A genome-wide association study identifies a genomic region for the polycerate phenotype in sheep (Ovis aries)

1 Scien i ic RepoR s | 6:21111 | DOI: 10.1038/s ep21111 www.na u e.com/scien i ic epo s A genome-wide associa ion s udy iden i ies a genomic egion o he polyce a e pheno ype in sheep (O is a ies) Xue Ren1,2,*, Guang-Li Yang1,3,*, Wei-Feng Peng1,2, Yong-Xin Zhao1,2, Min Zhang1, Ze-Hui Chen1,2, Fu-An Wu4, Juha Kan anen5,6, Min Shen7,8 & Meng-Hua Li1 Ho ns a e a c anial appendage ound exclusi ely in Bo idae, and play impo an oles in accessing esou ces and ma es. In sheep (O ies a ies), ho ns a y om polled o six-ho ned, and human ha e been selec ing polled animals in a ming and b eeding. He e, we conduc ed a genome-wide associa ion s udy on 24 wo-ho ned e sus 22 ou -ho ned pheno ypes in a na i e Chinese b eed o Sishui Fu sheep. Toge he wi h linkage disequilib ium (LD) analyses and haplo ype-based associa ion es s, we iden i ied a genomic egion comp ising 132.0–133.1 Mb on ch omosome 2 ha con ained he op 10 SNPs (including 4 signi ican SNPs) and 5 mos signi ican haplo ypes associa ed wi h he polyce a e pheno ype. In humans and mice, his genomic egion con ains he HOXD gene clus e and adjacen unc ional genes EVX2 and KIAA1715, which ha e a close associa ion wi h he o ma ion o limbs and geni al buds. Ou esul s p o ide new insigh s in o he gene ic basis unde lying a iable numbe s o ho ns and ep esen a new esou ce o use in sheep gene ics and b eeding. In mammals, mos membe s o he in ao de Peco a, such as dee , an elope, ca le, goa , and sheep (O ies a ies), ha e c anial appendages. These appendages ha e ou ex an o ms: an le s, ho ns, p ongho ns and ossicones1. Ho ns, being exclusi e o bo ids, consis o a bony ho n co e co e ed by a scabba d-like ke a inous shea h ha is ne e shed1,2. Ho ns se e as a o m o sexual weapon y o mos domes ic and wild animals; hus, la ge ho ns ha e an ad an age in s ong in a-sexual compe i ion3,4. Howe e , ho ns may be disad an ages in li es ock a m- ing and b eeding. Fo example, ho ned males may a ack o he animals o people and can ge hei heads s uck in ences and eede s. The e o e, ho nless (o polled) indi iduals a e p e e ed o animal a ming and b eeding. A numbe o ea lie s udies ha e iden i ied candida e genes o a ious ai s in sheep, including coa colou , ho n, mea p oduc ion, g ow h and ecundi y e c., using a genome-wide associa ion app oach5–8. In sheep, a majo i y o ams ha e s ong and no mal ho ns, while ewes can ha e smalle ho ns, scu s (de o med es igial ho ns) o no ho ns (polled). P e ious s udies ha e p ima ily ocused on polled e sus ho ned pheno ypes o on ho n ype, size and leng h in sheep and hei wild ela i es (e.g., O is canadensis)3,4,8–10. Ho ns, a single au osomal locus mapped o a 7.4 cM egion on ch omosome 10 (OAR10) in Soay sheep, has been demons a ed o con ol ho n- ype polymo phism and exe s a compa a i ely la ge e ec on ho n size3,8. Fine mapping s udies iden i ied elaxin-like ecep o 2 (RXFP2), which con ibu es o main sex cha ac e s in humans and mice, as he majo can- dida e gene o Ho ns in Soay sheep3,4. Qui e ecen ly, a 1.8-kb inse ion in he 3′ -UTR o RXFP2 was iden i ied o be associa ed wi h polledness in sheep9. Also, whole-genome esequencing o a wild sheep O is canadensis e ealed ha mul iple popula ions sha ed a common selec i e sweep a RXFP210. 1CAS Key Labo a o y o Animal Ecology and Conse a ion Biology, Ins i u e o Zoology, Chinese Academy o Sciences (CAS), Beijing 100101, China. 2Uni e si y o Chinese Academy o Sciences (UCAS), Beijing 100049, China. 3Depa men o Li e Sciences, Shangqiu No mal Uni e si y, Shangqiu 476000, China. 4Bu eau o Animal Husband y and Ve e ina y Medicine, Sishui Coun y, Jining 273200, China. 5G een Technology, Na u al Resou ces Ins i u e Finland (Luke), Jokioinen 31600, Finland. 6Depa men o Biology, Uni e si y o Eas e n Finland, Kuopio 70211, Finland. 7Ins i u e o Animal Husband y and Ve e ina y Medicine, Xinjiang Academy o Ag icul u al and Reclama ion Sciences, Shihezi 832000, China. 