Gene and pa hway le el analyses o
ge mline DNA- epai gene a ian s and
p os a e cance suscep ibili y using he
iCOGS-geno yping a ay
Edwa d J Saunde s
1
, Tokhi Dadae
1
, Daniel A Leongamo nle
1
, Ali Amin Al Olama
2
, Sa a Benlloch
2
,
G aham G Giles
3,4,29
, F ed ik Wiklund
5,29
, Hen ik G o
¨nbe g
5,29
, Ch is ophe A Haiman
6,29
,
Johanna Schleu ke
7,8,29
, Bø ge G No des gaa d
9,29
, Ru h C T a is
10,29
, Da id Neal
11,29
, No a Pasayan
12,29
,
Kay-Tee Khaw
13,29
, Jane L S an o d
14,29
, William J Blo
15,29
, S ephen N Thibodeau
16,29
, Ch is iane Maie
17,29
,
Adam S Kibel
18,29
, Ceza y Cybulski
19,29
, Lisa Cannon-Alb igh
20,29
, He mann B enne
21,29
, Jong Y Pa k
22,29
,
Radka Kane a
23,29
, Jyo sna Ba a
24,29
, Manuel R Teixei a
25,26,29
, Ha de Pandha
27,29
, Ko eela Go indasami
1
,
Ken Mui
28
, The UK Gene ic P os a e Cance S udy Collabo a o s
30
, The UK P o ecT S udy Collabo a o s
30
,
The PRACTICAL Conso ium
30
, Douglas F Eas on
2
, Rosalind A Eeles
1,31
and Zso ia Ko e-Ja ai*
,1,31
Backg ound: Ge mline mu a ions wi hin DNA- epai genes a e implica ed in suscep ibili y o mul iple o ms o cance . Fo
p os a e cance (P Ca), a e mu a ions in BRCA2 and BRCA1 gi e ise o mode a ely ele a ed isk, whe eas wo o B100 common,
low-pene ance P Ca suscep ibili y a ian s iden i ied so a by genome-wide associa ion s udies implica e RAD51B and RAD23B.
Me hods: Geno ype da a om he iCOGS a ay we e impu ed o he 1000 genomes phase 3 e e ence panel o 21 780 P Ca cases
and 21 727 con ols om he P os a e Cance Associa ion G oup o In es iga e Cance Associa ed Al e a ions in he Genome
(PRACTICAL) conso ium. We subsequen ly pe o med single a ian , gene and pa hway-le el analyses using 81 303 SNPs wi hin
20 Kb o a panel o 179 DNA- epai genes.
Resul s: Single SNP analyses iden i ied only he p e iously epo ed associa ion wi h RAD51B. Gene-le el analyses using he SKAT-C
es om he SNP-se (Sequence) Ke nel Associa ion Tes (SKAT) iden i ied a signi ican associa ion wi h P Ca o MSH5. Pa hway-
le el analyses sugges ed a possible ole o he anslesion syn hesis pa hway in P Ca isk and Homologous ecombina ion/Fanconi
Anaemia pa hway o P Ca agg essi eness, e en hough a e adjus men o mul iple es ing hese did no emain signi ican .
Conclusions: MSH5 is a no el candida e gene wa an ing addi ional ollow-up as a p ospec i e P Ca- isk locus. MSH5 has
p e iously been epo ed as a pleio opic suscep ibili y locus o lung, colo ec al and se ous o a ian cance s.
P os a e Cance (P Ca) is he mos equen ly diagnosed cance
among men in de eloped coun ies and despi e high su i al a es
also one o he highes o mo ali y (Cance Resea ch UK, 2014;
Qua esma e al, 2015). Howe e , as he majo i y o p os a e
neoplasms de elop ex emely slowly, many do no equi e clinical
in e en ion, which coupled wi h he low speci ici y o he
*Co espondence: D Z Ko e-Ja ai; E-mail: [email p o ec ed]
29
These au ho s con ibu ed equally o his wo k.
30
Fo a ull lis o conso ia membe s, see Supplemen a y No e.
31
Join senio au ho s.
Recei ed 20 No embe 2015; e ised 5 Feb ua y 2016; accep ed 9 Feb ua y 2016;
published online 10 Ma ch 2016
&2016 Cance Resea ch UK. All igh s ese ed 0007 – 0920/16
FULL PAPER
Keywo ds: DNA epai ; p os a e cance ; genome-wide associa ion s udy; GWAS; iCOGS
B i ish Jou nal o Cance (2016) 114, 945–952 | doi: 10.1038/bjc.2016.50
www.bjcance .com | DOI:10.1038/bjc.2016.50 945
p os a e-speci ic an igen es o clinically ele an o ms o he
disease could po en ially lead o conside able o e -diagnosis and
o e ea men o pa ien s o ela i ely modes educ ions in
mo ali y (Ilic e al, 2013). In conjunc ion wi h he es ablishmen o
imp o ed bioma ke s o le hal P Ca, he iden i ica ion o
indi iduals a g ea e isk o de eloping p os a e umou s ha
equi e clinical in e en ion would also help in o m mo e a ge ed
and app op ia e applica ion o ea men . The he i abili y o P Ca
is belie ed o be he highes o all he common o ms o cance
(Hjelmbo g e al, 2014). This is consis en wi h obse a ions om
genome-wide associa ion s udies (GWAS), which ha e o da e
iden i ied 4100 low-pene ance suscep ibili y a ian s o P Ca,
wo o which implica e he DNA- epai genes RAD51B and
RAD23B (Xu e al, 2012; Al Olama e al, 2014; Eeles e al, 2014;
Amin Al Olama e al, 2015). In addi ion, a e ge mline mu a ions
in a small numbe o genes ha e been epo ed, wi h a ying
deg ees o e idence, as po en ially con e ing g ea e isks o P Ca,
including he DNA- epai genes ATM,BRCA1,BRCA2,BRIP1,
CHEK2 and NBN (Dong e al, 2003; Ko e-Ja ai e al, 2009, 2011;
Leongamo nle e al, 2012, 2014; Robinson e al, 2015). Recen ly,
inc easing e idence has demons a ed ha hese ge mline DNA-
epai gene mu a ion ca ie s a e a inc eased likelihood o
expe iencing ad anced disease, me as a ic sp ead and poo e
su i al ou come; ye hese mu a ions also hold p omise as
po en ially clinically ac ionable and esponsi e o a ge ed ea -
men s (Cas o e al, 2013; Cybulski e al, 2013; Leongamo nle
e al, 2014; Robinson e al, 2015). In spi e o hese disco e ies, he
majo i y o he excess amilial isk o P Ca s ill emains o be
explained (A a d e al, 2015), wi h he con ibu ion o DNA- epai
gene a ian s iden i ied o da e making hem a ac i e candida es
o u he in es iga ion. In his s udy, using da a om he iCOGS
p ojec impu ed o he 1000 Genomes Phase 3 e e ence panel, we
ha e analysed a la ge panel o DNA- epai gene a ian s o 21 780
P Ca cases and 21 727 con ols o Eu opean ances y om he
P os a e Cance Associa ion G oup o In es iga e Cance Asso-
cia ed Al e a ions in he Genome (PRACTICAL) Conso ium
(Eeles e al, 2013). Analyses we e pe o med a single a ian , gene
and pa hway le els o maximise he powe o de ec pu a i e
associa ions wi h lowe equency a ian s o hose wi h modes
e ec sizes.
MATERIALS AND METHODS
Samples. Samples o he iCOGS s udy we e d awn om
25 s udies pa icipa ing in he PRACTICAL Conso ium. The
majo i y o s udies we e popula ion-based o hospi al-based
case–con ol s udies, o nes ed case–con ol s udies; some s udies
selec ed samples by age o o e sampled o cases wi h a amily
his o y o p os a e cance . Fu he in o ma ion ega ding he
samples om he PRACTICAL Conso ium included on
he iCOGS a ay may be ound wi hin he o iginal publica ion
(Eeles e al, 2013). Analyses o DNA- epai gene a ian s we e
es ic ed o samples o Eu opean ances y. In o al, geno ype da a
o 21 780 P Ca cases and 21 727 ma ched con ols we e a ailable
a e quali y con ol.
Geno yping and impu a ion. Geno yping was pe o med as pa
o he iCOGS p ojec . This u ilised a cus om geno yping a ay
designed in collabo a ion be ween he PRACTICAL, BCAC (B eas
Cance Associa ion Conso ium), OCAC (O a ian Cance Asso-
cia ion Conso ium) and CIMBA (Conso ium o In es iga o s o
Modi ie s o BRCA1/2) conso ia. De ailed in o ma ion abou he
design, geno yping and quali y con ol p ocedu es o iCOGS can
be ound wi hin he o iginal publica ion (Eeles e al, 2013). In o al
211 155 SNPs we e geno yped on he iCOGS a ay, o which 3510
we e si ua ed wi hin ou de ined DNA- epai gene egions.
