Genetic diversity, population structure and linkage disequilibrium in Nordic spring barley (Hordeum vulgare L. subsp. vulgare)

RESEARCH ARTICLE Gene ic di e si y, popula ion s uc u e and linkage disequilib ium in No dic sp ing ba ley (Ho deum ulga e L. subsp. ulga e) The e ´se Beng sson .The PPP Ba ley Conso ium .Ou i Manninen . Ahmed Jahoo .Jihad O abi Recei ed: 29 Sep embe 2016 / Accep ed: 24 Janua y 2017 ÓThe Au ho (s) 2017. This a icle is published wi h open access a Sp inge link.com Abs ac Gene ic di e si y, popula ion s uc u e and genome-wide linkage disequilib ium (LD) was es i- ma ed in No dic sp ing ba ley (Ho deum ulga e L. subsp. ulga e) by geno yping 180 b eeding lines wi h 48 SSR ma ke s and 7842 high-con idence SNPs using he Illumina In inium 9K assay. In o al 6208 SNPs we e polymo phic and selec ed o u he s a is ical analysis. A Man el es e ealed a s ong posi i e co ela ion wi h a Pea son’s co ela ion coe icien ( ) o 0.86, be ween he es ima es o gene ic dis ances based on SSR and SNP da a. Popula ion s uc u e analysis iden i ied wo g oups wi h a clea ances y and one g oup wi h an admixed ances y. The g oups we e p ima ily sepa a ed based on ow- ype and geog aphical o igin. A e age LD o he whole popula ion decayed below a c i ical le el o 2 =0.20 wi hin a ange o 0–4 cM. To a oid con ounding e ec s o he s ong popula ion s uc u e, LD decay o he di e en g oups was analysed sepa a ely and anged om 0 o 12 cM. A slowe LD decay was ound wi hin he wo- owed lines compa ed o he six- owed lines and he wo- owed lines o igina ing om he no he n pa , which could be he esul o s ong selec ion o mal ing quali y and yield in he sou he n pa . No la ge di e ence in gene ic di e si y was obse ed be ween popula ion sub-g oups, bu di e ences a ce ain ch omosomal egions we e e iden . Keywo ds Ho deum ulga e L. Linkage disequilib ium Mic osa elli es Molecula ma ke s  Plan b eeding SNP a ay In oduc ion Cul i a ed ba ley (Ho deum ulga e L. subsp. ul- ga e) is one o he mos impo an c ops in he No dic coun ies co e ing a o al a ea o 1.56 million ha in 2014 (h p:// aos a 3. ao.o g). No dic ba ley b eeding began mo e han 100 yea s ago, s a ing by selec ions in he land ace gene pool o he p esen mode n eli e cul i a s de eloped by c osses be ween pu e lines o ad anced ma e ial (Kolodinska B an es am e al. 2004; Fischbeck 1992). Exo ic sou ces a e used in mode n b eeding, bu hen p ima ily as dono s o single esis ance genes agains majo diseases Elec onic supplemen a y ma e ial The online e sion o his a icle (doi:10.1007/s10722-017-0493-5) con ains supple- men a y ma e ial, which is a ailable o au ho ized use s. T. Beng sson (&)A. Jahoo Depa men o Plan B eeding, Swedish Uni e si y o Ag icul u al Sciences, Box 101, 230 53 Alna p, Sweden e-mail: [email p o ec ed] O. Manninen Bo eal Plan B eeding L d, Mylly ie 10, 31600 Jokioinen, Finland A. Jahoo J. O abi No dic Seed A/S, Ko nma ken 1, 8464 Gal en, Denma k 123 Gene Resou C op E ol DOI 10.1007/s10722-017-0493-5 (Melchinge e al. 1994; Weibull e al. 2003). Today’s cul i a ion o gene ically uni o m cul i a s is aising conce ns abou loss o gene ic di e si y. A p e ious s udy o he gene ic di e si y in sp ing ba ley ge m- plasm in he No dic and Bal ic egion epo ed a sig- ni ican dec ease o gene ic di e si y in he sp ing ba ley om sou he n pa s o he in es iga ed egion in he middle o he wen ie h cen u y, bu no in he sp ing ba ley om he