8Key Labo a o y o Sheep B eeding and De elopmen Technology o Xinjiang P oduc ion and Cons uc ion C ops (XPCC), Shihezi 832000, China. *These au ho s con ibu ed equally o his wo k. Co espondence and eques s o ma e ials should be add essed o M.-H. L. (email: [email p o ec ed]) Recei ed: 24 Sep embe 2015 Accep ed: 18 Janua y 2016 Published: 17 Feb ua y 2016 OPEN www.na u e.com/scien i ic epo s/ 2 Scien i ic RepoR s | 6:21111 | DOI: 10.1038/s ep21111 Howe e , he gene ic basis o a di e en aspec o ho n pheno ype, he polyce a e ype (mul iple ho ns), has no been examined. Globally, he e a e only a small numbe o polyce a e na i e sheep b eeds, including Jacob, Manx Loagh an, Heb idean, Na ajo-Chu o, Icelandic sheep, Because nowadays he e a e no ou -ho ned Finnish Land ace sheep. Al ay sheep and Chinese Sishui Fu sheep (see Fig.1). Ho n pheno ypes o he b eeds ange om polled o as many as six ho ns. His o ically, ou -ho ned sheep we e widely dis ibu ed in bo h Eu ope and Asia11,12; howe e , hey ha e become a e due o long- e m a i icial selec ion. Mos o he ou -ho ned sheep a e ams, and ou ho ns a e a ely obse ed in ewes. The ou -ho nedness ai is dominan o e wo-ho nedness, bu ecessi e o polledness12. Sishui Fu sheep is a na i e sheep b eed in eas e n China, o igina ing om ancien Mongolian sheep. They a e a - ailed and mainly dis ibu ed in pas o al and ag icul u al a eas in Sishui Coun y o Shandong P o ince. They can p oduce high-quali y mea and ca pe wool. Also, hey ha e sound body con- o ma ion, s ong walking abili y and excellen adap a ion o local ecological en i onmen s13. Rams o he b eed ypically ha e 0–6 ho ns, and he occu ence o 2 and 4 ho ns is mos common (see Fig.1). Ewes a e always polled. Howe e , he unde lying mechanisms o he polyce a e pheno ype in ams o his and o he sheep b eeds ha e emained elusi e. He e, we ca ied ou a genome-wide associa ion (GWA) s udy and haplo ype-based associa ion es s o map a s ong candida e genomic egion o he polyce a e pheno ype in Sishui Fu sheep. Ou esul s will help o elucida e he molecula mechanism unde lying he ai and can be applied in molecula ma ke -assis ed b eeding p og ams. Me hods The me hods we e ca ied ou in acco dance wi h he app o ed guidelines o he Good Expe imen al P ac ices adop ed by he Ins i u e o Zoology, Chinese Academy o Sciences. All expe imen al p ocedu es and animal col- lec ions we e conduc ed unde a pe mi (No. IOZ13015) app o ed by he Commi ee o Animal Expe imen s o he Ins i u e o Zoology, Chinese Academy o Sciences, China. Sample collec ion and DNA ex ac ion. Ea issue samples we e collec ed om a o al o 60 ams o Sishui Fu sheep. Among hem, 26 indi iduals wi h ou ho ns we e assigned as he case g oup, and he o he 34 wo-ho ned sheep we e conside ed as he con ol g oup. All he samples we e collec ed om s ocks main ained on di e en a ms in Sishui Coun y, Shandong P o ince, China. Using bo h pedig ee eco ds and he a me s’ knowl- edge, pa icula e o s we e made o ensu e ha all animals we e ypical o he b eed and as un ela ed as possible. Genomic DNA was ex ac ed om he ea issues using a s anda d phenol/chlo o o m me hod14, and he DNA concen a ion was measu ed on a spec opho ome e (Nanod op 2000, Wilming on, DE, USA). The ex ac ed DNA was dilu ed o 100 ng/μ l o SNP BeadChip geno yping. SNP geno yping and quali y con ol. All sample DNA me he concen a ions ecommended o he Illumina O ine In inium HD SNP BeadChip geno yping acco ding o he manu ac u e ’s p o ocol. De ails on c ea ion o he o ine BeadChip (685,734 SNPs) and geno yping p ocedu es we e desc ibed in Ande son e al.15. Geno ypes o a o al o 606,006 sco able SNPs, which passed he manu ac u e ’s quali y con ol, we e a ailable o u he analysis. S ingen quali y con ol pa ame e s we e applied o bo h he samples and SNPs o ensu e eliabili y o esul s. We implemen ed he quali y con ol measu es using he so wa e PLINK 1.0716. SNPs o indi iduals who me any o he ollowing c i e ia