Impu a ion o he iCOGS PRACTICAL da a was pe o med based
on sequence da a o 2504 samples om he 1000 Genomes
phase 3 e e ence panel (IMPUTE2 haplo ype panel, Oc obe
2014 elease; h ps://ma hgen.s a s.ox.ac.uk/impu e/1000GP%20
Phase%203%20haplo ypes%206%20Oc obe %202014.h ml) using
SHAPEIT ( 2 778) and IMPUTE 2.3.1 in 588 chunks wi h a
median size o 5 Mb (Howie e al, 2009; Delaneau e al, 2013).
Impu ed da a o a ian s wi h INFO sco es X0.3 and
MAF 40.001 we e included in hese analyses, which e ained
a o al o 81 303 a ian s wi hin he s udied DNA- epai
gene egions.
Gene/ egion selec ion. We iden i ied a o al o 179 genes wi h
aco e unc ioninDNA-damage epai om heli e a u e
ha in e sec ed impu ed iCOGS geno ype da a. We anno a ed
DNA- epai genes o a single p ima y DNA- epai pa hway
acco ding o p e ious cu a ions (Wood e al, 2005; Kang e al,
2012). The genes analysed in his s udy ep esen he pa hways
Homologous ecombina ion/Fanconi Anaemia signalling ne -
wo k (HR/FA), base excision epai (BER), non-homologous end
joining (NHEJ), misma ch epai (MMR), nucleo ide excision
epai (NER), anslesion syn hesis (TLS), ATM signalling
(ATM), RECQ helicase amily (RECQ), c oss-link epai
(XLR), and o he miscellaneous DNA- epai genes wi h unc-
ions including endonuclease/exonuclease ac i i y and modi ica-
ion o ch oma in s uc u e (O he ). Gene coo dina es we e
assigned acco ding o GENCODE elease 19 (GRCh37.p13),
wi h a 20-kb lank added o de ine he s udy egion o each gene,
in o de o ocus p ima ily on cap u ing gene and p omo e
cen ic a ia ion o e ha wi hin egula o y elemen s, which can
be loca ed a a iable and po en ially ela i ely la ge dis ances
om he gene i sel . Va ian s we e anno a ed using wANNO-
VAR o acili a e designa ion as coding, in onic, UTR,
splice and in e genic (Wang e al, 2010; Chang and Wang,
2012). A ull lis o he DNA- epai genes analysed in his s udy,
hei pa hway anno a ions, egion coo dina es and he numbe
o yped and impu ed a ian s a ailable is included in
Supplemen a y Table 1.
S a is ical analyses. Analyses we e adjus ed o s udy g oups
and he i s eigh p incipal componen s. Fo single-SNP
analyses he genome-wide signi icance h eshold was employed
(Po510
8
), whe eas o gene and pa hway le el es s he
Bon e oni co ec ion was used o de e mine mul iple es ing
adjus ed signi icance h esholds (gene Po2.7 10
4
, pa hway
Po5.56 10
3
).
All analyses we e ca ied ou using R. Fo single-SNP analyses,
pe allele odds a ios we e es ima ed using logis ic eg ession.
SKAT es s we e pe o med using he SKAT package o R (h p://
CRAN.R-p ojec .o g/package=SKAT). We used he SKAT-O and
SKAT-C es s o op imal analyses o he combined e ec o
mul iple a e a ian s and common and a e a ian s, espec i ely
(Wu e al, 2011; Lee e al, 2012; Ioni a-Laza e al, 2013). Tes s we e
conduc ed using de aul pa ame e s and a common/ a e cu o
h eshold o MAF ¼0.01 o he SKAT-C es . S epAIC and SKAT
lea e one ou we e used o u he in e oga e he signi ican SKAT
signal a he MSH5 gene o he indi idual a ian s ha bes
desc ibed he signal.
Analyses o low-g ade s high-g ade P Ca we e ca ied ou
based on wo clinical c i e ia. Fo s ingen compa ison o non-
agg essi e and agg essi e P Ca, we de ined NCCN s age 1 pa ien s
as non-agg essi e P Ca and indi iduals wi h me as a ic disease
(M
þ
) o nodal sp ead (N
þ
) as agg essi e (395 NCCN1 s 1391
M
þ
/N
þ
), whe eas o enhance he sample panel a ailable o his
analysis we also compa ed pa ien s wi h Gleason S age (GS) p6
agains hose wi h GS X8 disease (9626 GS p6 s 2776 GS X8).
BRITISH JOURNAL OF CANCER DNA epai gene a ia ion and p os a e cance isk
946 www.bjcance .com | DOI:10.1038/bjc.2016.50
RESULTS
Using geno ype da a om he iCOGS s udy impu ed o he 1000
genomes phase 3 e e ence panel we analysed 81 303 SNPs wi hin a
20-kb lanking egion o 179 genes wi h a co e unc ion in DNA-
damage epai (Supplemen a y Table 1). Ra e and uncommon
a ian s ep esen ed a subs an ial p opo ion o he da a se , wi h
29 503 a ian s o MAF p1%, 16 689 wi h MAF 1–5% and 35 111
wi h MAF 45% (Supplemen a y Figu e 1a). Va ian s we e
ca ego ised as SNPs, inse ions and dele ions, anno a ed using
wANNOVAR (Wang e al, 2010; Chang and Wang, 2012), and
classi ied in o i e ca ego ies; coding, UTR, splice, in onic and
in e genic. Va ian s a ailable o his analysis we e p edominan ly
si ua ed wi hin non-coding (in onic o in e genic) egions, wi h
3943 a ian s anno a ed as coding, splice o UTR in o al, whe eas
mos we e single-base subs i u ions, wi h 3914 inse ions and 5576
dele ions, espec i ely. All o he inse ion and dele ion a ian s
we e impu ed, wi h he as majo i y loca ed wi hin non-coding
egions (Supplemen a y Figu e 1b–d, Supplemen a y Table 2). All
analyses we e adjus ed o s udy popula ion and he i s eigh
p incipal componen s. Fo single- a ian le el analyses he
genome-wide signi icance h eshold (Po510
8
) was used o
de e mine signi ican ly associa ed a ian s, whe eas o gene and
pa hway le el analyses he signi icance h eshold was de ined
acco ding o he Bon e oni co ec ion (gene Po2.7 10
4
,
pa hway Po5.56 10
3
).
Single- a ian analysis o associa ion o DNA- epai gene
a ian s wi h P Ca iden i ied only he p e iously epo ed
associa ion wi h RAD51B a Ch 14q24 ( s371311594, P¼1.29
10
10
). Se e al o he gene loci showed sugges i e associa ion
peaks; howe e , no o he a ian s we e wi hin one o de o
magni ude o genome-wide signi icance (Figu e 1, Supplemen a y
Table 3).
We obse ed e idence o modes in la ion wi hin ou associa ion
da a (l¼1.105); none heless, depa u e om he null was appa en
owa ds he ex emi y o he P- alue dis ibu ion and his pe sis ed o
a mo e modes ex en e en a e he RAD51B egion was excluded
(Supplemen a y Figu e 2). We subsequen ly pe o med gene le el
associa ion es s, in an a emp o asce ain whe he addi ional
pu a i e P Ca- isk signals migh be p esen among he genes wi hin
which no indi idual a ian achie ed signi icance a e adjus men
o mul iple es ing, a ising h ough a cumula i e e ec o se e al low
MAF o low pene ance a ian s. We pe o med wo gene-le el
associa ion es s using he SKAT; SKAT-C, which is op imised o
combined es ing o a e and common a ian s and SKAT-O, which
a emp s o maximise powe o a e a ian es ing (Lee e al, 2012;
Ioni a-Laza e al, 2013). Gene-le el analysis iden i ied a no el
signi ican associa ion wi h he MSH5 gene using he SKAT-C es
(Ch 6p21; P¼1.68 10
4
) (Figu e 2, Supplemen a y Table 4). We
used s epAIC and lea e one ou o SKAT o u he in e oga e he
MSH5 da a o he indi idual a ian s ha bes explain he signal.
This es selec ed h ee a ian s a he MSH5 locus, s61036903
(known as 6 : 31713892 wi hin he e e ence panel) in onic wi hin
he gene and wo a ian s 10-kb downs eam wi hin an adjacen gene
VWA7, s805825 and s185333600. These we e all among he
op- anking a ian s in he single-SNP analysis ( s61036903:
MAF ¼0.14, OR 0.92, P¼8.06 10
5
; s805825: MAF ¼0.45, OR
0.94, P¼4.05 10
5
; s185333600: MAF ¼0.003, OR 1.57,
P¼6.83 10
4
).
