JC polyoma i us58. Consis en wi h such a model is dimo phic a ia ion a posi ion 86 o HLA-DRB159, which we iden i y as influencing isk o NSHL and MCHL, modula ing he ancho ing pocke o he an igen binding si e, and influencing he con- o ma ion o pep ide–DR p o ein complexes while main aining a T-cell esponse60. In ou analysis we no ed a ecip ocal ela ionship be ween NSHL isk and ALL isk a 10p14 (GATA3)22. Since GATA3 plays a key ole in B-cell de elopmen and bo h ALL and NSHL a e malignancies de i ed om B-cells a di e en s ages o ma u a ion, ou obse a ion leads o specula ion o a significan empo al e ec o gene ic a ia ion a his locus in esponse o an en i onmen al o mu a ional insul . Al hough suppo ed by a con empo aneous s udy and equi - ing u he alida ion61, we ound e idence o common gene ic suscep ibili y be ween cHL and MS, hus aising he possibili y o sha ed en i onmen al isk ac o s. A po en ial biological basis o such a ela ionship may encompass abe an immune ac i a ion and cell p oli e a ion. In conclusion, ou s udy p o ides u he e idence o inhe - i ed suscep ibili y o cHL and suppo s a model whe eby isk loci influence disease h ough e ec s on B-cell egula o y ne wo ks, p o iding a mechanis ic link be ween suscep ibili y and biology. Ou findings also emphasise he di e ences be ween he majo sub ypes, which a e eflec i e o di e ences in umou ae iology. Me hods E hics. Collec ion o pa ien samples and associa ed clinico-pa hological in o - ma ion was unde aken wi h w i en in o med consen . Rele an e hical e iew boa ds app o ed he indi idual s udies in acco dance wi h he ene s o he Decla a ion o Helsinki (UK-GWAS MREC 03/1/096, Ge man-GWAS Uni e si y o Heidelbe g 104/2004 and UK-GWAS-NSHLG MREC 09/MRE00/72). The diagnosis o cHL (i.e., excluding cases wi h nodula lymphocy e p edominan HL), NSHL and MCHL (ICD-10-CM C81.1-3) in all cases was es ablished in acco dance wi h Wo ld Heal h O ganisa ion guidelines. Genome-wide associa ion s udies. P ima y s udy: We analysed cons i u ional DNA om 1,717 cases asce ained h ough he NSHLG (h p://www.public.ukc n. o g.uk) om 2010 o 2013. These a e de ailed in Supplemen a y Table 1. Cases we e geno yped using he Illumina Oncoa ay (Illumina Inc.). Con ols which we e also geno yped using he oncoa ay comp ised: (1) 2,976 cance - ee men ec ui ed by he PRACTICAL Conso ium— he UK Gene ic P os a e Cance S udy (UKGPCS) (age <65 yea s), a s udy conduc ed h ough he Royal Ma sden NHS Founda ion T us and SEARCH (S udy o Epidemiology & Risk Fac o s in Cance ), ec ui ed ia GP p ac ices in Eas Anglia (2003–2009), (2) 4,446 cance - ee women om ac oss he UK ia he B eas Cance Associa ion Conso ium (BCAC). Published s udies: We used GWAS da a gene a ed on wo non-o e lapping case–con ol se ies o No he n Eu opean ances y, which ha e been he subjec o p e ious analyses ha a e summa ised in Supplemen a y Table 1. B iefly: (1) The UK-GWAS was based on 622 cases asce ained h ough he Royal Ma sden Hospi al Na ional Heal h Se ice T us Family His o y s udy du ing 2004–20087, and 5,677 con ols om he UK Wellcome T us Case Con ol Conso ium 2 (WTCCC2)62. (2) The Ge man-GWAS comp ised 1,001 cases asce ained by he Ge man Hodgkin S udy G oup du ing 1998–20078, and 1,226 con ols om he Heinz Nixdo Recall (HNR) s udy. GWAS and me a-analysis. S anda d quali y con ol measu es we e applied o each o he h ee GWAS (Supplemen a y Tables 2,3and 4)7,8,63. Specifically, indi iduals wi h a low call a e (<95%) as well as all indi iduals e alua ed o be o non-Eu opean ances y (using he HapMap e sion III CEU, JPT/CHB and YRI popula ions as a e e ence, Supplemen a y Fig. 1) we e excluded. Fo appa en fi s -deg ee ela i e pai s, we excluded he con ol om a case–con ol pai o he indi idual wi h he lowe call a e. SNPs wi h a call a e <95% we e excluded as we e hose wi h a MAF <0.01 o displaying de ia ion om Ha dy–Weinbe g equilib ium (HWE) (i.e., P<10−6, Supplemen a y Table 4). GWAS da a we e impu ed o >10 million SNP wi h IMPUTE2 2.364 so wa e, using a me ged e e ence panel consis ing o da a om 1000 Genomes P ojec (phase 1 in eg a ed elease 3, Ma ch 2012)11 and UK10K (ALSPAC, EGAS00001000090/ EGAD00001000195 and TwinsUK EGAS00001000108/EGAS00001000194 s udies) 12. HLA impu a ion was conduc ed using SNP2HLA and he Type I Diabe es Gene ics Conso ium e e ence panel o 5,225 indi iduals o Eu opean descen 32. The numbe o a ian s in he HLA impu a ion eco e ed wi h an in o ma ion measu e o >0.80 we e 8,436 (94% o o al a ian s), 8506 (95% o o al a ian s) and 8599 (96% o o al a ian s) in he UK-GWAS, Ge