we e emo ed: (1) no ch omosomal o physical loca ion; (2) mino allele e- quency (MAF) < 0.05; (3) indi iduals call a e < 0.9; (4) missing geno ype equency o SNP > 0.05; (5) Fishe ’s exac es 17 P- alue o Ha dy-Weinbe g equilib ium (HWE) < 0.001. We u he es ima ed pai wise ela edness using KING18,19, and closely ela ed animals (e.g., ull-sibs) we e excluded om u he analyses. A e il e ing, a o al o 491,507 SNPs and 50 indi iduals ( wo-ho ned: 28 animals; ou -ho ned: 22 animals) we e le o he wi hin-b eed popula ion s a i ica ion analysis. Figu e 1. Sishui Fu sheep wi h he (a) wo-ho ned and (b) ou -ho ned pheno ypes. www.na u e.com/scien i ic epo s/ 3 Scien i ic RepoR s | 6:21111 | DOI: 10.1038/s ep21111 Popula ion s a i ica ion and genome-wide associa ion (GWA) analysis. Gene ic di e en ia- ion be ween he case and con ol g oups was assessed by Wei & Cocke ham’s FST me hod20 using he p og am GENEPOP 4.221, and mul idimensional scaling (MDS) analysis using he p og am PLINK 1.07. The MDS analysis we e implemen ed on a se o 15,911 independen SNPs p uned using he op ion o “indep-pai wise 50 5 0.05” in PLINK. This unc ion calcula es pai wise linkage disequilib ium (LD) in a 50-SNPs-window, shi s a a pace o 5 SNPs and excludes one o a pai o SNPs i he LD es ima e 2 > 0.05. Resul s o he MDS analysis we e plo ed in he GenABEL package22 o R 3.2.2 (h p://www. -p ojec .o g). We u he emo ed he 4 sepa a ed indi iduals om he con ol g oup in he ollowing analyses (see Resul s). Ou inal wo king da a se con ained 46 indi iduals (24 wo-ho ned; 22 ou -ho ned). Because he pheno ype (2 o 4 ho ns) is bina y, we pe o med a genome-wide associa ion (GWA) s udy using he case-con ol model in he GenABEL package o R 3.2.2. Gi en he 15,911 p uned independen SNPs, he s a is ical h eshold o genome-wide signi icance a e Bon e oni co ec ion was 3.14 × 10−6 (0.05/15,911) a he empi ical le el o 0.053,23. To accoun o he GWA analysis con ounded by popula ion s a i ica ion, he genome-wide associa ion P- alues we e co ec ed using he p incipal componen s (PCs)22,23. We used he op i e p inciple componen s om he MDS analyses as co a ia es o adjus he sys ema ic biases (see Resul s), which was also indica ed by he Q-Q plo (quan ile-quan ile plo ). We u he conduc ed a esampling es (1,000 imes) wi hou eplacemen o de e mine he alse posi i e a e (FPR) by chance using he unc ion o “q sco e” in GenABEL. The command is de ailed as ollows: q sco e(sample(phda a(da a1)$dm2, eplace = F), da a1, ai . ype = “binomial”), whe e “da a1” is he gwaa. da a objec and “dm2” ep esen s pheno ype. Linkage disequilib ium (LD) and haplo ype-based associa ion analyses. The game ic LD meas- u e o 2 among SNPs wi hin he candida e genomic egion (132–133.1Mb on OAR2, see Resul s) was calcu- la ed and isualised using he p og am Haplo iew 4.224. LD blocks we e de ined based on he ou -game e ule algo i hm25. Pai wise es s o LD o he mos signi ican SNP s420183358 wi h i s lanking SNPs wi hin app oxima ely 1Mb ups eam and downs eam we e ob ained using PLINK 1.07. A egional associa ion plo was gene a ed using he R 3.2.2. Fu he , we applied a haplo ype-based associa ion es o iden i y haplo ypes signi ican ly associa ed wi h he polyce a e pheno ype ac oss OAR2. We i s es ima ed he haplo ypes and hei equencies in he LD blocks using he maximum likelihood (ML) me hod and expec a ion-maximiza ion (EM) algo i hm. Then, we pe - o med he chi-squa e es o haplo ype-pheno ype associa ion26, and s a is ical signi icance was also de e mined. We conside ed aw P- alues (P aw) < 1 × 10−7 as signi ican and P- alues < 0.05 a e 100,000 pe mu a ions as genome-wide signi ican (i.e. Pgenome < 0.05). All he analyses we e ca ied ou using Haplo iew. Gene anno a ion. Genes wi hin he candida e genomic egion we e de e mined using he O is a ies assem- bly Oa _ 4.0 (h p://www.ncbi.nlm.nih.go /genome/? e m= o i s+ a ies/). Howe e , gi en he lagging esea ch o he sheep genome, he gene anno a ion in sheep is incomple e. We ob ained