no he n pa s (Kolodinska B an es am e al. 2007). Likewise, no signs o gene ic e osion we e obse ed in a ecen s udy o gene ic di e si y o ba leys om he no he n Eu opean a ea o e a hund ed yea s o ba ley b eeding (Rajala e al. 2016). This highligh s he impo ance o knowledge ega ding he le el o gene ic di e si y in b eeding ma e ial, since i enables he de ec ion o any changes in di e si y ha migh lead o gene ic e osion. Se e al ypes o molecula ma ke s such as ampli- ied agmen leng h polymo phism (AFLP), andom ampli ied polymo phic DNA (RAPD), simple sequence epea s (SSR), di e si y a ay echnology (DA T), and single nucleo ide polymo phism (SNP) ha e been used o s udy gene ic di e si y and s uc u e in c ops (Kesawa and Das Kuma 2009). SSRs ha e he ad an age o be abundan , highly polymo phic and mul i-allelic, and he e o e o en p o ide mo e in o - ma ion compa ed o biallelic ma ke s such as SNPs. On he o he hand, he new iSelec geno yping pla o m, based on he Illumina In inium assay, allows he simul aneous es ing o 7842 gene-de i ed SNPs (Comad an e al. 2012). The gene ic polymo phism o he SNP and SSR ma ke sys ems a e gene a ed h ough di e en mechanisms, hus hey could gi e di e en iews o he s uc u e o a popula ion. Analysing he gene ic s uc u e wi hin a popula ion is a c i ical s ep o he way o unde s and and e eal he complexi y wi hin his popula ion (P i cha d e al. 2000). Fac o s such as human o en i onmen ally d i en selec ion, gene ic d i , ma ing sys em and g ow h habi can ha e an e ec on he popula ion s uc u e (Buckle and Tho nsbe y 2002; Flin -Ga cia e al. 2003). S udies o wo ldwide (Malyshe a-O o e al. 2006), Eu opean (Ros oks e al. 2006), Ame ican (Hamblin e al. 2010) and No dic (Rajala e al. 2016) ba ley ge mplasm ha e shown ha cul i a ed ba ley has a clea le el o popula ion s uc u e wi h majo subpopula ions caused by di e ences in ea ype, i.e. wo- ow and six- ow, and seasonal g ow h habi , i.e. win e and sp ing (Hamblin e al. 2010; Malyshe a- O o e al. 2006; Ros oks e al. 2006). In addi ion o hese majo subpopula ions, i has been shown ha Ame ican ba ley accessions u he can be di ided in o mino sub-popula ions co esponding o he b eeding p og ams, which migh be alloca ed o limi ed exchange o ma e ial be ween he b eeding p og ams and/o a esul o local adap a ion (Hamblin e al. 2010). Ano he impo an ac o o conside is linkage disequilib ium (LD), which is he non- andom asso- cia ion o alleles be ween wo loci and shows he co ela ion be ween gene ic polymo phisms, e.g. SNPs, and hei his o y o mu a ions and ecombina- ion (Flin -Ga cia e al. 2003). LD is impo an since he a e o i s decay in a gi en species de e mines he numbe and densi y o he molecula ma ke s needed o pe o m GWAS (Ra alski 2002). In many sel - pollina ed species such as ba ley whe e LD ex ends o e long ch omosomal dis ances (Malyshe a-O o e al. 2006), ewe ma ke s a e needed o co e he whole genome, whe eas a highe ma ke densi y is needed when LD decays e y apidly in species such as maize whe e i declines o nominal le els wi hin 1.5 kb (Reming on e al. 2001). The aim o he Public P i a e Pa ne ship (PPP) o p e-b eeding in ba ley, pa ly unded by he No dic Council o Minis e s (NMR), is o lay a ounda ion o ba ley b eeding o disease esis ance and yield s abili y o mee cu en and u u e challenges in he No dic egion. This collabo a ion is be ween i e b eeding companies and h ee go e nmen al o gani- za