We subsequen ly examined he iCOGS da a se a he pa hway
le el unde he SKAT es o supplemen he gene-le el analyses. We
again used he Bon e oni co ec ion o de ine he signi icance
h eshold (pa hway Po5.56 10
3
). No pa hway achie ed sig-
ni icance a his h eshold, wi h sugges i e associa ions unde he
SKAT-O es obse ed wi h he anslesion syn hesis pa hway
(P¼6.18 10
3
) and misma ch- epai pa hway (P¼0.056).
Va ian s wi hin he coding sequence o DNA- epai genes could
be mo e likely o in luence P Ca isk han hose in non-coding
egions. We he e o e pe o med an addi ional SKAT es o assess
whe he he coding DNA- epai gene a ian s a ailable o his s udy,
when collapsed as a single en i y, could s a i y case and con ol
s a us. We obse ed a signi ican associa ion when using he SKAT-C
es (P¼0.003), which sugges s ha a ian s ha a ec he coding
sequence o genes pa icipa ing in DNA- epai p ocesses con ibu e
o P Ca isk. We a emp ed o u he elabo a e upon his inding by
analysing coding a ia ion wi hin each pa hway sepa a ely. Despi e
ela i ely modes numbe s o coding a ian s a ailable wi hin each
pa hway, we con inued o obse e sugges i e associa ions unde he
SKAT-O es o he anslesion syn hesis pa hway (P¼0.026) and
misma ch- epai pa hway (P¼0.055), in addi ion o he HR/FA
pa hway unde he SKAT-C es (P¼0.011).
To complemen he es s designed o iden i y po en ial P Ca
suscep ibili y a ian s and genes, we also pe o med case–case
analyses o in es iga e whe he indi idual o cumula i e ge mline
DNA- epai gene and pa hway a ian s in he iCOGS impu ed
da a se co ela ed wi h pheno ypic cha ac e is ics o mo e
agg essi e P Ca. This analysis was limi ed by lack o comple e
pheno ypic da a o all pa ien s wi hin he iCOGS sample se and
low numbe s o samples wi hin indi idual pheno ypic subg oups;
he e o e, we u ilised wo sepa a e c i e ia o de ine agg essi e and
non-agg essi e disease. Fo a s ingen compa ison o non-
agg essi e and agg essi e P Ca, we analysed NCCN s age 1
pa ien s agains indi iduals wi h me as a ic disease (M
þ
) o nodal
sp ead (N
þ
) (395 NCCN1 s 1391 M
þ
/N
þ
), whe eas o maximise
he numbe o samples a ailable we also compa ed pa ien s wi h
GS p6 disease agains hose wi h GS X8 (9626 GS p6 s 2776
GS X8). No signi ican associa ions wi h agg essi e P Ca
we e iden i ied a ei he he a ian o gene le els o ei he
o he pheno ypic c i e ia es ed. (Supplemen a y Figu e 3,
Supplemen a y Table 5). When we examined P Ca agg essi eness
a he pa hway le el, we obse ed associa ions a Po0.05 o he
HR/FA pa hway unde bo h es s o he GS p6 s GS X8
pheno ype coho (SKAT-C P¼0.011, SKAT-O P¼0.040).
This pa hway was also he highes anked o he NCCN1 s
M
þ
/N
þ
pheno ype coho unde he SKAT-C es (P¼0.052).
When hese analyses we e es ic ed o only coding a ian s, an
associa ion a Po0.05 emained o he HR/FA pa hway o he
NCCN1 s M
þ
/N
þ
coho and he SKAT-O es (P¼0.021).
These sugges i e associa ions we e no howe e signi ican a e
adjus men o mul iple es ing (Supplemen a y Table 5).
DISCUSSION
DNA- epai genes ha e a c ucial ole in he co ec ion o damage
o he genome o a cell and he e o e hei impai men can lead o
ca cinogenesis. Al hough hese de imen al gene ic al e a ions
equen ly o igina e wi hin soma ic cells du ing an indi idual’s
li e ime, a numbe o a e, he edi a y mu a ions wi hin speci ic
DNA- epai genes ha e been iden i ied ha con e subs an ially
inc eased isks o he indi idual o P Ca and o he cance s. GWAS
ha e also p e iously iden i ied common, low-pene ance a ian s
in close p oximi y o he DNA- epai genes RAD51B and RAD23B
ha con ibu e o P Ca suscep ibili y (Xu e al, 2012; Eeles e al,
2013; Amin Al Olama e al, 2015). Howe e , e en ela i ely well-
powe ed gene ic associa ion s udies may ha e been limi ed in hei
abili y o eliably in e oga e a ian s wi h lowe MAFs o
associa ions wi h modes e ec sizes; he e o e, addi ional- isk
a ian s ha con e hei unc ional e ec hough DNA- epai
genes may emain o be disco e ed. We ha e ecen ly impu ed
P Ca da a om he iCOGS s udy o he 1000 Genomes phase 3
e e ence panel, he eby enhancing he capabili y o in e oga e his
DNA epai gene a ia ion and p os a e cance isk BRITISH JOURNAL OF CANCER
www.bjcance .com | DOI:10.1038/bjc.2016.50 947
da a se o un yped a ian s wi hin agged egions. In pa icula , a
a g ea e numbe o lowe MAF and inse ion and dele ion
a ian s we e a ailable o analysis, al hough hese a e p edomi-
nan ly si ua ed in non-coding egions. Impu a ion pe o mance o
lowe MAF a ian s is imp o ed by la ge e e ence panel size and
e hnic di e si y and highe ma ke densi y on he geno yping
a ay; howe e , a e a ian s s ill egula ly emain challenging o
impu e wi hou an addi ional e e ence panel en iched o speci ic
low- equency a ian s o known in e es , and may also be mo e
sensi i e o di e ences in he impu a ion app oach employed
(Ho mann and Wi e, 2015). Ou ela i ely la ge sample size
p o ided good powe o de ec associa ions wi h P Ca o a e
a ian s wi h g ea e e ec sizes (e.g., o a a ian a ou 0.1%
MAF cu o wi h OR 2.5, we had 78% powe ) as well as common,
low-pene ance a ian s (e.g., o a a ian wi h OR 1.1 and a MAF
o 20%, powe was 86%). We we e howe e limi ed wi h espec o
he de ec ion o a ian s wi h he combina ion o bo h modes
allele equency and e ec size.
We ha e examined all a ian s in he impu ed iCOGS da a se
si ua ed wi hin 20-kb o a panel o 179 DNA- epai genes o
associa ion wi h P Ca o mo e agg essi e pheno ypic p esen a ion.
No no el isk a ian s we e iden i ied in ou single-SNP analysis,
wi h only he p e iously epo ed signal a RAD51B on Ch 14q24
genome-wide signi ican (Figu e 1, Supplemen a y Table 3).
Ou analysis did no de ec he p e iously epo ed signal a he
RAD23B locus on Ch 9q31, which was o iginally iden i ied in he
Chinese popula ion and ecen ly also con i med in Eu opeans wi h
he mos signi ican ly associa ed a ian s1771718 and he signal
also an eQTL o RAD23B in no mal p os a e issue in he TCGA
da a se (Xu e al, 2012; Amin Al Olama e al, 2015). s1771718 is
loca ed B57 kb downs eam o RAD23B, which is he closes
neighbou ing gene bu loca ed in a dis inc ecombina ion block
5
AB
4
3
–log10(p)
2
1
0
5
4
3
–log10(p)
2
1
0
1
2
3
4
5
6
7
8
9
Ch omosome
SKAT_C SKAT_O
10
11
12
13
14
15
16
17
18
20
1
2
3
4
5
6
7
8
9
Ch omosome
10
11
12
13
14
15
16
17
18
20
Figu e 2. Case–con ol Manha an Plo s o he 179 DNA- epai genes analysed by SKAT. (A) A signi ican associa ion was obse ed o he MSH5
gene using he SKAT-C es ha examines he combined e ec o common and a e a ian s. (B) No signi ican associa ion was de ec ed o any
gene unde he SKAT-O es ha p ima ily ocuses on a e a ian associa ion es ing.
10 Case-con ol single SNP
Ch omosome
8
6
4
–log10(p)
2
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
20
Figu e 1. Single SNP case–con ol Manha an Plo . In o al, 81 303 SNPs om 179 DNA- epai genes we e analysed o associa ion wi h P Ca.
Only he p e iously epo ed associa ion wi hin he RAD51B gene was iden i ied, wi h sugges i e, non-signi ican associa ion peaks obse ed a a
small numbe o o he loci.