addi ional unc ional in o ma ion o o hologous genes in o he species (e.g. human, mouse, chimpanzee, chicken and bo ine) om UniP o (h p:// www.unip o .o g/) and published esea ch ega ding he candida e genes iden i ied in sheep. Resul s Popula ion s a i ica ion and signi ican SNPs by GWA analysis. We ob ained a FST alue o 0.011 o he gene ic di e en ia ion be ween he case and con ol g oups. Also, he MDS plo (see Supplemen a y Fig. S1 online) indica ed wi hin-popula ion s a i ica ion, which may bias he associa ion es . In pa icula , 4 wo-ho ned animals o he con ol g oup showed appa en gene ic di e en ia ion om he es animals (see Supplemen a y Fig. S1 online) and, hus, we e emo ed om he GWA analysis. The aw genomic in la ion ac o lambda was 1.352 in he ini ial GWA analysis and u ned o be 1.103 a e applying he i s i e PCs in he popu- la ion s a i ica ion co ec ion (Fig.2a). All he op en associa ed SNPs we e iden i ied as being loca ed on OAR2 (Fig.2b, Table1) and ex emely low FPR alues (FPR < 0.001) we e obse ed o he SNPs a e he boo s apping es . We also de ec ed s a is ically signi ican di e ences in allele equencies o he en SNPs be ween he case and con ol g oups (pai ed - es ; P = 3.366 × 10−7, = 13.128, d. . = 9). O he op en SNPs, ou ( s420183358, s422753866, s428812278 and s428938943) showed signi ican associa ions wi h he polyce a e pheno ype a he genome-wide h eshold o P < 0.05 (Table1). SNP s420183358 showed he mos signi ican associa ion (P = 1.74 × 10−6). LD and haplo ype-based associa ion es . Pai wise es s o LD o he mos signi ican SNP s420183358 wi h neighbou ing SNPs in i s ups eam and downs eam egions indica ed ela i ely s ong LD ( 2 > 0.6) wi h wo SNPs ( s403275219 and s416536940) (Fig.3). A ela i ely high le el o LD was obse ed ac oss he pai wise SNPs in he candida e genomic egion (Fig.3). Howe e , he wo linked SNPs ( s403275219 and s416536940) we e dis ibu ed in a di e en haplo ype block om ha which con ained he mos signi ican SNP s420183358. Fou -game e ule algo i hm iden i ied a o al o 1,552 LD blocks con aining 44,373 haplo ypes on OAR2. The chi-squa e es e ealed ha a e 100,000 pe mu a ions 20 haplo ypes we e signi ican (P < 0.05) a he genome-wide h eshold on he ch omosome (Table2). The mos signi ican i e haplo ypes we e dis ibu ed in 5 adjacen and s ongly linked blocks (Table2, Supplemen a y Fig. S2 online). O he 5 mos signi ican haplo ypes, he op h ee haplo ype o blocks 6085 (AGAG, P < 1 × 10−5), 6086 (AGGGA, P < 1 × 10−5) and 6084 (GAGCA, P < 1 × 10−5) had he same equencies o 0.727 in he case g oups e sus 0.104, 0.104 and 0.125 in he con ol g oups, espec i ely (Table2). The esul s indica ed ha he egion 132.9–133.1 Mb mos p obably ha bou s he causa i e mu a ions accoun ing o he polyce a e pheno ype. www.na u e.com/scien i ic epo s/ 4 Scien i ic RepoR s | 6:21111 | DOI: 10.1038/s ep21111 Gene anno a ion. Based on he lis s o signi ican SNPs iden i ied by GWA s udy and he haplo ype-based associa ion es , he 132.0–133.1Mb egion on OAR2 was iden i ied as a s ong candida e genomic egion o he polyce a e pheno ype. Thi een genes ha e been anno a ed wi hin his genomic egion (see Supplemen a y Table S1 online). Me axin 2 (MXT2) is a p o ein ha is loca ed on he cy osolic ace o he mi ochond ial ou e memb ane and wo ks wi h me axin 1 o impo mi ochond ial p ep o eins in o mammalian mi ochond ia27. The HOXD gene clus e belongs o he homeobox amily o genes, which encode a highly conse ed amily o an- sc ip ion ac o s ha speci y di e ences in mo phogenesis in all mul icellula o ganisms28,29. Ea lie s udies in humans, mice and chicks ha e concluded ha 5′ HOXD genes a e closely associa ed wi h he de elopmen o limbs and geni alia30–34. EVX2 is adjacen o HOXD13, i.e., 8 kb ups eam, and o ms he EVX2-HOXD13 in e genic egion, which beha es as a spa io- empo al bounda y elemen 35. KIAA1715 is an o hologous gene o LUNPARK (LNP) in he human and mouse genomes, and i plays an impo an ole in he de elopmen o digi s and he cen- al