man-GWAS and UK- NSHLG-GWAS da a se s, espec i ely. Impu a ion was conduc ed sepa a ely o each s udy, and in each, he da a we e p uned o a common se o SNPs be ween cases and con ols p io o impu a ion. Poo ly impu ed SNPs defined by an in o ma ion measu e <0.80 we e excluded. Tes s o associa ion be ween SNPs and cHL we e pe o med using logis ic eg ession unde an addi i e gene ic model in SNPTEST 2.565. The adequacy o he case–con ol ma ching was e alua ed using Q–Q plo s o es s a is ics (Supplemen a y Fig. 2). The infla ion ac o λwas based on he 90% leas -significan SNP66. Whe e app op ia e, p incipal componen s, gene a ed using common SNPs, we e included in he analysis o limi he e ec s o c yp ic popula ion s a ifica ion ha o he wise migh cause infla ion o es s a- is ics. Eigen ec o s o he GWAS da a se s we e in e ed using sma pca (pa o EIGENSOFT) by me ging cases and con ols wi h Phase III HapMap samples. LD me ics we e calcula ed in c ools 0.1.12b67, using UK10K me ged 1000 Genomes P ojec da a and plo ed using isPIG68. Replica ion s udies and echnical alida ion. The eigh SNPs in he mos p o- mising loci (Table 1; Supplemen a y Table 6), we e aken o wa d o de no o eplica ion in an addi ional 1,284 cases om he NSHLG and 2,504 con ols om he UK eplica ion 1 se ies (Supplemen a y Table 1). A e his six SNPs we e geno yped in an addi ional eplica ion se ies, (UK eplica ion 2 se ies) comp ising 953 cases and 565 con ols om he Sco land and Newcas le Epidemiological S udy o Hodgkin Disease (SNEHD), he Young Adul Hodgkin Case–Con ol S udy (YHCCS) and he Epidemiology and Cance S a is ics G oup Lymphoma Case–Con ol S udy (ELCCS; h p://www.elccs.in o) (Supplemen a y Table 1). Full de ails o he SNEHD, YHCCS and ELCCS s udies ha e been p e iously epo ed69–71. B iefly, SNEHD in ol ed asce ainmen o inciden cases om Sco land and No he n England du ing 1993–1997. YHCCS was based on newly diagnosed cases aged 16–24 yea s om No he n England du ing 1991–1995. ELCCS comp ised cases esiding in he no h o pa s o sou hwes o England aged 16–69 yea s wi h newly diagnosed, non-human immunodeficiency i us- ela ed cHL du ing 1998–2003. UK popula ion con ols ma ched o cases on age, sex and a ea o esidence we e ob ained om SNEHD, YHCCS and ELCCS. The EBV s a us o cHL umou s in he UK eplica ion 2 se ies was de e mined by immunohis ochemical s aining o EBV la en memb ane an igen-1 and/o EBV EBV-encoded RNA in si u hyb idisa ion using sec ions o pa a fin-embedded ma e ial72,73. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 6NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions The fideli y o GWAS impu a ion was assessed by he conco dance be ween impu ed and di ec ly geno yped SNP in a subse o samples (Supplemen a y Table 7). Replica ion geno yping o UK samples was pe o med using compe i i e allele-specific PCR KASP chemis y (LGC, He o dshi e, UK). P ime s, p obes and condi ions a e de ailed in Supplemen a y Table 14. Call a es o SNP geno ypes we e >95% in each o he eplica ion se ies. To ensu e he quali y o geno yping in assays, a leas wo nega i e con ols and a se o duplica es we e geno yped (conco dance >99%). Me a-analysis. Me a-analyses we e pe o med unde a fixed-e ec s model using META 1.674. Coch an’s Q-s a is ic o es o he e ogenei y and he I2s a is ic o quan i y he p opo ion o he o al a ia ion due o he e ogenei y we e calcula ed; an I2 alue ≥75% is conside ed o be cha ac e is ic o la ge he e ogenei y75.We used he es -based me hod o Higgins e al.76 o de i e 95% CIs o I2 alues (Supplemen a y Table 9). To es ima e s udy powe o he disco e y GWAS phase, we made use o he CaTS online calcula o 77, assuming a isk allele equency o 0.2 and geno ype ela i e isk o 1.20. Exp ession quan i a i e ai locus analysis. To examine he ela ionship be ween SNP geno ype and gene exp ession, we ca ied ou SMR analysis as pe Zhu e al., 201625. B iefly, i b xy is he e ec size o x(gene exp ession) on y(slope o y eg essed on he gene ic alue o x), b zx is he e ec o zon x, and b zy is he e ec o zon y. The e o e b xy (b zy /b zx ) is he e ec o xon y. To dis inguish pleio opy om linkage whe e he op associa ed cis-eQTL is in LD wi h wo causal a ian s, one a ec ing gene exp ession he o he a ec ing ai , we es ed o he - e ogenei y in dependen ins umen s, using mul iple SNPs in each cis-eQTL egion. Unde he hypo hesis o pleio opy b xy alues o SNPs in LD wi h he causal a ian will be iden ical. Thus es ing agains he null hypo hesis ha he e is a single causal a ian is equi alen o es ing he e ogenei y in he b xy alues es i- ma ed o