BRITISH JOURNAL OF CANCER DNA epai gene a ia ion and p os a e cance isk
948 www.bjcance .com | DOI:10.1038/bjc.2016.50
om hese isk a ian s. As no a ian among he 509 wi hin he
gene cen ic egion ha we in e oga ed in his s udy showed
subs an ial e idence o associa ion (PX2.94 10
3
), i appea s
likely ha isk a his locus is modula ed h ough a nea by
egula o y elemen con olling exp ession o he gene as opposed o
in agenic causal unc ional a ian s (Supplemen a y Figu e 4).
We conduc ed wo gene-le el analyses in an a emp o iden i y
whe he he e may be addi ional signals among he se e al loci ha
demons a ed sugges i e bu non-signi ican associa ion peaks
in ou single-SNP analysis, bu o which no indi idual a ian
had achie ed signi icance. SKAT-C es s o he combined e ec s o
common and a e a ian s, whe eas SKAT-O adap i ely combines
he bu den es and SKAT es in an a emp o maximise powe o
a e a ian associa ion es ing (Lee e al, 2012; Ioni a-Laza e al,
2013). We iden i ied a signi ican P Ca- isk associa ion a e
adjus men o mul iple es ing a he MSH5 gene a Ch 6p21 using
he SKAT-C es , implying ha mul iple common, o a combina ion
o common and a e a ian s wi hin his gene may con ibu e o
P Ca isk. Al hough cau ion mus be aken wi h espec o his
inding un il eplica ed and decons uc ed, his e idence implica es
MSH5 as a p ospec i e P Ca suscep ibili y locus ha wa an s
addi ional ollow-up. MSH5 had p e iously been epo ed as a
plausible candida e gene o he lung cance - isk locus a Ch 6p21.33,
o which he mos s ongly associa ed a ian s3117582 is in onic
in BAT3; howe e , is highly co ela ed o s3131379 in in on 10 o
MSH5 (Wang e al, 2008; Kazma e al, 2012). A ecen s udy
examining cance pleio opy among DNA- epai and DNA-damage
signalling pa hway a ian s has also epo ed a highly signi ican
associa ion wi h lung cance o s3115672, a synonymous a ian
wi hin MSH5, in addi ion o weake associa ions wi h colon and
se ous o a ian cance s (pleio opic OR 1.18, 95% CI 1.12-1.24,
P¼2.53 10
8
) (Sca b ough e al, 2016). This a ian was howe e
non-signi ican o P Ca wi hin hei s udy o 14 160 P Ca cases and
12 724 con ols (OR 0.96, P¼0.21).Wi hinou la ge s udy(o
which 2614 cases and 2679 con ols o e lapped wi h hose o
Sca b ough e al), in he single-SNP analysis, s3115672 emained
non-signi ican a e adjus men o mul iple es ing (OR 0.94, 95%
CI 0.90–0.98, P¼5.69 10
3
). Howe e , a numbe o o he a ian s
among he 312 wi hin he MSH5 gene in ou analysis we e mo e
s ongly associa ed, he op indi idual a ian o which was
s9281573 (OR 0.94, P¼4.01 10
5
). S epAIC combined wi h
SKAT lea e one ou selec ed wo common and one a e a ian as
bes explaining he SKAT-C associa ion, all o which we e among he
op a ian s in he single-SNP analysis. This implies ha a
combina ion o common and a e a ian s could po en ially
unde pin his signal.
We anno a ed hese h ee a ian s o e idence o unc ionali y
using HaploReg 4.1 (Wa d and Kellis, 2016); his anno a ion
included ch oma in s a e da a o cell lines de i ed om mul iple
issue ypes p o ided by he Roadmap Epigenomics Conso ium
(Roadmap Epigenomics e al, 2015); howe e , no da a o p os a e
issue we e a ailable. s61036903, which is in onic o MSH5,
showed limi ed di ec e idence o unc ionali y i sel . Bo h o he
a ian s si ua ed a ound he MSH5 p omo e egion, wi hin
VWA7, showed s ong e idence o being loca ed wi hin enhance
elemen s ha a e ac i e ac oss a wide ange o issue ypes.
In addi ion, exp ession da a om he GTEx Conso ium indica es
ha s805825 is an eQTL o a numbe o genes om he MHC
egion (HLA-DRB1,HLA-DRB5,LY6G5C,DDAH2,LY6G6C,
HSPA1B and C4B) (GTEx Conso ium, 2015). These genes a e
clus e ed closely cen ome ic and elome ic o MSH5 and VWA7
wi hin a gene dense locus; howe e , no eQTL wi h MSH5 o
VWA7 was obse ed o his a ian .
Al hough he MSH5 gene is ou inely classi ied as a membe o
he MMR pa hway along wi h all o he homologues o Mu S
(Wood e al, 2005; Ji e al, 2012; Sca b ough e al, 2016), unc ional
e idence o da e p o ides limi ed suppo o a ole in MMR o
MSH5 i sel . Ins ead, his gene has been implica ed p ima ily in he
p ocesses o meio ic ecombina ion, main enance o ch omosome
in eg i y and DNA double-s and b eak epai (Cla k e al, 2013;
Wu e al, 2013). RNA-seq da a om GTEx Analysis Release V6 o
2712 o al samples ac oss 51 no mal human issues (including 106
p os a e issue samples) demons a es ha MSH5 is exp essed a
b oadly simila le els ac oss a wide ange o issue ypes, including
p os a e (GTEx Conso ium, 2015; accessed ia. h p://www.g ex-
po al.o g/home/gene/MSH5). Da a om TCGA u he suppo
his exp ession p o ile ac oss a ange o no mal issues and also
indica es ha MSH5 is consis en ly o e exp essed o almos all
umou ypes in compa ison wi h hei espec i e no mal issues.
Fo TCGA p os a e issue, a median RSEM (log2) alue o 8.08 was
obse ed ac oss 498 umou samples compa ed wi h 6.85 om 52
no mal samples (h p://cance genome.nih.go /; accessed ia.
h p:// i eb owse.o g/ iewGene.h ml?gene=msh5).
Taken oge he , hese in o ma ion demons a e ha al hough he
MSH5 gene ep esen s a s ong biological candida e o he P Ca- isk
associa ion ha we ha e obse ed, addi ional unc ional ollow-up
s udies will be equi ed o dissec he p ecise unc ional a ian s,
genes, egula o y elemen s o p ocesses ha unde pin his signal.
I is wo h no ing ha he gene le el analyses in his s udy did
no iden i y signi ican associa ions wi h any genes p e iously
implica ed in P Ca suscep ibili y. This was i espec i e o whe he
he known isk mechanisms a e belie ed o ope a e h ough
mul iple common, low-pene ance a ian s (e.g., RAD51B; SKAT-
OP¼0.05, SKAT-C P¼2.76 10
3
) o a e coding a ian s (e.g.,
BRCA2; SKAT-O P¼0.46, SKAT-C P¼0.15). In he case o
BRCA2 and o he genes in which a e, mode a e pene ance,
p o ein unca ing P Ca suscep ibili y a ian s had p e iously been
iden i ied, his is likely o e lec he ac ha e en using he la es
1000 Genomes e e ence panel, a e a ian s expec ed o con e
g ea e pheno ypic consequences may emain absen om he
e e ence panel and consequen ly unimpu able. This is consis en
wi h he poo ep esen a ion o coding inse ion and dele ion
a ian s wi hin ou da a se and would ha e ende ed us
unde powe ed o de ec he e ec s o his class o a ia ion in
ou analysis. Ou obse a ions do howe e imply ha any
addi ional con ibu ion om common, lowe pene ance a ia ion
a hese genes may be minimal. This includes he s11571833
nonsense polymo phism in he e minal exon o BRCA2, which is a
epo ed lung cance suscep ibili y a ian , bu was no associa ed
wi h P Ca in his s udy (OR 1.03, 95% CI 0.89–1.19, P¼0.74)
(Wang e al, 2014). I is pe haps mo e su p ising ha RAD51B did
no achie e signi icance unde he SKAT-C es , which conside s
he po en ial con ibu ion owa ds associa ion o bo h common
and a e a ian s wi hin a egion, gi en ha h ee independen
associa ions ha e p e iously been iden i ied a his locus (Amin Al
Olama e al, 2015). Howe e , a sugges i e associa ion was obse ed
unde his es , which may be an indica ion ha he cumula i e
e ec size o he independen low-pene ance- isk a ian s wi hin
his egion we e insu icien o be conclusi ely disambigua ed
h ough his me hodology.