ne ous sys em, along wi h EVX2 and he HOXD clus e 30. The mos signi ican SNP s420183358 iden i ied he e lies downs eam o he HOXD genes, and he o he h ee signi ican SNPs ( s422753866, s428812278 and s428938943) lie wi hin he gene KIAA1715. Figu e 2. Q-Q plo and manha an plo o genome-wide associa ion (GWA) es . The “p” in he labels ep esen s P- alues o GWA analysis. (a) Q-Q plo : he g ey and black do s deno e associa ion s a is ics be o e and a e co ec ion o popula ion s a i ica ion, espec i ely; (b) Manha an plo : he 5% genome-wide signi ican h eshold alue (P = 3.14 × 10−6) is p esen ed by a black dashed line. Ch . SNP Posi ion Allele MAF F equency P- alue FPRCase Con ol 2 s420183358 132042535 A:G 0.222 0.405 0.062 1.74E-06 < 0.001 2 s422753866 133065296 A:G 0.402 0.727 0.104 2.221E-06 < 0.001 2 s428812278 133084414 G:A 0.402 0.727 0.104 2.221E-06 < 0.001 2 s428938943 133008470 G:A 0.413 0.727 0.125 2.601E-06 < 0.001 2 s429526398 132934057 G:A 0.413 0.705 0.146 6.876E-06 < 0.001 2 s419153948 132994229 G:A 0.413 0.705 0.146 6.876E-06 < 0.001 2 s416536940 132519432 A:G 0.217 0.386 0.062 7.482E-06 < 0.001 2 s403336550 133065371 G:A 0.446 0.75 0.167 7.518E-06 < 0.001 2 s418381175 132200980 G:A 0.391 0.818 0.417 8.912E-06 < 0.001 2 s408254270 132201113 G:A 0.391 0.818 0.417 8.912E-06 < 0.001 Table 1. The op en SNPs associa ed wi h he polyce a e pheno ype. Ch . is he abb e ia ion o ma o ch omosome. MAF ep esen s mino allele equency. P- alue ep esen s he co ec ed signi icance o GWA a e p inciple componen adjus men . FPR means alse posi i e a e. www.na u e.com/scien i ic epo s/ 5 Scien i ic RepoR s | 6:21111 | DOI: 10.1038/s ep21111 Discussion Unde s anding he gene ic a chi ec u e o ho n pheno ypic a ia ion in wild and domes ic e eb a es is one o he undamen al goals in e olu iona y gene ics. In his s udy, we iden i ied a genomic egion o he polyce a e pheno ype using a GWA s udy app oach. The mos signi ican SNP ( s420183358) is no loca ed in any anno a ed gene, bu downs eam o he HOXD gene clus e . Thus, s420183358 may be he causal mu a ion, which has a egula o y unc ion. Al e na i ely, i is also possible ha he causal mu a ion is loca ed inside he HOXD gene clus e and is closely linked o he a ge SNP s420183358. The candida e genomic egion (132–133.1 Mb) e ealed in his in es iga ion is in acco dance wi h an ongoing in es iga ion on he gene ic cause o he polyce a e ai in Na ajo-Chu o and Jacob sheep, whe e a genomic egion o 132Mb on OAR2 was iden i ied as being associa ed wi h his ai 36. This egion is well-known o he HOXD gene clus e , which comp ises nine genes (HOXD1, HOXD3, HOXD4 and HOXD8-HOXD13) in o de om 3′ o 5′ . HOXD genes a e he p ima y de e minan s o he an e io -pos e io body axis in all bila e ians, including he unk o appendicula axis37,38. HOXD11 is exp essed along he p i- ma y body axis, and HOXD11 mu an mice (o mice wi h a dele ion o HOXD13 o HOXD11) exhibi a supe nume a y lumba e eb a39,40. HOXD13 and HOXA13 a e mos s ongly exp essed a he dis al ends o limbs and geni al buds, and he absence o he wo genes could a ec he de elopmen o hands and ee 31,32,34. Synpolydac yly, a limb mal o ma ion, is caused by HOXD13 mu a ions41, and la e esea ch sugges ed ha emo - ing HOXD9-HOXD13 and EVX2 also led o synpolydac yly35. EVX2 is loca ed 8kb ups eam o HOXD13 and esponds o he digi enhance in a simila way as he HOXD genes35,42. The signi ican ly associa ed haplo ypes we e dis ibu ed in se e al LD blocks (see Supplemen a y Fig. S2 online). The h ee mos signi ican haplo ypes (in blocks 6084–6086) con aining h ee signi ican SNPs Figu e 3. Plo o egional associa ion esul s o loci su ounding he mos signi ican SNP s420183358 ( ed hombus). Di e en colou s ep esen he 2 alues o pai -wise LD es ima es. Func ional genes in his egion a e plo ed in he box. The hombus ep esen s he op en SNPs o GWA analysis. The yellow do ep esen s s403275219. LD Block Haplo ype Posi ion F equency P- alue P100,000 i s las Case Con ol 6085 AGAG 133040292 