he SNPs in he cis-eQTL egion. Fo each p obe ha passed significance h eshold o he SMR es , we es ed he he e ogenei y in he b xy alues es ima ed o mul iple SNPs in he cis-eQTL egion using HEIDI. We used publicly a ailable LCL exp ession da a om he MuTHER (n=825)23 and GTEx conso ium (n=114)24. B iefly, GWAS summa y s a is ics files we e gene a ed om he me a-analysis. Re e ence files we e gene a ed om me ging 1000 Genomes P ojec phase 3 and UK10K (ALSPAC and TwinsUK) c s11,12.As p e iously ad oca ed, only p obes wi h a leas one eQTL P- alue o <5.0 × 10−8 we e conside ed o SMR analysis25. We se a h eshold o he SMR es o P SMR <5.49 × 10−4co esponding o a Bon e oni co ec ion o 91 es s (91 p obes wi h a op eQTL P<5.0 × 10−8ac oss he 12 loci and wo LCL eQTL da a se s). Fo all genes passing his h eshold, we gene a ed plo s o he eQTL and GWAS associa ions a he locus, as well as plo s o GWAS and eQTL e ec sizes (i.e., co esponding o inpu o he HEIDI he e ogenei y es ). HEIDI es P- alues <0.05 we e conside ed as being eflec i e o he e ogenei y. This h eshold is conse a i e o gene disco e y because i e ains ewe genes han when co ec ing o mul iple es ing. SMR plo s o significan eQTLs a e shown in Supplemen a y Fig. 5. Ch oma in s a e dynamics. En ichmen o cHL isk SNPs wi h DNAse and enhance s is conduc ed using Haplo eg 431. The o e lap o cHL isk SNPs wi h enhance s in GM12878 cell is compa ed o a backg ound model o all 1000 Genomes P ojec a ian s wi h a equency abo e 5% in any popula ion. The en ichmen ela i e o hese backg ound equencies was pe o med using a binomial es and a FDR P- alue was subsequen ly calcula ed; we conside ed an FDR <0.05 as being significan . To examine en ichmen in specific TF binding ac oss isk loci, we adap ed he a ian se en ichmen me hod o Cowpe -Sal la i e al.78. Fo each isk locus, a egion o s ong LD (defined as 2>0.8 and D′>0.8) was de e mined, and hese SNPs we e e med he associa ed a ian se (AVS). TF ChIP-seq uni o m peak da a we e ob ained om ENCODE o he GM12878 cell line, and included da a o 82 TFs. Fo each o hese ma ks, he o e lap o he SNP in he AVS and he binding si es was de e mined o p oduce a mapping ally. A null dis ibu ion was p oduced by andomly selec ing SNP wi h he same LD s uc u e (gene a ed om 1000 Genomes P ojec and UK10K da a) as he isk associa ed SNP, and he null mapping ally calcula ed. This p ocess was epea ed 10,000 imes, and app oxima e P- alues we e calcula ed as he p opo ion o pe mu a ions whe e he null mapping ally was g ea e o equal o he AVS mapping ally. An en ichmen sco e was calcula ed by he allies o he median o he null dis ibu ion. Thus he en ichmen sco e is he numbe o s anda d de ia ions o he AVS mapping ally om he mean o he null dis ibu ion allies. P omo e cap u e Hi-C da a. To map isk SNPs o in e ac ions in ol ing p o- mo e con ac s and iden i y genes in ol ed in cHL suscep ibili y, we analysed p omo e cap u e Hi-C da a on he LCL cell line GM12878 as a model B-cell26. Reads om echnical eplica es (E-MTAB-2323) we e combined be o e p ocessing wi h HiCUP79. Significan in e ac ions (i.e., sco e ≥5) on wo biological eplica es we e de e mined using CHiCAGO80. Co-he i abili y o Hodgkin lymphoma wi h au oimmune disease. We u ilised LD eg ession o es ima e gene ic co ela ion be ween indi idual au oimmune diseases and cHL, NSHL and MCHL35. Summa y s a is ics o published s udies o coeliac disease41, sys emic lupus e y hema osus37, p ima y bilia y ci hosis39, heuma oid a h i is36,MS 38 and UC40 we e downloaded om he ImmunoBase websi e (h p://www.immunobase.o g/). He i abili y analysis. We used genome-wide complex ai analysis o es ima e he polygenic a iance (i.e., he i abili y) asc ibable o all geno yped and impu ed GWAS SNPs33. SNPs we e excluded based on low MAF <0.01, poo impu a ion (in o sco e <0.9) and e idence o depa u e om HWE (P<0.05). Indi iduals we e excluded o poo impu a ion and whe e wo indi iduals we e closely ela ed. A gene ic ela ionship ma ix o pai s o samples was used as inpu o he es ic ed maximum likelihood analysis o es ima e he he i abili y explained by he selec ed se o SNPs. To ans o m he es ima ed he i abili y o he liabili y scale, we used he li e ime isk, o cHL, which is es ima ed o be 0.002 by SEER (h ps://see . cance .go /s a ac s/h ml/hodg.h ml). Da a a ailabili y. Geno ype da a ha suppo he findings o his s udy ha e been deposi ed in he Eu opean Genome-phenome A chi e (EGA) unde accession codes EGAD00000000022 and EGAD00000000024. Sequencing da a, which o ms he e e ence panel o impu a ion, ha e been deposi ed