Ou pa hway-le el analysis iden i ied sugges i e bu non-
signi ican associa ions o wo pa hways unde he SKAT-O es ;
anslesion syn hesis and misma ch epai . Al hough his s udy did
no he e o e p o ide su icien e idence o implica e genes wi hin
hese pa hways in P Ca suscep ibili y, gi en he inhe en ly
conse a i e na u e o he Bon e oni co ec ion wi h espec o
ype II e o and he ela i ely low p opo ion o coding a ian s
wi hin ou da a se , hese obse a ions may s ill jus i y u he
e alua ion. In pa icula , as hese sugges i e associa ions we e
obse ed unde he SKAT-O es ha maximises powe o a e
a ian associa ion analyses and we e no ab oga ed when he
analyses we e es ic ed only o coding a ian s, i subs an ia ed,
hese nascen obse a ions could be unde pinned by di ec e ec s
o a e a ian s on he p o ein s uc u e and unc ion.
DNA epai gene a ia ion and p os a e cance isk BRITISH JOURNAL OF CANCER
www.bjcance .com | DOI:10.1038/bjc.2016.50 949

Consequen ly, sequencing s udies designed o comp ehensi ely
analyse he en i e coding sequence o genes wi hin he anslesion
syn hesis and misma ch epai pa hways could po en ially yield
u he insigh owa ds he mechanisms o suscep ibili y o
de eloping P Ca. I is also wo h no ing ha soma ic mu a ions
in anslesion syn hesis pa hway genes, in pa icula he POLK
gene, ha e been obse ed in p os a e umou s p e iously
(Mak idakis and Reicha d , 2012; Yada e al, 2015), whe eas a
a e ge mline nonsynonymous a ian in he POLI gene has also
been epo ed o p edispose owa ds he occu ence o he
TMPRSS2-ERG usion in P Ca pa ien s (Luedeke e al, 2009).
Inc easing e idence sugges s ha mode a e-pene ance ge mline
mu a ions wi hin DNA- epai genes also co ela e wi h a mo e
agg essi e pheno ypic p esen a ion o P Ca and poo e p ognosis
(Cas o e al, 2013; Cybulski e al, 2013; Leongamo nle e al,
2014; Robinson e al, 2015). This could in u n signi y ha DNA-
epai gene a ian s migh exis ha do no con e g ea e isk o
de eloping P Ca pe se, ye do modi y he likelihood o de eloping
mo e agg essi e disease in indi iduals ha de elop P Ca owing o
o he isk ac o s o exposu es. We he e o e also pe o med case–
case analyses o u he explo e his hypo hesis using wo dis inc
pheno ypic c i e ia. No signi ican o sugges i e associa ions wi h
agg essi e disease we e iden i ied a he indi idual a ian o gene
le els unde ei he de ini ion; howe e , sugges i e non-signi ican
associa ions wi h he HR/FA pa hway we e obse ed. These analyses
we e, howe e , limi ed by ela i ely low sample numbe s wi hin each
compa ison g oup, which would ha e educed ou powe o de ec
associa ions, pa icula ly o a e and uncommon a ian s. We
canno he e o e exclude he exis ence o addi ional DNA- epai gene
a ian s ha p omo e inc eased P Ca agg essi eness a he han isk
o he disease i sel ; howe e , ou da a would sugges ha any ha
exis a e mo e likely o be a e han common.
O e all, his s udy ep esen s he mos comp ehensi e in e oga-
ion o he ole o DNA- epai gene a ian s in P Ca suscep ibili y
ha we a e awa e o o da e. We con i med he p esence o
low-pene ance suscep ibili y loci si ua ed a he RAD51B locus and
ound e idence o implica e a no el gene, MSH5,inP Ca
suscep ibili y. We also sha e p elimina y obse a ions ha a e
ge mline a ia ion in genes wi hin he anslesion syn hesis pa hway,
in pa icula a ian s wi hin he coding sequence, could be wo hy o
u he in es iga ion as candida es o P Ca isk.
The main limi a ions o ou s udy ela e o he challenges in
impu ing a e, po en ially pa hogenic a ian s o a ay geno ype da a
om popula ion-based e e ence panels and in pe o ming associa-
ion es s on low- equency a ian s in a la ge mul i-popula ion s udy
while con olling o popula ion s a i ica ion. The e o e, addi ional
sequencing s udies would s ill be wa an ed o u he explo e he
con ibu ion o a e DNA- epai gene a ian s o P Ca isk. In
addi ion, incomple e a ailabili y o pheno ypic da a and he ac ha
he iCOGS s udy did no speci ically selec indi iduals wi h low- o
high-g ade disease may ha e educed ou abili y o examine any
po en ial in luence o hese a ian s on P Ca agg essi eness. Fu u e
s udies, whe he a ay o sequencing based, ha speci ically selec
pa ien s om hese coho s o inclusion would acili a e in es iga-
ion o his aspec ; which migh in u n help o enhance s a i ica ion
o pa ien s ha equi e al e ed clinical managemen pa hways.
CONFLICT OF INTEREST
The au ho s decla e no con lic o in e es .
REFERENCES
Al Olama AA, Ko e-Ja ai Z, Be nd SI, Con i DV, Schumache F, Han Y,
Benlloch S, Hazele DJ, Wang Z, Saunde s E, Leongamo nle D,
Linds om S, Jugu nau h-Li le S, Dadae T, Tym akiewicz M, S am DO,
Rand K, Wan P, S am A, Sheng X, Poole LC, Pa k K, Xia L, Ty e J,
Kolonel LN, Le Ma chand L, Hoo e RN, Machiela MJ, Yeage M,
Bu de e L, Chung CC, Hu chinson A, Yu K, Goh C, Ahmed M,
Go indasami K, Guy M, Tammela TL, Au inen A, Wahl o s T,
Schleu ke J, Visako pi T, Leinonen KA, Xu J, Aly M, Dono an J,
T a is RC, Key TJ, Siddiq A, Canzian F, Khaw KT, Takahashi A, Kubo M,
Pha oah P, Pashayan N, Weische M, No des gaa d BG, Nielsen SF,
Kla sko P, Rode MA, I e sen P, Thibodeau SN, McDonnell SK,
Schaid DJ, S an o d JL, Kolb S, Hol S, Knudsen B, Coll AH, Gaps u SM,
Di e WR, S e ens VL, Maie C, Luedeke M, He komme K, Rinckleb AE,
S om SS, Pe away C, Yeboah ED, Te ey Y, Bi i wum RB, Adjei AA,
Tay E, T uelo e A, Niwa S, Chokkalingam AP, Cannon-Alb igh L,
Cybulski C, Wokolo czyk D, Kluzniak W, Pa k J, Selle s T, Lin HY,
Isaacs WB, Pa in AW, B enne H, Die enbach AK, S egmaie C, Chen C,
Gio annucci EL, Ma J, S amp e M, Penney KL, Mucci L, John EM,
Ingles SA, Ki les RA, Mu phy AB, Pandha H, Michael A, Kie zek AM,
Blo W, Signo ello LB, Zheng W, Albanes D, Vi amo J, Weins ein S,
Nemesu e B, Ca p en J, Leske C, Wu SY, Hennis A, Kibel AS, Rybicki BA,
Neslund-Dudas C, Hsing AW, Chu L, Goodman PJ, Klein EA, Zheng SL,
Ba a J, Clemen s J, Spu dle A, Teixei a MR, Paulo P, Maia S, Sla o C,
Kane a R, Mi e V, Wi e JS, Casey G, Gillande s EM, Semina a D,
Riboli E, Hamdy FC, Coe zee GA, Li Q, F eedman ML, Hun e DJ,
Mui K, G onbe g H, Neal DE, Sou hey M, Giles GG, Se e i G.
B eas , P os a e Cance Coho C, Conso ium P, Conso ium C,
Conso ium G-OECook MB, Nakagawa H, Wiklund F, K a P,
Chanock SJ, Hende son BE, Eas on DF, Eeles RA, Haiman CA (2014)
A me a-analysis o 87 040 indi iduals iden i ies 23 new suscep ibili y loci
o p os a e cance . Na Gene 46(10): 1103–1109.
Amin Al Olama A, Dadae T, Hazele DJ, Li Q, Leongamo nle D,
Saunde s EJ, S ephens S, Cieza-Bo ella C, Whi mo e I, Benlloch Ga cia S,
Giles GG, Sou hey MC, Fi zge ald L, G onbe g H, Wiklund F, Aly M,
Hende son BE, Schumache F, Haiman CA, Schleu ke J, Wahl o s T,
Tammela TL, No des gaa d BG, Key TJ, T a is RC, Neal DE, Dono an JL,
Hamdy FC, Pha oah P, Pashayan N, Khaw KT, S an o d JL,
Thibodeau SN, McDonnell SK, Schaid DJ, Maie C, Vogel W, Luedeke M,
He komme K, Kibel AS, Cybulski C, Wokolo czyk D, Kluzniak W,
Cannon-Alb igh L, B enne H, Bu e bach K, A nd V, Pa k JY, Selle s T,
Lin HY, Sla o C, Kane a R, Mi e V, Ba a J, Clemen s JA, Spu dle A,
Teixei a MR, Paulo P, Maia S, Pandha H, Michael A, Kie zek A,
Go indasami K, Guy M, Lopha onanon A, Mui K, Vinuela A, B own AA.