ions in he No dic egion. One o he goals wi h his p og am is o iden i y ma ke s linked wi h ai s o in e es ia genome-wide associa ion s udies (GWAS). The main objec i e wi h he p esen s udy is o de e mine he popula ion s uc u e and he LD decay in a No dic ba ley panel, in o de o es ima e he ela ionships among indi iduals. Ma e ials and me hods Plan ma e ials A o al o 134 and 46 sp ing ba ley Ho deum ulga e L. subsp. ulga e b eeding lines and cul i a s, espec- i ely, we e included in his s udy. The selec ed sp ing ba ley b eeding lines and cul i a s a e he ea e e e ed o as lines. Equal numbe o lines was selec ed Gene Resou C op E ol 123 by b eede s om each o Bo eal Plan B eeding (Finland), G amino B eeding AS (No way), Ag icul- u al Uni e si y o Iceland (AUI Iceland), Lan ma ¨n- nen Lan b uk (LSW Sweden), No dic Seed and Seje Plan e o aedling I/S (Denma k). The lines we e cho- sen o ep esen he a ailable gene ic a ia ion in cu en eli e No dic ba ley ge mplasm. Ou o he 180 lines, ele en lines we e emo ed om u he analyses since hey we e duplica es, o due o incomple e geno yping. Ou o he emaining 169 lines, 124 we e wo- owed and 45 six- owed. DNA ex ac ion DNA was ex ac ed om 2-week-old seedlings, using a CTAB (Ce yl T ime hyl Ammonium B omide) me hod as desc ibed ea lie by (O abi e al. 2014). The DNA was p ecipi a ed wi h isop opanol, washed wo imes wi h 75% e hanol, ai -d ied and inally dilu ed in TE bu e (pH 8.0). Mic osa elli e geno yping All lines we e geno yped using 48 mic osa elli e ma ke s e enly dis ibu ed o e all ch omosomes. PCR ampli ica ions we e pe o med on a GeneAmp Ò PCR Sys em 2700 he mal cycle (Applied Biosys- ems, Fos e Ci y, CA, USA) using a single uni e sal ouchdown PCR p og am as p e iously desc ibed in (O abi e al. 2014). The o wa d p ime s we e 50- labeled wi h luo escen dyes o achie e he maximum mul iplex capaci y o he ABI 3130xl sequence . Di ec and M13-labelling we e used o he mic osa elli e agmen s, wi h 6-ca boxy luo escein (6-FAM, blue) o hexachlo o-6-ca boxy luo escein (HEX o VIC, g een) and 50- luo escein phospho- amidi e (NED, yellow) o he di ec labelling. Fo he M-13 labelling, 6-FAM (blue), VIC (g een) and NED (yellow) we e used. Fo agmen de ec ion he ABI 3130xl DNA analyze (Applied Biosys ems, Fos e Ci y, CA, USA) was used and he agmen analysis and geno yping we e pe o med using he GeneMa - ke geno yping so wa e p og am, e sion 1.85 (So gene ics, S a e College, PA, USA). SNP geno yping All lines we e geno yped wi h he ba ley iSelec SNP chip based on he Illumina In inium 9K assay. The geno yping o he lines was ou sou ced o T ai Gene - ics. The chip consis s o 7842 high-con idence SNPs de i ed om exp essed genes (Comad an e al. 2012). Da a analysis Gene di e si y and ma ke allele equency Gene ic dis ances be ween geno ypes, gene ic di e - si y, allele equency and p i a e alleles (alleles p esen only in one g oup) we e calcula ed using an in-house p og am w i en in VBA (Visual Basic o Applica ions) and implemen ed in Mic oso Excel 2007 (Mic oso , Redmond, WA, USA). The p og am u ilises R language so wa e .2.14.2 (R De elopmen Co e Team 2012), which includes he Mode n Applied S a is ics wi h S-plus (MASS) package (Ripley 2002). The a e age numbe o alleles pe locus pe g oup ep esen s how polymo phic a gi en ma ke was wi hin each g oup, and his alue was calcula ed o each SNP ma ke . The a e age numbe o alleles pe ma ke is be ween 