133065679 0.727 0.104 1.14E-09 < 0.00001 6086 AGGGA 133072339 133084414 0.727 0.104 1.14E-09 < 0.00001 6084 GAGCA 133008423 133017771 0.727 0.125 4.61E-09 < 0.00001 6081 ACCGA 132934057 132951006 0.705 0.146 5.43E-08 0.0005 6083 AAAG 132978078 132994332 0.705 0.146 5.43E-08 0.0005 11095 GA 240198021 240198140 0.136 0.646 6.42E-07 0.0096 6125 GGAG 133917957 133928107 0.455 0.021 7.38E-07 0.0105 6107 GGGG 133560053 133564962 0.841 0.333 8.60E-07 0.0120 11110 AGACC 240440866 240460254 0.592 0.106 8.82E-07 0.0120 6097 GCGGGG 133339051 133365628 0.523 0.062 9.72E-07 0.0128 Table 2. The op en signi ican haplo ypes wi h he polyce a e pheno ype. P100,000 ep esen s P- alue wi h 100,000 pe mu a ions in he associa ion es s; LD deno es linkage disequilib ium. www.na u e.com/scien i ic epo s/ 6 Scien i ic RepoR s | 6:21111 | DOI: 10.1038/s ep21111 ( s422753866, s428812278 and s428938943) we e loca ed wi hin he gene KIAA1715, which has he unc ions associa ed wi h he o ma ion o limb digi s and geni al buds30,41,43. The HOXD gene clus e and genes EVX2 and KIAA1715 a e loca ed in he candida e genomic egion in a successi e o de (Fig.3). All hese genes ha e analo- gous unc ions in limb and digi de elopmen , and he numbe o digi s and lumba e eb as a e con olled by his egion30. Also, p e ious wo k localized a “digi enhance ” ups eam om he HOXD clus e 39. Subsequen s udies demons a ed ha a global con ol egion (GCR) de ined a ch omosomal egula o y landscape con aining he HOXD gene clus e , wi h LNP, EVX2 and HOXD-speci ic pa e ns30. The e o e, he GCR could also egula e he polyce a e pheno ype. Ou app oaches o de ec ing he candida e genomic egion unde lying he polyce a e pheno ype we e e i- cien , al hough he ela i ely small sample size was one possible po en ial limi a ion o his s udy. Fu u e in es- iga ions wi h a la ge sample size in o he sheep b eeds and deep-sequencing o he candida e genomic egion a e necessa y o con i m and ex end ou indings. In conclusion, ou esul s p o ided s ong e idence ha he genomic egion 132.0–133.1 Mb on ch omosome 2 is associa ed wi h he polyce a e pheno ype. Re e ences 1. Da is, E. B., B ako a, K. A. & Lee, A. H. E olu ion o uminan headgea : a e iew. P oc. R. Soc. Lond. B Biol. Sci. 278, 2857–2865 (2011). 2. Do e, W. F. The physiology o ho n g ow h: a s udy o he mo phogenesis, he in e ac ion o issues, and he e olu iona y p ocesses o a Mendelian ecessi e cha ac e by means o ansplan a ion o issues. J. Exp. Zool. 69, 347–405 (1935). 3. Johns on, S. E. e al. Genome-wide associa ion mapping iden i ies he gene ic basis o disc e e and quan i a i e a ia ion in sexual weapon y in a wild sheep popula ion. Mol. Ecol. 20, 2555–2566 (2011). 4. Johns on, S. E. e al. Li e his o y ade-o s a a single locus main ain sexually selec ed gene ic a ia ion. Na u e 502, 93–96 (2013). 5. Li, M. H., Tii ikka, T. & Kan anen, J. A genome-wide scan s udy iden i ies a single nucleo ide subs i u ion in ASIP associa ed wi h whi e e sus non-whi e coa -colou a ia ion in sheep (O is a ies). He edi y 112, 122–131 (2014). 6. Dema s, J. e al. Genome-wide associa ion s udies iden i y wo no el BMP15 mu a ions esponsible o an a ypical hype p oli icacy pheno ype in sheep. PLoS Gene . 9, e1003482 (2013). 7. Zhang, L. e al. Genome-wide associa ion s udies o g ow h and mea p oduc ion ai s in sheep. PLoS One 8, e66569 (2013). 8. Johns on, S. E., Be aldi, D., McRae, A. F., Pembe on, J. M. & Sla e, J. Ho n ype and ho n leng h genes map o he same ch omosomal egion in Soay sheep. He edi y 104, 196–205 (2010). 9. Wiedema , N. & D ögemülle , C. A 1.8-kb inse ion in he 3′ -UTR o RXFP2 is associa ed wi h polledness in sheep. Anim. Gene . 46, 457–461 (2015). 10. Ka dos, M. e al. Whole-genome esequencing unco e s molecula signa u es o na u al and sexual selec ion in wild bigho n sheep. Mol. Ecol. 24, 5616–5632 (2015). 11. Ryde , M. L. in Sheep and man, (Duckwo h, London, 1983). 