Conso ium P, Ini ia i e C-CG-E, Aus alian P os a e Cance B,
Collabo a o s UKGPCS, Collabo a o s UKPSF eedman M, Con i DV,
Eas on D, Coe zee GA, Eeles RA, Ko e-Ja ai Z (2015) Mul iple no el
p os a e cance suscep ibili y signals iden i ied by ine-mapping
o known isk loci among Eu opeans. Hum Mol Gene 24(19):
5589–5602.
A a d G, Pa ke C, Eeles RA, Sch ode F, Tomlins SA, Tannock I, D ake CG,
de Bono JS (2015) P os a e cance . Lance 387(10013): 70–82.
Cance Resea ch UK (2014) Cance S a is ics Repo : Cance Incidence and
Mo ali y in he UK, Sep embe 2014.
Cas o E, Goh C, Olmos D, Saunde s E, Leongamo nle D, Tym akiewicz M,
Mahmud N, Dadae T, Go indasami K, Guy M, Sawye E, Wilkinson R,
A de n-Jones A, Ellis S, F os D, Peock S, E ans DG, Tischkowi z M,
Cole T, Da idson R, Eccles D, B ewe C, Douglas F, Po eous ME,
Donaldson A, Do kins H, Iza L, Cook J, Hodgson S, Kennedy MJ,
Side LE, Eason J, Mu ay A, An oniou AC, Eas on DF, Ko e-Ja ai Z,
Eeles R (2013) Ge mline BRCA mu a ions a e associa ed wi h highe isk
o nodal in ol emen , dis an me as asis, and poo su i al ou comes in
p os a e cance . J Clin Oncol 31(14): 1748–1757.
Chang X, Wang K (2012) wANNOVAR: anno a ing gene ic a ian s o
pe sonal genomes ia he web. J Med Gene 49(7): 433–436.
Cla k N, Wu X, He C (2013) Mu S homologues hMSH4 and hMSH5: gene ic
a ia ions, unc ions, and implica ions in human diseases. Cu Genomics
14(2): 81–90.
Cybulski C, Wokolo czyk D, Kluzniak W, Jakubowska A, Go ski B,
G onwald J, Huza ski T, Kashyap A, By ski T, Debniak T, Golab A,
Gliniewicz B, Siko ski A, Swi ala J, Bo kowski T, Bo kowski A, An czak A,
Wojna L, P zybyla J, Sosnowski M, Malkiewicz B, Zd ojowy R,
Siko ska-Radek P, Ma ych J, Wilkosz J, Rozanski W, Kis J, Ba K,
B ynia ski P, Pa adysz A, Je sak K, Niemi owicz J, Slupski P, Ja zemski P,
Sk zypczyk M, Dob uch J, Domagala P, Na od SA, Lubinski J.
Polish He edi a y P os a e Cance C (2013) An inhe i ed NBN mu a ion
BRITISH JOURNAL OF CANCER DNA epai gene a ia ion and p os a e cance isk
950 www.bjcance .com | DOI:10.1038/bjc.2016.50
is associa ed wi h poo p ognosis p os a e cance . B J Cance 108(2):
461–468.
Delaneau O, Zagu y JF, Ma chini J (2013) Imp o ed whole-ch omosome
phasing o disease and popula ion gene ic s udies. Na Me hods 10(1):
5–6.
Dong X, Wang L, Taniguchi K, Wang X, Cunningham JM, McDonnell SK,
Qian C, Ma ks AF, Slage SL, Pe e son BJ, Smi h DI, Che ille JC,
Blu e ML, Jacobsen SJ, Schaid DJ, Tindall DJ, Thibodeau SN, Liu W (2003)
Mu a ions in CHEK2 associa ed wi h p os a e cance isk. Am J Hum
Gene 72(2): 270–280.
Eeles R, Goh C, Cas o E, Banc o E, Guy M, Al Olama AA, Eas on D,
Ko e-Ja ai Z (2014) The gene ic epidemiology o p os a e cance and i s
clinical implica ions. Na Re U ol 11(1): 18–31.
Eeles RA, Olama AA, Benlloch S, Saunde s EJ, Leongamo nle DA,
Tym akiewicz M, Ghoussaini M, Lucca ini C, Dennis J,
Jugu nau h-Li le S, Dadae T, Neal DE, Hamdy FC, Dono an JL, Mui K,
Giles GG, Se e i G, Wiklund F, G onbe g H, Haiman CA, Schumache F,
Hende son BE, Le Ma chand L, Linds om S, K a P, Hun e DJ,
Gaps u S, Chanock SJ, Be nd SI, Albanes D, And iole G, Schleu ke J,
Weische M, Canzian F, Riboli E, Key TJ, T a is RC, Campa D,
Ingles SA, John EM, Hayes RB, Pha oah PD, Pashayan N, Khaw KT,
S an o d JL, Os ande EA, Signo ello LB, Thibodeau SN, Schaid D,
Maie C, Vogel W, Kibel AS, Cybulski C, Lubinski J, Cannon-Alb igh L,
B enne H, Pa k JY, Kane a R, Ba a J, Spu dle AB, Clemen s JA,
Teixei a MR, Dicks E, Lee A, Dunning AM, Baynes C, Con oy D,
Ma anian MJ, Ahmed S, Go indasami K, Guy M, Wilkinson RA,
Sawye EJ, Mo gan A, Dea naley DP, Ho wich A, Hudda RA, Khoo VS,
Pa ke CC, Van As NJ, Woodhouse CJ, Thompson A, Dudde idge T,
Ogden C, Coope CS, Lopha ananon A, Cox A, Sou hey MC, Hoppe JL,
English DR, Aly M, Adol sson J, Xu J, Zheng SL, Yeage M, Kaaks R,
Di e WR, Gaude MM, S e n MC, Co al R, Joshi AD, Shahabi A,
Wahl o s T, Tammela TL, Au inen A, Vi amo J, Kla sko P,
No des gaa d BG, Rode MA, Nielsen SF, Bojesen SE, Siddiq A,
Fi zge ald LM, Kolb S, Kwon EM, Ka yadi DM, Blo WJ, Zheng W, Cai Q,
McDonnell SK, Rinckleb AE, D ake B, Coldi z G, Wokolo czyk D,
S ephenson RA, Tee link C, Mulle H, Ro henbache D, Selle s TA,
Lin HY, Sla o C, Mi e V, Lose F, S ini asan S, Maia S, Paulo P,
Lange E, Cooney KA, An oniou AC, Vincen D, Baco F, Tessie DC.
Ini ia i e CO-CRUG-E, Aus alian P os a e Cance BOncology
UKGPCSCBAoUSSo, Collabo a o s UKPS, Conso ium PKo e-Ja ai Z,
Eas on DF (2013) Iden i ica ion o 23 new p os a e cance suscep ibili y
loci using he iCOGS cus om geno yping a ay. Na Gene 45(4):
385–391e1–2.
GTEx Conso ium (2015) Human genomics. he geno ype- issue exp ession
(GTEx) pilo analysis: mul i issue gene egula ion in humans. Science
348(6235): 648–460.
Hjelmbo g JB, Scheike T, Hols K, Sky he A, Penney KL, G a RE, Pukkala E,
Ch is ensen K, Adami HO, Holm NV, Nu all E, Hansen S, Ha man M,
Czene K, Ha is JR, Kap io J, Mucci LA (2014) The he i abili y o p os a e
cance in he No dic Twin S udy o Cance . Cance Epidemiol Bioma ke s
P e 23(11): 2303–2310.
Ho mann TJ, Wi e JS (2015) S a egies o impu ing and analyzing a e
a ian s in associa ion s udies. T ends Gene 31(10): 556–563.
Howie BN, Donnelly P, Ma chini J (2009) A lexible 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(6): e1000529.
Ilic D, Neube ge MM, Djulbego ic M, Dahm P (2013) Sc eening o p os a e
cance . Coch ane Da abase Sys Re 1: CD004720.
Ioni a-Laza I, Lee S, Maka o V, Buxbaum JD, Lin X (2013) Sequence
ke nel associa ion es s o he combined e ec o a e and common
a ian s. Am J Hum Gene 92(6): 841–853.
Ji G, Long Y, Zhou Y, Huang C, Gu A, Wang X (2012) Common a ian s in
misma ch epai genes associa ed wi h inc eased isk o spe m DNA
damage and male in e ili y. BMC Med 10: 49.