1 and 2, whe e ma ke s wi h a numbe o 1 we e conside ed monomo phic. Modi ied Roge ’s dis ances (MRD) based on (W igh 1978) we e calcula ed o he SSR da a based on he ollowing equa ion: MRD ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 2mXm i¼1Xai k¼1ðpij qijÞ2 whe e p ij and q ij a e he allele equencies o he j h allele a he i h ma ke o he wo ba ley lines in conside a ion; a i is he numbe o alleles a he i h ma ke ; and mis he numbe o SSR loci. The gene ic dis ances (D SM ) based on he SNP da a we e de e - mined by es ima ing a simple ma ching coe icien (S SM ) (Rei e al. 2005): DSM ¼1SSM;SSM ¼ ij þyij ij þwij þxij þyij whe e ij e e s o he allele in common be ween wo lines, iand j;w ij is he numbe o alleles p esen in i and absen in j;x ij is he numbe o alleles p esen in j and absen in i; and y ij is he numbe o alleles absen in bo h iand j. Co ela ions be ween he MRD (SSR) and D SM (SNP) ma ices we e calcula ed using Man el es (Man el 1967). The polymo phic in o ma ion con en (PIC) o he indi idual ma ke s was calcula ed as explained by Bo s ein e al. (1980): Gene Resou C op E ol 123 PIC ¼1X n i¼1 p2 iX n1 i¼1X n j¼iþ1 2p2 ip2 j whe e p i is he equency o he i h allele, and nis he numbe o alleles pe ma ke . To show gene ic di e si y and di e en ia ion along he ba ley ch omosomes Shannon’s di e si y indices we e calcula ed using GenAlEx . 6.5.0.1 (Peakall and Smouse 2006,2012) based on he SNP da a. Popula ion s uc u e analysis To de e mine popula ion s uc u e o he ba ley lines using SNP ma ke s, he so wa e package STRUC- TURE .2.3.4 based on a Bayesian clus e ing app oach, was used (P i cha d e al. 2000). STRUC- TURE was un 10 imes o each hypo he ical numbe o subpopula ions (K) be ween 1 and 12 wi h he ploidy le el se as 2. The Ma ko chain Mon e Ca lo (MCMC) was se o 9999 bu n-in phases ollowed by 9999 i e a ions. S uc u e Ha es e .0.6.94 (Ea l and on Hold 2012), was used o es ima e he mos likely numbe o g oups (K) using he Del aK me hod (E anno e al. 2005). Popula ion s uc u e based on he SSR ma ke s was calcula ed as desc ibed abo e wi h he same se ings bu by using he p e iously desc ibed VBA p og am in Excel. Analysis o Molecula Va iance (AMOVA), Nei’s unbiased gene ic dis ance and P incipal Coo dina es Analysis (PCoA) we e calcula ed using GenAlEx . 6.5.0.1. Linkage disequilib ium analysis The TASSEL 3.0 so wa e (h p.//www. maizegene ics.ne ) was used o calcula e he LD (al- lele equency co ela ion, 2 ) es ima es be ween he SNP ma ke pai s using he ull ma ix op ion. Only in a-ch omosomal compa isons we e included and ma ke s wi h mino allele equency (MAF) below 0.05 we e excluded. Thus 4884 ou o he o al 6280 polymo phic ma ke s we e subjec ed o analysis. To es ima e he LD decay, he in a-ch omosomal 2 alues we e plo ed agains he gene ic dis ance wi h a second-deg ee smoo hed loess cu e i ed using he p og am R (R De elopmen Co e Team 2012) and a baseline based on he c i ical alue o 2 was d awn. The c i ical alue o 2 , as an e idence o linkage, was calcula ed based on he me hod desc ibed in B e- seghello and So ells (2006), by squa e oo ans- o ming he 2 - alues and aking he 95 h pe cen ile o unlinked 2 - alues. In he analysis ma ke s loca ed mo e han 50 cM apa we e conside ed unlinked. Resul s Popula ion s uc u e in No dic sp ing ba ley The STRUCTURE analysis indica ed ha he ba ley panel could be di ided in o wo g oups wi h common ances y ([0.7) K1 