in he Eu opean Genome-phenome A chi e (EGA) unde accession codes EGAS00001000090, EGAD00001000195, EGAS00001000108. T ansc ip ional p ofiling da a om he MuTHER conso ium ha suppo he findings o his wo k ha e been deposi ed in he Eu opean Bioin o ma ics Ins i u e (Pa o he Eu opean Molecula Biology Labo a o y, EMBL-EBI) unde accession code E-TABM-1140. T ansc ip ional p ofiling da a om he Geno ype-Tissue Exp ession (GTEx) p ojec , ha suppo he findings o his wo k a e a ailable he e: h ps://www. g expo al.o g/ T ansc ip ion ac o binding da a ha suppo he findings o his wo k a e a ailable he e: h p://genome.ucsc.edu/ENCODE/downloads.h ml. P omo e cap u e Hi-C da a in GM12878 cells ha suppo he findings o his wo k ha e been deposi ed in he Eu opean Bioin o ma ics Ins i u e (Pa o he Eu opean Molecula Biology Labo a o y, EMBL-EBI) unde accession code E-MTAB-2323. The emaining da a con ained wi hin he pape and supplemen a y files a e a ailable om he au ho upon eques . Recei ed: 10 Janua y 2017 Accep ed: 20 June 2017 Re e ences 1. Kuppe s, R. The biology o Hodgkin’s lymphoma. Na . Re . Cance 9,15–27 (2009). 2. Smi h, A. e al. Lymphoma incidence, su i al and p e alence 2004-2014: sub- ype analyses om he UK’s Haema ological Malignancy Resea ch Ne wo k. B . J. Cance 112, 1575–1584 (2015). 3. Hjalg im, H. On he Ae iology o Hodgkin Lymphoma (S a ens Se um Ins i u , 2011). 4. Mack, T. M. e al. Conco dance o Hodgkin’s disease in iden ical wins sugges ing gene ic suscep ibili y o he young-adul o m o he disease. N. Engl. J. Med. 332, 413–419 (1995). 5. Kha azmi, E. e al. Risk o amilial classical Hodgkin lymphoma by ela ionship, his ology, age, and sex: a join s udy om fi e No dic coun ies. Blood 126, 1990–1995 (2015). 6. Cozen, W. e al. A me a-analysis o Hodgkin lymphoma e eals 19p13.3 TCF3 as a no el suscep ibili y locus. Na . Commun.5, 3856 (2014). 7. Enciso-Mo a, V. e al. A genome-wide associa ion s udy o Hodgkin’s lymphoma iden ifies new suscep ibili y loci a 2p16.1 (REL), 8q24.21 and 10p14 (GATA3). Na . Gene . 42, 1126–1130 (2010). 8. F amp on, M. e al. Va ia ion a 3p24.1 and 6q23.3 influences he isk o Hodgkin’s lymphoma. Na . Commun. 4, 2549 (2013). 9. U ayama, K. Y. e al. Genome-wide associa ion s udy o classical Hodgkin lymphoma and Eps ein–Ba i us s a us–defined subg oups. J. Na l. Cance Ins . 104, 240–253 (2012). 10. Thomsen, H. e al. He i abili y es ima es on Hodgkin/‘s lymphoma: a genomic- e sus popula ion-based app oach. Eu . J. Hum. Gene . 23, 824–830 (2015). 11. The 1000 Genomes P ojec Conso ium. A map o human genome a ia ion om popula ion-scale sequencing. Na u e 467, 1061–1073 (2010). 12. Huang, J. e al. Imp o ed impu a ion o low- equency and a e a ian s using he UK10K haplo ype e e ence panel. Na . Commun.6, 8111 (2015). 13. Fla ell, J. R. e al. Down- egula ion o he TGF-be a a ge gene, PTPRK, by he Eps ein-Ba i us encoded EBNA1 con ibu es o he g ow h and su i al o Hodgkin lymphoma cells. Blood 111, 292–301 (2008). NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 ARTICLE NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions 7 14. Jiang, X. e al. De egula ed exp ession in Ph + human leukemias o AHI-1, a gene ac i a ed by inse ional mu agenesis in mouse models o leukemia. Blood 103, 3897–3904 (2004). 15. Kennah, E. e al. Iden ifica ion o y osine kinase, HCK, and umo supp esso , BIN1, as po en ial media o s o AHI-1 oncogene in p ima y and ans o med CTCL cells. Blood 113, 4646–4655 (2009). 16. Chan, W.-K. e al. A UPF3-media ed egula o y swi ch ha main ains RNA su eillance. Na . S uc . Mol. Biol. 16, 747–753 (2009). 17. Sch eibe , A. e al. S uc u al basis o he subuni assembly o he anaphase- p omo ing complex. Na u e 470, 227–232 (2011). 18. S eidl, C. e al. Gene exp ession p ofiling o mic odissec ed Hodgkin Reed- S e nbe g cells co ela es wi h ea men ou come in classical Hodgkin lymphoma. Blood 120, 3530–3540 (2012). 19. Li, J. e al. Associa ion o CLEC16A wi h human common a iable immunodeficiency diso de and ole in mu ine B cells. Na . Commun.6, 6804 (2015). 20. Bane jee, A., No h up, D., Bouka abila, H., Jacobsen, S. E. & Allman, D. T ansc ip ional ep ession o Ga a3 is essen ial o ea ly B cell commi men . Immuni y 38, 930–942 (2013). 21. S anelle, J., Dö ing, C., Hansmann, M.-L. & Küppe s, R. Mechanisms o abe an GATA3 exp ession in classical Hodgkin lymphoma and i s consequences o he cy okine p ofile o Hodgkin and Reed/S e nbe g cells. Blood 116, 4202–4211 (2010). 22. Vijayak ishnan, J. e al. A genome-wide associa ion s udy iden ifies isk loci o childhood acu e lymphoblas ic leukemia a 10q26.13 and 12q23.1. Leukemia 31, 573–579 (2017). 