Kang J, D’And ea AD, Kozono D (2012) A DNA epai pa hway- ocused
sco e o p edic ion o ou comes in o a ian cance ea ed wi h pla inum-
based chemo he apy. J Na l Cance Ins 104(9): 670–681.
Kazma R, Bab on MC, Gabo ieau V, Genin E, B ennan P, Hung RJ,
McLaughlin JR, K okan HE, El es ad MB, Sko pen F, Ande ssen E,
Voode T, Valk K, Me spalu A, Field JK, La h op M, Sa asin A,
Benhamou S, conso ium I (2012) Lung cance and DNA epai genes:
mul ile el associa ion analysis om he In e na ional Lung Cance
Conso ium. Ca cinogenesis 33(5): 1059–1064.
Ko e-Ja ai Z, Jugu nau h S, Mulholland S, Leongamo nle DA, Guy M,
Edwa ds S, Tym akiewi cz M, O’B ien L, Hall A, Wilkinson R,
Al Olama AA, Mo ison J, Mui K, Neal D, Dono an J, Hamdy F,
Eas on DF, Eeles R. Collabo a o s U, B i ish Associa ion o U ological
Su geons’ Sec ion o O (2009) A ecu en unca ing ge mline mu a ion
in he BRIP1/FANCJ gene and suscep ibili y o p os a e cance .
B J Cance 100(2): 426–430.
Ko e-Ja ai Z, Leongamo nle D, Saunde s E, Tym akiewicz M, Cas o E,
Mahmud N, Guy M, Edwa ds S, O’B ien L, Sawye E, Hall A,
Wilkinson R, Dadae T, Goh C, Eas on D. Collabo a o s U, Goldga D,
Eeles R (2011) BRCA2 is a mode a e pene ance gene con ibu ing o
young-onse p os a e cance : implica ions o gene ic es ing in p os a e
cance pa ien s. B J Cance 105(8): 1230–1234.
Lee S, Emond MJ, Bamshad MJ, Ba nes KC, Riede MJ, Nicke son DA.
Team NGESP-ELPCh is iani DC, Wu el MM, Lin X (2012) Op imal
uni ied app oach o a e- a ian associa ion es ing wi h applica ion o
small-sample case-con ol whole-exome sequencing s udies. Am J Hum
Gene 91(2): 224–237.
Leongamo nle D, Mahmud N, Tym akiewicz M, Saunde s E, Dadae T,
Cas o E, Goh C, Go indasami K, Guy M, O’B ien L, Sawye E, Hall A,
Wilkinson R, Eas on D. Collabo a o s UGoldga D, Eeles R, Ko e-Ja ai Z
(2012) Ge mline BRCA1 mu a ions inc ease p os a e cance isk.
B J Cance 106(10): 1697–1701.
Leongamo nle D, Saunde s E, Dadae T, Tym akiewicz M, Goh C,
Jugu nau h-Li le S, Koza ewa I, Fenwick K, Assio is I, Ba owdale D,
Go indasami K, Guy M, Sawye E, Wilkinson R. Collabo a o s U
An oniou AC, Eeles R, Ko e-Ja ai Z (2014) F equen ge mline dele e ious
mu a ions in DNA epai genes in amilial p os a e cance cases a e
associa ed wi h ad anced disease. B J Cance 110(6):
1663–1672.
Luedeke M, Linne CM, Ho e MD, Su owy HM, Rinckleb AE, Hoegel J,
Kue e R, Rubin MA, Vogel W, Maie C (2009) P edisposi ion o
TMPRSS2-ERG usion in p os a e cance by a ian s in DNA epai genes.
Cance Epidemiol Bioma ke s P e 18(11): 3030–3035.
Mak idakis NM, Reicha d JK (2012) T anslesion DNA polyme ases and
cance . F on Gene 3: 174.
Qua esma M, Coleman MP, Rache B (2015) 40-yea ends in an index o
su i al o all cance s combined and su i al adjus ed o age and sex o
each cance in England and Wales, 1971-2011: a popula ion-based s udy.
Lance 385(9974): 1206–12018.
Roadmap Epigenomics C, Kundaje A, Meuleman W, E ns J, Bilenky M,
Yen A, He a i-Moussa i A, Khe adpou P, Zhang Z, Wang J, Zille MJ,
Amin V, Whi ake JW, Schul z MD, Wa d LD, Sa ka A, Quon G,
Sands om RS, Ea on ML, Wu YC, P enning AR, Wang X, Claussni ze M,
Liu Y, Coa a C, Ha is RA, Sho esh N, Eps ein CB, Gjoneska E, Leung D,
Xie W, Hawkins RD, Lis e R, Hong C, Gasca d P, Mungall AJ, Moo e R,
Chuah E, Tam A, Can ield TK, Hansen RS, Kaul R, Sabo PJ, Bansal MS,
Ca les A, Dixon JR, Fa h KH, Feizi S, Ka lic R, Kim AR, Kulka ni A, Li D,
Lowdon R, Ellio G, Me ce TR, Neph SJ, Onuchic V, Polak P,
Rajagopal N, Ray P, Salla i RC, Sieben hall KT, Sinno -A ms ong NA,
S e ens M, Thu man RE, Wu J, Zhang B, Zhou X, Beaude AE, Boye LA,
De Jage PL, Fa nham PJ, Fishe SJ, Haussle D, Jones SJ, Li W,
Ma a MA, McManus MT, Sunyae S, Thomson JA, Tls y TD, Tsai LH,
Wang W, Wa e land RA, Zhang MQ, Chadwick LH, Be ns ein BE,
Cos ello JF, Ecke JR, Hi s M, Meissne A, Milosa lje ic A, Ren B,
S ama oyannopoulos JA, Wang T, Kellis M (2015) In eg a i e analysis
o 111 e e ence human epigenomes. Na u e 518(7539): 317–330.
Robinson D, Van Allen EM, Wu YM, Schul z N, Lonig o RJ, Mosque a JM,
Mon gome y B, Taplin ME, P i cha d CC, A a d G, Bel an H, Abida W,
B adley RK, Vinson J, Cao X, Va s P, Kunju LP, Hussain M, Feng FY,
Tomlins SA, Cooney KA, Smi h DC, B ennan C, Siddiqui J, Meh a R,
Chen Y, Ra hkop DE, Mo is MJ, Solomon SB, Du ack JC, Reu e VE,
Gopalan A, Gao J, Loda M, Lis RT, Bowden M, Balk SP, Ga iola G,
Sougnez C, Gup a M, Yu EY, Mos aghel EA, Cheng HH, Mulcahy H, T ue
LD, Plyma e SR, D inge H, Fe aldeschi R, Floh P, Mi anda S,
Za ei iou Z, Tuna iu N, Ma eo J, Pe ez-Lopez R, Demichelis F,
Robinson BD, Schi man M, Nanus DM, Tagawa ST, Siga as A, Eng KW,
Elemen o O, Sbone A, Hea h EI, Sche HI, Pien a KJ, Kan o P,
de Bono JS, Rubin MA, Nelson PS, Ga away LA, Sawye s CL,
Chinnaiyan AM (2015) In eg a i e clinical genomics o ad anced
p os a e cance . Cell 161(5): 1215–1228.
Sca b ough PM, Webe RP, I e sen ES, B hane Y, Amos CI, K a P, Hung RJ,
Selle s TA, Wi e JS, Pha oah P, Hende son BE, G ube SB, Hun e DJ,
DNA epai gene a ia ion and p os a e cance isk BRITISH JOURNAL OF CANCER
www.bjcance .com | DOI:10.1038/bjc.2016.50 951
Ga be JE, Joshi AD, McDonnell K, Eas on DF, Eeles R, Ko e-Ja ai Z,
Mui K, Dohe y JA, Schildk au JM (2016) A c oss-cance gene ic
associa ion analysis o he dna epai and dna damage signaling pa hways
o lung, o a y, p os a e, b eas , and colo ec al cance . Cance Epidemiol
Bioma ke s P e 25(1): 193–200.
Wang K, Li M, Hakona son H (2010) ANNOVAR: unc ional anno a ion o
gene ic a ian s om high- h oughpu sequencing da a. Nucleic Acids Res
38(16): e164.
Wang Y, B ode ick P, Webb E, Wu X, Vijayak ishnan J, Ma akidou A,
Qu eshi M, Dong Q, Gu X, Chen WV, Spi z MR, Eisen T, Amos CI,
Houls on RS (2008) Common 5p15.33 and 6p21.33 a ian s in luence
lung cance isk. Na Gene 40(12): 1407–1409.