SSR (n =109) and K2 SSR (n =50) and one g oup (n =10) wi h an admixed ances y ( 0.7), based on he SSR da a. AMOVA analysis e ealed ha he K1 SSR ,K2 SSR and admixed SSR g oup we e signi ican ly sepa a ed (p 0.001) and explained 35% o o al molecula a iance (Online Resou ce 1). This is simila o he esul ob ained om he PCoA analysis, whe e he i s wo p incipal coo dina es combined explained 32.1% (21.7 and 10.4%) o he a ia ion (Fig. 1a). The subdi ision o he popula ion along he i s p incipal coo dina e (PC1) co esponded o he sepa a ion o he ba ley popula ion in o K1 SSR and K2 SSR . The K1 SSR and K2 SSR g oup co esponded o wo- owed lines and six- owed lines, espec i ely, wi h he excep ion o six wo- owed lines ound in K2 SSR . In he small admixed SSR g oup, nine wo- owed lines and one six- owed line om mainly he no he n pa s we e ound. Also o he SNP da a, wo g oups wi h a common ances y ([0.7) (K1 SNP :n =109; K2 SNP :n =47) and one g oup (n =13) wi h an admixed ances y ( 0.7) we e in e ed by he STRUCTURE analysis. Th ee clus e s we e obse ed in he PCoA analysis, whe e he i s and second p incipal componen s explained 24.6 and 5.9% o he a ia ion, espec i ely, o 30.5% combined (Fig. 1b). AMOVA analysis e ealed ha he K1 SNP ,K2 SNP and admixed SNP g oups we e signi ican ly sepa a ed (p alue 0.001) and explained 42% o he o al molecula a iance (Online Resou ce 1). The wo- owed lines we e dis ibu ed be ween K1 SNP and he admixed SNP g oup and he six- owed lines we e ound in g oup K2 SNP , wi h he excep ion o wo wo- owed lines ha we e ound in he la e g oup. The lines wi h an admixed ances y in he admixed SNP g oup we e wo- owed and we e, jus as Gene Resou C op E ol 123 he lines in he admixed SSR g oup, mainly om he no he n pa s. Mo e de ailed in o ma ion ega ding he o igin o he lines and he in e ed popula ion s uc u e g oups can be ound in Online Resou ce 2. Gene ic di e si y in No dic sp ing ba ley Ma ke sys em compa ison The ba ley lines we e geno yped using 48 SSR ma ke s. The same lines we e also geno yped using he iSelec 9K SNP ba ley chip esul ing in a o al o 6208 polymo phic SNP ma ke s. The SSR ma ke s p oduced 234 sco able loci. The numbe o alleles pe ma ke anged om 1 o 15 wi h an a e age o 4.9 alleles pe SSR ma ke . The a e age polymo phic in o ma ion con en (PIC) alue was 0.46 o he SSR ma ke s and 0.28 o he SNP ma ke s. The gene ic di e si y index was highe o he SSR ma ke s (0.514) han o he SNP ma ke s (0.359). Man el es showed a s ong co ela ion be ween he SSR-based Modi ied Roge ’s gene ic dis ances and he SNP-based simple ma ching coe icien , wi h a Pea - son’s alue o 2 =0.86 (p alue 0.0000). Gene ic di e si y based on ea ow ype Allelic ichness pa ame e s in he o m o a e age numbe o alleles pe locus, numbe and p opo ion o p i a e alleles and gene ic di e si y o each ma ke ype based on ea ow ype a e p esen ed in Table 1. The e we e no majo di e ences in he a e age numbe o alleles pe locus ound in he wo- owed lines (SSR, 4.0; SNP: 1.9) compa ed o he six- owed lines (SSR: 3.5, SNP: 1.8).). The gene ic di e si y was highe in he wo- owed lines (SSR: 0.431; SNP: 0.305) compa ed o he six- owed lines (SSR: 0.386; SNP: 0.225). The wo- owed lines had also a highe numbe o p i a e alleles (SSR: 64; SNP: 1145) compa ed o he six- owed lines (SSR: 41; SNP: 447). AMOVA analysis e ealed ha 35 and 40% o he molecula a iance be ween he lines based on he SSRs and SNPs, espec