23. G undbe g, E. e al. Mapping cis- and ans- egula o y e ec s ac oss mul iple issues in wins. Na . Gene . 44, 1084–1089 (2012). 24. Lonsdale, J. e al. The Geno ype-Tissue Exp ession (GTEx) p ojec . Na . Gene . 45, 580–585 (2013). 25. Zhu, Z. e al. In eg a ion o summa y da a om GWAS and eQTL s udies p edic s complex ai gene a ge s. Na . Gene . 48, 481–487 (2016). 26. Mi sud, B. e al. Mapping long- ange p omo e con ac s in human cells wi h high- esolu ion cap u e Hi-C. Na . Gene . 47, 598–606 (2015). 27. Ramachand a eddy, H. e al. BCL6 p omo e in e ac s wi h a ups eam sequences wi h g ea ly enhanced ac i a ing his one modifica ions in ge minal cen e B cells. P oc. Na l Acad. Sci. USA 107, 11930–11935 (2010). 28. Schneide , C. e al. mic oRNA 28 con ols cell p oli e a ion and is down- egula ed in B-cell lymphomas. P oc. Na l Acad. Sci. USA 111, 8185–8190 (2014). 29. Ha , D. O., San a, M. K., Raha, T. & G een, M. R. Selec i e in e ac ion be ween T 3 and Ta 3 equi ed o ea ly de elopmen and hema opoiesis. De . Dyn. 238, 2540–2549 (2009). 30. Singh, T. R. e al. BLAP18/RMI2, a no el OB- old-con aining p o ein, is an essen ial componen o he Bloom helicase–double Holliday junc ion dissol asome. Genes De . 22, 2856–2868 (2008). 31. Wa d, L. D. & Kellis, M. HaploReg 4: sys ema ic mining o pu a i e causal a ian s, cell ypes, egula o s and a ge genes o human complex ai s and disease. Nucleic Acids Res. 44, D877–D881 (2016). 32. Jia, X. e al. Impu ing amino acid polymo phisms in human leukocy e an igens. PLoS ONE 8, e64683 (2013). 33. Yang, J., Lee, S. H., Godda d, M. E. & Vissche , P. M. GCTA: a ool o genome- wide complex ai analysis. Am. J. Hum. Gene . 88,76–82 (2011). 34. K is insson, S. Y. e al. Au oimmuni y and isk o Hodgkin’s lymphoma by sub ype. Haema ologica 94, 1468–1469 (2009). 35. Bulik-Sulli an, B. K. e al. LD sco e eg ession dis inguishes con ounding om polygenici y in genome-wide associa ion s udies. Na . Gene . 47,291–295 (2015). 36. Okada, Y. e al. Gene ics o heuma oid a h i is con ibu es o biology and d ug disco e y. Na u e 506, 376–381 (2014). 37. Ben ham, J. e al. Gene ic associa ion analyses implica e abe an egula ion o inna e and adap i e immuni y genes in he pa hogenesis o sys emic lupus e y hema osus. Na . Gene . 47, 1457–1464 (2015). 38. Sawce , S. e al. Gene ic isk and a p ima y ole o cell-media ed immune mechanisms in mul iple scle osis. Na u e 476, 214–219 (2011). 39. Co dell, H. J. e al. In e na ional genome-wide me a-analysis iden ifies new p ima y bilia y ci hosis isk loci and a ge able pa hogenic pa hways. Na . Commun. 6, 8019 (2015). 40. Ande son, C. A. e al. Me a-analysis iden ifies 29 addi ional ulce a i e coli is isk loci, inc easing he numbe o confi med associa ions o 47. Na . Gene . 43, 246–252 (2011). 41. Dubois, P. C. A. e al. Mul iple common a ian s o celiac disease influencing immune gene exp ession. Na . Gene . 42, 295–302 (2010). 42. Chen, F. E., Huang, D. B., Chen, Y. Q. & Ghosh, G. C ys al s uc u e o p50/p65 he e odime o ansc ip ion ac o NF-kappaB bound o DNA. Na u e 391, 410–413 (1998). 43. Izban, K. F. e al. Cha ac e iza ion o NF-kappaB exp ession in Hodgkin’s disease: inhibi ion o cons i u i ely exp essed NF-kappaB esul s in spon aneous caspase-independen apop osis in Hodgkin and Reed-S e nbe g cells. Mod. Pa hol. 14, 297–310 (2001). 44. Ba gou, R. C. e al. Cons i u i e nuclea ac o -kappaB-RelA ac i a ion is equi ed o p oli e a ion and su i al o Hodgkin’s disease umo cells. J. Clin. In es . 100, 2961–2969 (1997). 45. Lin, Y. C. e al. A global ne wo k o ansc ip ion ac o s, in ol ing E2A, EBF1 and Foxo1, ha o ches a es B cell a e. Na . Immunol. 11, 635–643 (2010). 46. Tiacci, E. e al. Analyzing p ima y Hodgkin and Reed-S e nbe g cells o cap u e he molecula and cellula pa hogenesis o classical Hodgkin lymphoma. Blood 120, 4609–4620 (2012). 47. He el, C. B., Zhou, X. G., Hamil on-Du oi , S. J. & Junke , S. Loss o B cell iden i y co ela es wi h loss o B cell-specific ansc ip ion ac o s in Hodgkin/ Reed-S e nbe g cells o classical Hodgkin lymphoma. Oncogene 21, 4908–4920 (2002). 48. Whi eman, H. J. & Fa ell, P. J. RUNX exp ession and unc ion in human B cells. C i . Re . Euka yo . Gene Exp . 16,31–44 (2006). 49. B ady, G., Whi eman, H. J., Spende , L. C. & Fa ell, P. J. Down egula ion o RUNX1 by RUNX3 Requi es he RUNX3 VWRPY sequence and is essen ial o Eps ein-Ba i us-d i en B-cell p oli e a ion. J. Vi ol. 83, 6909–6916 (2009). 