Wang Y, McKay JD, Ra na T, Wang Z, Timo ee a MN, B ode ick P,
Zong X, Laplana M, Wei Y, Han Y, Lloyd A, Delahaye-Sou deix M,
Chubb D, Gabo ieau V, Wheele W, Cha e jee N, Tho lei sson G,
Sulem P, Liu G, Kaaks R, Hen ion M, Kinne sley B, Vallee M,
LeCal ez-Kelm F, S e ens VL, Gaps u SM, Chen WV, Za idze D,
Szeszenia-Dab owska N, Lissowska J, Rudnai P, Fabiano a E, Ma es D,
Bencko V, Fo e o a L, Janou V, K okan HE, Gab ielsen ME, Sko pen F,
Va en L, Njols ad I, Chen C, Goodman G, Benhamou S, Voode T,
Valk K, Nelis M, Me spalu A, Lene M, Lubinski J, Johansson M,
Vineis P, Agudo A, Cla el-Chapelon F, Bueno-de-Mesqui a HB,
T ichopoulos D, Khaw KT, Johansson M, Weide pass E, Tjonneland A,
Riboli E, La h op M, Scelo G, Albanes D, Capo aso NE, Ye Y, Gu J,
Wu X, Spi z MR, Dienemann H, Rosenbe ge A, Su L, Ma akidou A,
Eisen T, S e ansson K, Risch A, Chanock SJ, Ch is iani DC, Hung RJ,
B ennan P, Landi MT, Houls on RS, Amos CI (2014) Ra e a ian s o
la ge e ec in BRCA2 and CHEK2 a ec isk o lung cance . Na Gene
46(7): 736–741.
Wa d LD, Kellis M (2016) 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(D1): D877–D881.
Wood RD, Mi chell M, Lindahl T (2005) Human DNA epai genes, 2005.
Mu a Res 577(1-2): 275–283.
Wu MC, Lee S, Cai T, Li Y, Boehnke M, Lin X (2011) Ra e- a ian associa ion
es ing o sequencing da a wi h he sequence ke nel associa ion es .
Am J Hum Gene 89(1): 82–93.
Wu X, Xu Y, Feng K, Tompkins JD, He C (2013) Mu S homologue hMSH5:
ecombina ional DSB epai and non-synonymous polymo phic a ian s.
PLoS One 8(9): e73284.
Xu J, Mo Z, Ye D, Wang M, Liu F, Jin G, Xu C, Wang X, Shao Q, Chen Z,
Tao Z, Qi J, Zhou F, Wang Z, Fu Y, He D, Wei Q, Guo J, Wu D, Gao X,
Yuan J, Wang G, Xu Y, Wang G, Yao H, Dong P, Jiao Y, Shen M, Yang J,
Ou-Yang J, Jiang H, Zhu Y, Ren S, Zhang Z, Yin C, Gao X, Dai B, Hu Z,
Yang Y, Wu Q, Chen H, Peng P, Zheng Y, Zheng X, Xiang Y, Long J,
Gong J, Na R, Lin X, Yu H, Wang Z, Tao S, Feng J, Sun J, Liu W, Hsing A,
Rao J, Ding Q, Wiklund F, G onbe g H, Shu XO, Zheng W, Shen H, Jin L,
Shi R, Lu D, Zhang X, Sun J, Zheng SL, Sun Y (2012) Genome-wide
associa ion s udy in Chinese men iden i ies wo new p os a e
cance isk loci a 9q31.2 and 19q13.4. Na Gene 44(11): 1231–1235.
Yada S, Mukhopadhyay S, Anbalagan M, Mak idakis N (2015)
Soma ic mu a ions in ca aly ic co e o POLK epo ed in p os a e
cance al e anslesion DNA syn hesis. Hum Mu a 36(9): 873–880.
This wo k is licensed unde he C ea i e Commons
A ibu ion 4.0 In e na ional License. To iew a
copy o his license, isi h p://c ea i ecommons.o g/licenses/by/4.0/
Supplemen a y In o ma ion accompanies his pape on B i ish Jou nal o Cance websi e (h p://www.na u e.com/bjc)
1
The Ins i u e o Cance Resea ch & Royal Ma sden NHS Founda ion T us , 123 Old B omp on Rd, London SW7 3RP, UK;
2
Cen 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, S angeways
Labo a o y, Wo s Causeway, Camb idge CB1 8RN, UK;
3
Cance Epidemiology Cen e, The Cance Council Vic o ia, 1 Ra hdowne
S ee , Ca l on Vic o ia, Aus alia;
4
Cen e o Molecula , En i onmen al, Gene ic and Analy ic Epidemiology, The Uni e si y o
Melbou ne 3053, Vic o ia, Aus alia;
5
Depa men o Medical Epidemiology and Bios a is ics, Ka olinska Ins i u e, S ockholm
17177, Sweden;
6
Depa men o P e en i e Medicine, Keck School o Medicine, Uni e si y o Sou he n Cali o nia & No is
Comp ehensi e Cance Cen e , Los Angeles, CA 90089, USA;
7
Depa men o Medical Biochemis y and Gene ics, Uni e si y o
Tu ku, Tu ku, Finland;
8
Ins i u e o Biomedical Technology and BioMediTech, Uni e si y o Tampe e and FimLab Labo a o ies,
Tampe e 33520, Finland;
9
Depa men o Clinical Biochemis y, He le and Gen o e Hospi al, Copenhagen Uni e si y Hospi al,
He le Ring ej 75 DK-2730, He le , Denma k;
10
Cance Epidemiology Uni , Nu ield Depa men o Popula ion Heal h, Uni e si y
o Ox o d, Ox o d OX3 7LF, UK;
11
Su gical Oncology (U o-Oncology: S4), Uni e si y o Camb idge, Addenb ooke’s Hospi al,
Hills Road, Camb idge & Cance Resea ch UK Camb idge Resea ch Ins i u e, Li Ka Shing Cen e, Camb idge CB2 2QQ,
UK;
12
Uni e si y College London, Depa men o Applied Heal h Resea ch, 1-19 To ing on Place, London WC1E 7HB, UK;
13
Camb idge Ins i u e o Public Heal h, Uni e si y o Camb idge, Fo ie Si e, Robinson Way, Camb idge CB2 0SR, UK;
14
Depa men o Epidemiology, School o Public Heal h, Uni e si y o Washing on & Di ision o Public Heal h Sciences,
F ed Hu chinson Cance Resea ch Cen e , Sea le, WA, USA;
15
In e na ional Epidemiology Ins i u e, 1455 Resea ch Bl d., Sui e
550, Rock ille MD 20850, USA;
16
Mayo Clinic, Roches e , MN 55905, USA;
17
Ins i u e o Human Gene ics, Uni e si y Hospi al Ulm,
Ulm 89075, Ge many;
18
Di ision o U ologic Su ge y, B igham and Women’s Hospi al, Dana-Fa be Cance Ins i u e, 45 F ancis
S ee - ASB II-3 Bos on, MA, 02245, USA;
19
In e na ional He edi a y Cance Cen e , Depa men o Gene ics and Pa hology,
Pome anian Medical Uni e si y, Szczecin 70-115, Poland;
20
Di ision o Gene ic Epidemiology, Depa men o Medicine, Uni e si y o
U ah School o Medicine & Geo ge E. Wahlen Depa men o Ve e ans A ai s Medical Cen e , Sal Lake Ci y, UT 84132, USA;
21
Di ision o Clinical Epidemiology and Aging Resea ch, Ge man Cance Resea ch Cen e (DKFZ), Heidelbe g & Di ision o
P e en i e Oncology, Ge man Cance Resea ch Cen e (DKFZ) and Na ional Cen e o Tumo Diseases (NCT), Heidelbe g &
Ge man Cance Conso ium (DKTK), Ge man Cance Resea ch Cen e (DKFZ), Heidelbe g, Ge many;
22
Depa men o Cance
Epidemiology, H. Lee Mo i Cance Cen e , 12902 Magnolia D i e, Tampa, FL 33612, USA;
23
Molecula Medicine Cen e and
Depa men o Medical Chemis y and Biochemis y, Medical Uni e si y - So ia, 2 Zd a e S ee , So ia 1431, Bulga ia;
24
Aus alian
P os a e Cance Resea ch Cen e-Qld, Ins i u e o Heal h and Biomedical Inno a ion & School o Biomedical Science, Queensland
Uni e si y o Technology, B isbane 4102, Aus alia;
25
Biomedical Sciences Ins i u e (ICBAS), Po o Uni e si y, Po o, Po ugal;
26
Depa men o Gene ics, Po uguese Oncology Ins i u e, Po o, Po ugal 4200-072, Po ugal;
27
The Uni e si y o Su ey,
Guild o d, Su ey GU2 7XH, UK and
28
Wa wick Medical School, Uni e si y o Wa wick, Co en y CV4 7AL, UK
BRITISH JOURNAL OF CANCER DNA epai gene a ia ion and p os a e cance isk
952 www.bjcance .com | DOI:10.1038/bjc.2016.50