i ely, could be explained by he ow- ypes (da a no shown). Gene ic di e si y based on popula ion s uc u e Allelic ichness pa ame e s in he o m o a e age numbe o alleles pe locus, numbe and p opo ion o p i a e alleles and gene ic di e si y o each ma ke ype based on he in e ed popula ion s uc u e g oups a e p esen ed in Table 2. Fo he SSRs he highes a e age numbe o alleles pe locus and gene di e si y was ound in he admixed SSR g oup (3.7; 0.404). Howe e , his g oup also had he lowes p opo ion o p i a e alleles (0.30), whe eas he highes p opo ion o p i a e alleles was ound wi hin he six- owed lines in g oup K2 SSR (0.86). No majo di e ences in he a e age numbe o alleles pe locus we e ound be ween he h ee subg oups, based on he SNP da a. The highes p opo ion o p i a e alleles (10.2) was, no di e en om he esul s ob ained wi h he SSRs, ound wi hin he six- owed lines in K2 SNP whe eas he lowes Fig. 1 a Associa ions be ween s uc u e g oups e ealed by p incipal coo dina e analysis o he No dic sp ing ba ley collec ion based on he SSR da a. bAssocia ions be ween s uc u e g oups e ealed by p incipal coo dina e analysis o he No dic sp ing ba ley collec ion based on he SNP da a Gene Resou C op E ol 123 p opo ion o p i a e alleles (0.54) we e ound in he admixed SNP g oup. The highes gene di e si y (0.279) was ound in K1 SNP , whe eas he lowes (0.198) was ound in he admixed SNP g oup. Ma ices showing ela ionships be ween he s uc- u e g oups we e gene a ed based on Nei’s unbiased gene ic dis ance o SSR and SNP da a (Table 3). In he SSR ma ix, he smalles dis ance (0.272) was ound be ween he K1 SSR and admixed SSR g oup (bo h mainly wo- owed lines om he sou he n and no h- e n pa s, espec i ely). The la ges dis ance (0.566) was ound be ween g oups K2 SSR (mainly six- owed lines) and admixed SSR (mainly wo- owed lines om he sou he n and no he n pa s, espec i ely). Simila esul s we e seen in he SNP ma ix, whe e he smalles dis ance (0.217) was ound be ween g oups K1 SNP and admixed SNP (mainly wo- owed lines om he sou he n and no he n pa s, espec i ely) and he la ges (0.338) be ween K2 SNP (mainly six- owed lines) and admixed SNP (mainly wo- owed lines om he no he n pa s). When compa ing gene ic di e si y along he ba ley ch omosomes, di e ences be ween he popula ion s uc u e g oups we e e iden in se e al genomic egions (Fig. 2). Bo h g oup K2 SNP and admixed SNP we e low in di e si y on ch omosome 2H (a ound 83–113 cM), 3H (a ound 37–64 cM) and 5H (a ound 108–127 cM), whe eas he K1 SNP lines we e e y di e se in hese egions. A la ge di e ence in gene ic di e si y we e also obse ed on ch omosome 6H (a ound 55–58 cM), whe e he K1 SNP and K2 SNP lines we e much mo e di e se compa ed o admixed SNP lines. Righ be o e his egion on ch omosome 6H (a ound 49 cM), a egion wi h low di e si y was seen o he K1 SNP lines, bu he e he di e si y was main ained o he wo o he g oups. On ch omosome 7H (a ound 68–88 cM) he lines in g oup K2 SNP we e e y di e se in con as o he low di e si y obse ed o he K1 SNP and admixed SNP lines in he same egion. Linkage disequilib ium in No dic sp ing ba ley LD o he whole popula ion, he popula ion s uc u e g oups and he ea ow- ypes we e calcula ed based on he SNP ma ke s. The pe cen age o unlinked ma ke pai s anged be ween 40 and 42% and no la ge di e ences we e ound be ween he di e en g oups. The numbe o in a-ch omosomal ma ke -pai s and he numbe o unlinked