50. Be z, B. C. e al. Ba coo dina es mul iple aspec s o B and T cell unc ion equi ed o no mal an ibody esponses. J. Exp. Med. 207, 933–942 (2010). 51. Lo enzo, Y. e al. Di e en ial gene ic and unc ional ma ke s o second neoplasias in Hodgkin’s disease pa ien s. Clin. Cance Res. 15,4823–4828 (2009). 52. Mu ay, P. G., Cons andinou, C. M., C ocke , J., Young, L. S. & Ambinde , R. F. Analysis o majo his ocompa ibili y complex class I, TAP exp ession, and LMP2 epi ope sequence in Eps ein-Ba i us–posi i e Hodgkin’s disease. Blood 92, 2477–2483 (1998). 53. Lenz, T. L. e al. Widesp ead non-addi i e and in e ac ion e ec s wi hin HLA loci modula e he isk o au oimmune diseases. Na . Gene . 47, 1085–1090 (2015). 54. P a ama, A. & Vinuesa, C. G. Con ol o TFH cell numbe s: why and how? Immunol. Cell. Biol. 92,40–48 (2014). 55. Sha on, E. e al. Gene ic a ia ion in MHC p o eins is associa ed wi h T cell ecep o exp ession biases. Na . Gene . 48, 995–1002 (2016). 56. G ea es, P. e al. Defining cha ac e is ics o classical Hodgkin lymphoma mic oen i onmen T-helpe cells. Blood 122, 2856–2863 (2013). 57. Dieps a, A. e al. HLA class II exp ession by Hodgkin Reed-S e nbe g Cells is an independen p ognos ic ac o in classical Hodgkin’s lymphoma. J. Clin. Oncol. 25, 3101–3108 (2007). 58. Hamme , C. e al. Amino acid a ia ion in HLA class II p o eins is a majo de e minan o humo al esponse o common i uses. Am. J. Hum. Gene . 97, 738–743 (2015). 59. Apple, R. J. & E lich, H. A. Two new HLA DRB1 alleles ound in A ican Ame icans: implica ions o balancing selec ion a posi ions 57 and 86. Tissue An igens 40,69–74 (1992). 60. Busch, R., Hill, C. M., Hayball, J. D., Lamb, J. R. & Ro hba d, J. B. E ec o na u al polymo phism a esidue 86 o he HLA-DR be a chain on pep ide binding. J. Immunol. 147, 1292–1298 (1991). 61. Khankhanian, P. e al. Me a-analysis o genome-wide associa ion s udies e eals gene ic o e lap be ween Hodgkin lymphoma and mul iple scle osis. In . J. Epidemiol. 45, 728–740 (2016). 62. The Wellcome T us Case Con ol Conso ium. Genome-wide associa ion s udy o 14,000 cases o se en common diseases and 3,000 sha ed con ols. Na u e 447, 661–678 (2007). 63. Ande son, C. A. e al. Da a quali y con ol in gene ic case-con ol associa ion s udies. Na . P o oc. 5, 1564–1573 (2010). 64. Howie, B. N., Donnelly, P. & Ma chini, J. A flexible and accu a e geno ype impu a ion me hod o he nex gene a ion o genome-wide associa ion s udies. PLoS Gene . 5, e1000529 (2009). 65. 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, 906–913 (2007). 66. Clay on, D. G. e al. Popula ion s uc u e, di e en ial bias and genomic con ol in a la ge-scale, case-con ol associa ion s udy. Na . Gene . 37, 1243–1246 (2005). 67. Danecek, P. e al. The a ian call o ma and VCF ools. Bioin o ma ics 27, 2156–2158 (2011). 68. Scales, M., Jäge , R., Miglio ini, G., Houls on, R. S. & Hen ion, M. Y. R. isPIG - a web ool o p oducing mul i- egion, mul i- ack, mul i-scale plo s o gene ic da a. PLoS ONE 9, e107497 (2014). 69. Alexande , F. E. e al. Risk ac o s o Hodgkin’s disease by Eps ein-Ba i us (EBV) s a us: p io in ec ion by EBV and o he agen s. B . J. Cance 82, 1117–1121 (2000). 70. Ja e , R. F. e al. The Sco land and Newcas le epidemiological s udy o Hodgkin’s disease: impac o his opa hological e iew and EBV s a us on incidence es ima es. J. Clin. Pa hol. 56, 811–816 (2003). 71. Wille , E. V. & Roman, E. Obesi y and he isk o Hodgkin lymphoma (Uni ed Kingdom). Cance Causes Con ol 17, 1103–1106 (2006). ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 8NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions 72. Lake, A. e al. Mu a ions o NFKBIA, encoding IkappaB alpha, a e a ecu en finding in classical Hodgkin lymphoma bu a e no a uni ying ea u e o non-EBV-associa ed cases. In . J. Cance 125, 1334–1342 (2009). 73. Hjalg im, H. & Engels, E. A. In ec ious ae iology o Hodgkin and non-Hodgkin lymphomas: a e iew o he epidemiological e idence. J. In e n. Med. 264, 537–548 (2008). 74. Liu, J. Z. e al. Me a-analysis and impu a ion efines he associa ion o 15q25 wi h smoking quan i y. Na . Gene . 42, 436–440 (2010). 75. Higgins, J. P. & Thompson, S. G. Quan i ying he e ogenei y in a me a-analysis. S a . Med. 21, 1539–1558 (2002). 76. Tho lund, K. e al. E olu ion o he e ogenei y (I2) es ima es and hei 95% confidence in e als in la ge me a-analyses. PLoS ONE 7, e39471 (2012). 77. Skol, A. D., Sco , L. J., Abecasis, G. R. & Boehnke, M. Join analysis is mo e e ficien han eplica ion-based analysis o wo-s age genome-wide associa ion s udies. Na . Gene . 38, 209–213 (2006). 