12. Dý mundsson, Ó. R. Fou -ho nedness; a a e peculia i y s ill ound in Icelandic sheep. The Icelandic Sheep B eede s o No h Ame ica Newsle e 9, 6–8 (2005). 13. China Na ional Commission o Animal Gene ic Resou ces. in Animal gene ic esou ces in China sheep and goa s, 56–58 (China Ag icul u e P ess, Beijing, 2011). 14. Köchl, S., Niede s ä e , H. & Pa son, W. DNA ex ac ion and quan i a ion o o ensic samples using he phenol-chlo o o m me hod and eal- ime PCR. Me hods Mol. Biol. 297, 13–29 (2005). 15. Ande son, R. De elopmen o a high densi y (600K) Illumina O ine SNP chip and i s use o ine map he yellow a locus. in Plan and Animal Genome XXII Con e ence (Plan and Animal Genome, 2014). 16. Pu cell, S. e al. PLINK: a ool se o whole-genome associa ion and popula ion-based linkage analyses. Am. J. Hum. Gene . 81, 559–575 (2007). 17. Raymond, M. & Rousse , F. An exac es o popula ion di e en ia ion. E olu ion 49, 1280–1283 (1995). 18. Manichaikul, A. e al. Robus ela ionship in e ence in genome-wide associa ion s udies. Bioin o ma ics 26, 2867–2873 (2010). 19. Li, M. H., S anden I., Tii ikka T., Se on-Aimonen, M. L. & Kan anen J. A compa ison o app oaches o es ima e he inb eeding coe icien and pai wise ela edness using genomic and pedig ee da a in a sheep popula ion. PLoS One 6, e26256 (2011). 20. Wei , B. S. & Cocke ham, C. C. Es ima ing F-s a is ics o he analysis o popula ion s uc u e. E olu ion 38, 1358–1370 (1984). 21. Raymond, M. & Rousse , F. GENEPOP ( e sion 1.2): popula ion gene ics so wa e o exac es s and ecumenicism. J. He ed. 86, 248–249 (1995). 22. Aulchenko, Y. S., Ripke, S., Isaacs, A. & an Duijn, C. M. GenABEL: an R lib a y o genome-wide associa ion analysis. Bioin o ma ics 23, 1294–1296 (2007). 23. Johns on, S. E. e al. Genome-wide SNP analysis e eals a gene ic basis o sea-age a ia ion in a wild popula ion o A lan ic salmon (Salmo sala ). Mol. Ecol. 23, 3452–3468 (2014). 24. Ba e , J. C., F y, B., Malle , J. & Daly, M. J. Haplo iew: analysis and isualiza ion o LD and haplo ype maps. Bioin o ma ics 21, 263–265 (2005). 25. Wang, N., Akey, J. M., Zhang, K., Chak abo y, R. & Jin, L. Dis ibu ion o ecombina ion c osso e s and he o igin o haplo ype blocks: he in e play o popula ion his o y, ecombina ion, and mu a ion. Am. J. Hum. Gene . 71, 1227–1234 (2002). 26. Wang, Q. e al. Genome-wide haplo ype associa ion s udy iden i ies BLM as a isk gene o p os a e cance in Chinese popula ion. Tumo Biol. 36, 2703–2707 (2015). 27. A ms ong, L. C., Saenz, A. J. & Bo ns ein, P. Me axin 1 in e ac s wi h me axin 2, a no el ela ed p o ein associa ed wi h he mammalian mi ochond ial ou e memb ane. J. Cell Biochem. 74, 11–22 (1999). 28. K umlau , R. Hox genes in e eb a e de elopmen . Cell 78, 191–201 (1994). 29. Acampo a, D. e al. The human HOX gene amily. Nucleic acids Res. 17, 10385–10402 (1989). 30. Spi z, F., Gonzalez, F. & Duboule, D. A global con ol egion de ines a ch omosomal egula o y landscape con aining he HoxD clus e . Cell 113, 405–417 (2003). 31. Delp e i, S., Zakany, J. & Duboule, D. A unc ion o all pos e io Hoxd genes du ing digi de elopmen . De . Dynam. 241, 792–802 (2012). 32. And ey, G. e al. A swi ch be ween opological domains unde lies HoxD genes collinea i y in mouse limbs. Science 340, 1234167 (2013). 33. Kmi a, M., F audeau, N., He aul , Y. & Duboule, D. Se ial dele ions and duplica ions sugges a mechanism o he collinea i y o Hoxd genes in limbs. Na u e 420, 145–150 (2002). 34. Cobb, J. & Duboule, D. Compa a i e analysis o genes downs eam o he Hoxd clus e in de eloping digi s and ex e nal geni alia. De elopmen 132, 3055–3067 (2005). 35. Goodman, F. R., Majewski, F., Collins, A. L. & Scamble , P. J. A 117-kb mic odele ion emo ing HOXD9-HOXD13 and EVX2 causes synpolydac yly. Am. J. Hum. Gene . 70, 547–555 (2002). www.na u e.com/scien i ic epo s/ 7 Scien i ic RepoR s | 6:21111 | DOI: 10.1038/s ep21111 36. Had ield, T. The polyce a e ai in Jacob and Na ajo-Chu o sheep may be loca ed on O ine Ch omosome 2. in Plan and Animal Genome XXIII Con e ence (Plan and Animal Genome, 2015). 