pai s in he o al popula ion and in he di e en g oups a e p esen ed in Table 4.A o al o 1,795,852 in a-ch omosomal ma ke pai s we e ound in he en i e popula ion. The mean 2 - alue o he en i e popula ion, he ea ow- ypes and popula ion s uc u e g oups we e calcula ed o he whole genome and o he se en ch omosomes sepa a ely (Table 5). The mean 2 - alue o he whole genome o he en i e popula ion was ound o be 0.10. The highes and lowes mean 2 - alue o he whole genome was ound in he admixed SNP and K1 SNP g oup ( 2 : 0.19; 0.07), espec i ely. The in e al in which he Loess cu e in e cep s he c i ical alue (backg ound LD) was conside ed as he LD decay. The backg ound LD and he LD decay o he en i e popula ion, he ea ow- ypes and popula ion s uc u e g oups we e calcula ed o he whole genome and o he se en ch omosomes sepa a ely (Table 6, Online Resou ce 3). A e age LD o he whole popula ion decayed below he c i ical le el ( 2 =0.20) wi hin a ange o 0–4.0 cM. Howe e he LD decay o he di e en ch omosomes, ea ow- ypes and popula ion s uc u e g oups anged om 0 o 12, wi h he mos ex ended LD decay ound o 5H in he six- owed lines and in g oup K1 SNP wi h an a e age 2 o 8–12 Table 1 Allele equency and gene ic di e si y in wo- and six- owed ba ley, based on SSR and SNP da a Popula ion size (n) A e age numbe o alleles pe locus Numbe o p i a e alleles P opo ion o p i a e alleles Gene ic di e si y (D) S anda d de ia ion (D) SSR SNP SSR SNP SSR SNP SSR SNP SSR SNP Two- ow 124 4 1.9 64 1145 0.5 9.2 0.431 0.305 0.063 0.100 Six- ow 45 3.5 1.8 41 447 0.9 9.9 0.386 0.225 0.123 0.233 To al 169 0.514 0.359 0.434 0.448 Gene Resou C op E ol 123 (Table 6). When compa ing he whole genome, he six- owed lines, g oup K2 SNP and he lines om he no he n pa s in g oup admixed SNP had a mo e apid LD decay wi h an a e age 2 be ween 0 and 4 compa ed o he wo- owed lines and g oup K1 SNP , wi h lines om he sou he n pa s, whe e he a e age LD decay we e mo e slow wi h an a e age 2 be ween 4–8 and 8–12, espec i ely. Discussion Compa ison o ma ke sys ems I is o in e es o compa e i he in o ma ion ega ding popula ion s uc u e and gene ic di e si y is a ec ed by he ma ke sys em o choice, since he e a e di e en mu a ional mechanisms behind SSR ( epli- ca ion slippage) and SNP (poin mu a ion) ma ke s. SSRs ha e been he mos commonly used o s udies o gene ic di e si y, mainly due o hei abundance in he genome, ep oducibili y and high le el o poly- mo phism. Howe e , he inc eased a ailabili y o he SNP ma ke s and he as and highly au oma ed geno yping echnologies, ha e ecen ly mo ed he a en ion o he use o he SNPs in s udies o gene ic di e si y and popula ion s uc u e. This s udy showed ha he a e age PIC and gene ic di e si y alues we e highe o he SSRs compa ed o he SNPs. Howe e , hese alues ha e a maximum alue o 0.5 o bi- allelic ma ke s such as SNPs when he ma ke s sco es a e 50% (0) and 50% (1). Taking his in o Table 2 P i a e alleles and gene ic di e si y in he s uc u e g oups based on SSR and SNP da a G oup SSR da a G oup SNP da a Popula ion size A e age numbe o alleles pe locus No. o p i a e alleles P opo ion o p i a e alleles Gene ic Di e si y (D) S anda d de ia ion (D) Popula ion size A e age numbe o alleles pe locus No o p i a e alleles P opo ion o p i a e alleles Gene ic Di e si y (D) S anda d de ia ion (D) K1 SSR 109 2.5 42 0.39 0.392 0.275 K1 SNP 109 1.9 639 5.86 0.279 0.153 K2 SSR 50 3.5 43 0.86 0.397 0.063 