78. Cowpe -Salla i, R. e al. B eas cance isk-associa ed SNPs modula e he a fini y o ch oma in o FOXA1 and al e gene exp ession. Na . Gene . 44, 1191–1198 (2012). 79. Winge , S. e al. HiCUP: pipeline o mapping and p ocessing Hi-C da a. F1000Res 4, 1310 (2015). 80. Cai ns, J. e al. CHiCAGO: obus de ec ion o DNA looping in e ac ions in cap u e Hi-C da a. Genome Biol.17, 127 (2016). Acknowledgemen s In he Uni ed Kingdom, Bloodwise (LLR; 10021) p o ided p incipal unding o he s udy. Suppo om Cance Resea ch UK (C1298/A8362 suppo ed by he Bobby Moo e Fund) and he Lymphoma Resea ch T us is also acknowledged. A.S. is suppo ed by a clinical ellowship om Cance Resea ch UK. Fo he UK-GWAS, sample and da a acquisi ion we e suppo ed by B eas Cance Now, he Eu opean Union and he Lym- phoma Resea ch T us . The UK-GWAS made use o con ol geno yping da a gene a ed by he WTCCC. We acknowledge use o geno ype da a om he B i ish 1958 Bi h Coho DNA collec ion, which was unded by he Medical Resea ch Council G an G0000934 and he Wellcome T us G an 068545/Z/02. A ull lis o he in es iga o s who con ibu ed o he gene a ion o he da a is a ailable om h p://www.w ccc.o g.uk. Funding o his p ojec was p o ided by he Wellcome T us unde awa ds 076113 and 085475. Pa ien s o he new GWAS we e asce ained h ough he Na ional S udy o Hodgkin Lymphoma Gene ics (h p://www.public.ukc n.o g.uk) and we hank he High- Th oughpu Genomics G oup a he Wellcome T us Cen e o Human Gene ics ( unded by Wellcome T us g an e e ence 090532/Z/09/Z) o he gene a ion o Gen- o yping da a. The BCAC s udy would no ha e been possible wi hou he con ibu ions o he ollowing: Manjee K. Bolla, Qin Wang, Ky iaki Michailidou and Joe Dennis. BCAC is unded by Cance Resea ch UK (C1287/A10118, C1287/A16563). Fo he BBCS s udy, we hank Eileen Williams, Elaine Ryde -Mills, Ka a Sa gus. The BBCS is unded by Cance Resea ch UK and B eas Cance Now and acknowledges NHS unding o he Na ional Ins i u e o Heal h Resea ch (NIHR) Biomedical Resea ch Cen e (BRC) and he Na ional Cance Resea ch Ne wo k (NCRN). We hank he pa icipan s and he in es iga o s o EPIC (Eu opean P ospec i e In es iga ion in o Cance and Nu i ion). The coo dina ion o EPIC is financially suppo ed by he Eu opean Commission (DG- SANCO) and he In e na ional Agency o Resea ch on Cance . The na ional coho s a e suppo ed by: Ligue Con e le Cance , Ins i u Gus a e Roussy, Mu uelle Géné ale de l’Educa ion Na ionale, Ins i u Na ional de la San é e de la Reche che Médicale (INSERM) (F ance); Ge man Cance Aid, Ge man Cance Resea ch Cen e (DKFZ), Fede al Minis y o Educa ion and Resea ch (BMBF) (Ge many); he Hellenic Heal h Founda ion, he S a os Nia chos Founda ion (G eece); Associazione I aliana pe la Rice ca sul Canc o-AIRC-I aly and Na ional Resea ch Council (I aly); Du ch Minis y o Public Heal h, Wel a e and Spo s (VWS), Ne he lands Cance Regis y (NKR), LK Resea ch Funds, Du ch P e en ion Funds, Du ch ZON (Zo g Onde zoek Nede land), Wo ld Cance Resea ch Fund (WCRF), S a is ics Ne he lands (The Ne he lands); Heal h Resea ch Fund (FIS), PI13/00061 o G anada, PI13/01162 o EPIC-Mu cia, Regional Go e nmen s o Andalucía, As u ias, Basque Coun y, Mu cia and Na a a, ISCIII RETIC (RD06/0020) (Spain); Cance Resea ch UK (14136 o EPIC-No olk; C570/ A16491 and C8221/A19170 o EPIC-Ox o d), Medical Resea ch Council (1000143 o EPIC-No olk, MR/M012190/1 o EPIC-Ox o d) (Uni ed Kingdom). We hank he SEARCH and EPIC eams, which we e unded by a p og amme g an om Cance Resea ch UK (C490/A10124) and suppo ed by he UK NIHR BRC a he Uni e si y o Camb idge. We hank B eas Cance Now and he Ins i u e o Cance Resea ch (ICR) o suppo and unding o he UKBGS, and he s udy pa icipan s, s udy s a , and he doc o s, nu ses and o he heal h-ca e p o ide s and heal h in o ma ion sou ces who ha e con ibu ed o he s udy. We acknowledge NHS unding o he Royal Ma sden/ICR NIHR BRC. UKGPCS would like o hank The Ins i u e o Cance Resea ch and The E e yman Campaign o unding suppo . The UKGPCS acknowledges The P os a e Cance Resea ch Founda ion, P os a e Ac ion, The O chid Cance Appeal, The Na ional Cance Resea ch Ne wo k UK, The Na ional Cance Resea ch Ins i u e (NCRI), he NIHR unding o he NIHR Biomedical Resea ch da a manage s and consul an s o hei wo k in he UKGPCS s udy and u ologis s and o he pe sons in ol ed in he planning, and da a collec ion o he CAPS s udy. Geno yping o he OncoA ay was unded by he US Na ional Ins i u es o Heal h (NIH) (U19 CA 148537 o ELucida ing Loci In ol ed in P os a e cance SuscEp ibili y (ELLIPSE) p ojec and X01HG007492 o he Cen e o Inhe i ed