37. Zha, Y. e al. Func ional dissec ion o HOXD clus e genes in egula ion o neu oblas oma cell p oli e a ion and di e en ia ion. PLoS One 7, e40728 (2012). 38. So dino, P. & Duboule, D. A molecula app oach o he e olu ion o e eb a e pai ed appendages. T ends Ecol. E ol. 11, 114–119 (1996). 39. Spi z, F. e al. La ge scale ansgenic and clus e dele ion analysis o he HoxD complex sepa a e an ances al egula o y module om e olu iona y inno a ions. Genes De . 15, 2209–2214 (2001). 40. Da is, A. P. & Capecchi, M. R. Axial homeosis and appendicula skele on de ec s in mice wi h a a ge ed dis up ion o hoxd-11. De elopmen 120, 2187–2198 (1994). 41. Aka su, A. N., S oilo , I., Yilmaz, E., Sayli, B. S. & Sa a azi, M. Genomic s uc u e o HOXD13 gene: a nine polyalanine duplica ion causes synpolydac yly in wo un ela ed amilies. Hum. Mol. Gene . 5, 945–952 (1996). 42. Yamagishi, T., Ozawa, M., Oh suka, C., Ohyama-Go o, R. & Kondo, T. E x2-Hoxd13 in e genic egion es ic s enhance associa ion o Hoxd13 p omo e . PLoS One 2, e175 (2007). 43. Dlugaszewska, B. e al. B eakpoin s a ound he HOXD clus e esul in a ious limb mal o ma ions. J. Med. Gene . 43, 111–118 (2006). Acknowledgemen s This wo k was suppo ed by he 100- alen P og am o he Chinese Academy o Sciences (CAS), he Na ional High Technology Resea ch and De elopmen P og am o China (i.e., 863 P og am, g an No. 2013AA102506), g an s om he Na ional Na u al Science Founda ion o China (g an s Nos. 31272413 and U1303284), he Na ional T ansgenic B eeding P ojec o China (2014ZX0800952B) and he Academy o Finland (g an s No. 250633 and No. 256077). Au ho Con ibu ions M.-H.L. concei ed and designed he p ojec . G.-L.Y. and F.-A.W. collec ed he samples. W.-F.P., Y.-X.Z., M.Z. and Z.-H.C ex ac ed he DNA. J.K. helped o p oduce he SNP chip da a. X.R., G.-L.Y. and M.S. analysed he da a. X.R. and G.-L.Y. w o e he pape , wi h con ibu ions om M.-H.L. All au ho s e iewed and app o ed he inal manusc ip . Addi ional In o ma ion Supplemen a y in o ma ion accompanies his pape a h p://www.na u e.com/s ep Compe ing inancial in e es s: The au ho s decla e no compe ing inancial in e es s. How o ci e his a icle: Ren, X. e al. A genome-wide associa ion s udy iden i ies a genomic egion o he polyce a e pheno ype in sheep (O is a ies). Sci. Rep. 6, 21111; doi: 10.1038/s ep21111 (2016). This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. 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 he c edi line; i he ma e ial is no included unde he C ea i e Commons license, use s will need o ob ain pe mission om he license holde o ep oduce he ma e ial. To iew a copy o his license, isi h p://c ea i ecommons.o g/licenses/by/4.0/ 1 Scien i ic RepoR s | 6:25322 | DOI: 10.1038/s ep25322 www.na u e.com/scien i ic epo s E a um: A genome-wide associa ion s udy iden i ies a genomic egion o he polyce a e pheno ype in sheep (O is a ies) Xue Ren, Guang-Li Yang, Wei-Feng Peng, Yong-Xin Zhao, Min Zhang, Ze-Hui Chen, Fu-An Wu, Juha Kan anen, Min Shen & Meng-Hua Li Scien i ic Repo s 6:21111; doi: 10.1038/s ep21111; published online 17 Feb ua y 2016; upda ed on 20 May 2016 In his A icle, he e is an e o in he In oduc ion sec ion. “Globally, he e a e only a small numbe o polyce a e na i e sheep b eeds, including Jacob, Manx Loagh an, Heb idean, Na ajo-Chu o, Icelandic sheep, Because nowadays he e a e no ou -ho ned Finnish Land ace sheep. Al ay sheep and Chinese Sishui Fu sheep (see Fig. 1)”. should ead: “Globally, he e a e only a small numbe o polyce a e na i e sheep b eeds, including Jacob, Manx Loagh an, Heb idean, Na ajo-Chu o, Icelandic sheep, Al ay sheep and Chinese Sishui Fu sheep (see Fig. 1)”. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. 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 he c edi line; i he ma e ial is no included unde he C ea i e Commons license, use s will need o ob ain pe mission om he license holde o ep oduce he ma e ial. To iew a copy o his license, isi h p://c ea i ecommons.o g/licenses/by/4.0/ OPEN