Disease Resea ch (CIDR) unde con ac numbe HHSN268201200008I). Addi ional analy ic suppo was p o ided by NIH NCI U01 CA188392 (PI: Schumache ). The PRACTICAL conso ium was suppo ed by Cance Resea ch UK G an s C5047/ A7357, C1287/A10118, C1287/A16563, C5047/A3354, C5047/A10692, C16913/A6135, Eu opean Commission's Se en h F amewo k P og amme g an ag eemen no. 223175 (HEALTH-F2-2009-223175), and The Na ional Ins i u e o Heal h (NIH) Cance Pos - Cance GWAS ini ia i e g an : No. 1 U19 CA 148537-01 ( he GAME-ON ini ia i e). We would also like o hank he ollowing o unding suppo : The Ins i u e o Cance Resea ch and The E e yman Campaign, The P os a e Cance Resea ch Founda ion, P os a e Resea ch Campaign UK (now P os a e Ac ion), The O chid Cance Appeal, The Na ional Cance Resea ch Ne wo k UK, The Na ional Cance Resea ch Ins i u e (NCRI) UK. We a e g a e ul o suppo o NIHR unding o he NIHR Biomedical Resea ch Cen e a The Ins i u e o Cance Resea ch and The Royal Ma sden NHS Founda ion T us . The APBC BioResou ce, which o m pa o he PRACTICAL conso ium, consis s o he ollowing membe s: Wayne Tilley, Gail Risb idge , Renea Taylo , Judi h A Clemen s, Lisa Ho a h, Vanessa Hayes, Lisa Bu le , T ina Yeadon, Allison Ecke , Pamela Saunde s, Anne-Ma ee Haynes, Melissa Papa gi is. A he MRC Uni e si y o Glasgow Cen e o Vi us Resea ch, unding was p o ided by Leukaemia Lymphoma Resea ch (12022). The Sco land and Newcas le Epidemiological S udy o Hodgkin Dis- ease (SNEHD) was unded by he Kay Kendall Leukaemia Fund and he Young Adul Hodgkin Case–Con ol S udy (YHCCS) and he Epidemiology and Cance S a is ics G oup Lymphoma Case–Con ol S udy (ELCCS) we e unded by Bloodwise. Ge man unding was p o ided by he Ge man Cance Aid, he Ha ald Huppe Founda ions, The Ge man Fede al Minis y o Educa ion and Resea ch (eMed, Cliommics 01ZX1309B), he Mul iple Myeloma Resea ch Founda ion, he Heinz Nixdo Founda ion (Ge many), he Minis e ium ü Inno a ion, Wissenscha und Fo schung des Landes No d hein- Wes alen and he Facul y o Medicine Uni e si y Duisbu g–Essen. Fo hei help wi h UK sample collec ion we hank Hayley E ans, James G i fin, Joanne Micic, Susan Blackmo e, Be e ley Smi h, Debo ah Hogben, Alison Bu lin, Jill Wood, Ma go Pele in, Alison Ha , Ka a zyna Tomczyk and Sa ah Chilco -Bu ns. Finally, we a e g a e ul o all he pa ien s and indi iduals o hei pa icipa ion and he clinicians, in es iga o s, o he s a who con ibu ed o sample and da a collec ion. Au ho con ibu ions A.S., R.S.H. and K.H. designed and p o ided o e all p ojec managemen . A.S. and R.S.H. d a ed he manusc ip . In he UK, R.S.H. and A.S. pe o med da abase de elopmen and o e saw labo a o y analyses; R.H.A.J.S. and N.O. p o ided samples o UK-GWAS and UK-NSHLG-GWAS, L.W. and R.C. p o ided da a on samples o UK-GWAS and UK- NSHLG-GWAS. D.E., P.P., A.D., J.P., F.C., R.E., Z.K.-J, K.M. and N.P. p o ided con ol samples o he UK-NSHLG-GWAS. A.S. pe o med bioin o ma ic and s a is ical ana- lysis. In he UK P.L., G.O. and O.L. pe o med addi ional bioin o ma ic analyses. P.B. and A.S. pe o med sample and labo a o y coo dina ion. A.H. pe o med alida ion geno yping. Fo UK Replica ion 2, A.L. and D.M. p epa ed samples, T.L., E.K., E.R., D.M. and R.F.J. p o ided samples and da a. In Ge many, A.F., H.T. and M.I.d.S.F. pe o med bioin o ma ic and s a is ical analyses; P.H. and M.M.N. we e esponsible o Ge man- GWAS analysis; K.-H.J. p o ided he Ge man con ol samples; E.P. .S. and A.E. we e esponsible o Ge man cHL pa ien s. All au ho s con ibu ed o he final pape . Addi ional in o ma ion Supplemen a y In o ma ion accompanies his pape a doi:10.1038/s41467-017-00320-1. Compe ing in e es s: The au ho s decla e no compe ing financial in e es s. Rep in s and pe mission in o ma ion is a ailable online a h p://npg.na u e.com/ ep in sandpe missions/ Publishe 's no e: Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a filia ions. Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License, which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his a icle a e included in he a icle’s C ea i e Commons license, unless indica ed o he wise in a c edi line o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons license and you in ended use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om he copy igh holde . To iew a copy o his license, isi h p://c ea i ecommons.o g/ licenses/by/4.0/. © The Au ho (s) 2017 NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-00320-1 ARTICLE NATURE COMMUNICATIONS |8: 1892 |DOI: 10.1038/s41467-017-00320-1 |www.na u e.com/na u ecommunica ions 9