K2 SNP 47 1.8 481 10.23 0.235 0.208 Admixed SSR 10 3.7 3 0.30 0.404 0.120 Admixed SNP 13 1.6 7 0.54 0.198 0.495 To al 169 0.514 0.434 To al 169 0.359 0.448 Table 3 Nei’s unbiased gene ic dis ance be ween di e en s uc u e g oups, based on he (a) SSR da a, (b) SNP da a K1 SSR K2 SSR Admixed SSR (a) K1 SSR 0.000 K2 SSR 0.292 0.000 Admixed SSR 0.272 0.566 0.000 K1 SNP K2 SNP Admixed SNP (b) K1 SNP 0.000 K2 SNP 0.250 0.000 Admixed SNP 0.217 0.338 0.000 Gene Resou C op E ol 123 conside a ion, he SNPs would be jus as o e en mo e in o ma i e han he SSRs. A highe numbe o p i a e alleles we e ound wi h he SNPs compa ed o he SSRs (Tables 1,2). Howe e , conside ing he di e - en numbe o ma ke s, he SSRs ac ually had he highes p i a e allele equency, which could be expec ed since he SNPs a e bi-allelic and ha e a lowe mu a ion a e compa ed o he SSRs (Ma inez- A ias e al. 2001; Li e al. 1981; K uglyak e al. 1998). I has ea lie been epo ed ha he e is s ong co ela ion be ween hese wo ma ke sys ems in ba ley (Va shney e al. 2008; Va shney e al. 2010). Fig. 2 Shannon’s di e si y index calcula ed as olling means o e 20 adjacen loci. The s a and end posi ion o each ch omosome a e indica ed wi h e ical lines a he bo om o he igu e Table 4 Numbe o in a- ch omosomal ma ke -pai s in he o al popula ion and in he di e en g oups Unlinked ma ke -pai s e e s o a ma ke -pai dis ance [50 cM Popula ion size To al pai s Unlinked pai s Unlinked pai s (%) To al popula ion 169 1,795,852 744,332 41 Two- ow lines 124 1,464,051 609,353 42 Six- ow lines 45 923,896 367,226 40 K1 SNP 109 1,166,044 482,154 41 K2 SNP 47 981,129 393,735 40 Admixed SNP 13 663,021 262,178 40 Table 5 Mean 2 - alues o in a-ch omosomal ma ke -pai s in he whole genome and o each ch omosome Ch omosome No. o SNPs To al popula ion Two- owed lines (n =124) Six- owed lines (n =45) K1 SNP (n =109) K2 SNP (n =47) Admixed SNP (n =13) 1H 437 0.12 0.05 0.11 0.05 0.11 0.15 2H 768 0.11 0.07 0.12 0.06 0.11 0.23 3H 724 0.07 0.08 0.10 0.08 0.10 0.17 4H 578 0.09 0.07 0.11 0.06 0.10 0.18 5H 1025 0.10 0.10 0.10 0.05 0.10 0.21 6H 726 0.09 0.09 0.14 0.10 0.13 0.17 7H 626 0.09 0.07 0.08 0.07 0.09 0.19 Whole genome 4884 0.10 0.08 0.11 0.07 0.10 0.19 Gene Resou C op E ol 123 This was also demons a ed he e wi h a s ong and posi i e co ela ion ( =0.86) be ween he modi ied Roge ’s dis ances based on he SSR da a and he simple ma ching coe icien based on he SNP da a. Also he PCoA and s uc u e analysis g ouped he lines in a simila way wi h he wo ma ke sys ems. Howe e , he PCoA clus e ing based on he SNP da a showed clea e and mo e dis inc g oupings han he SSR-based PCoA, bu no majo di e ence in he amoun o molecula a iance explained was ound be ween he wo ma ke sys ems. Nei he did he AMOVA analyses o he s uc u e g oups e eal any la ge di e ences be ween he wo ma ke sys ems (Table S1). Howe e , supe imposing he esul s om he popula ion s uc u e analysis on he esul s om he PCoA p o ided a clea e image and highe esolu ion o he popula ion s uc u e based on he SNPs compa ed o he SSRs (Fig. 1a, b). Tha e eals he abili y o SNPs o explain he popula ion s uc u e a a mo e speci ic le el. This was also seen in a s udy o gene ic di e si y and popula ion s uc u e o 375 ice a ie ies (Singh e al. 2013), whe e a compa ison o he SSR and SNP ma ke sys ems e ealed ha a he s uc u e le el he SNPs we e be e a desc ibing gene ic ela edness whe eas a he di e si y le el he SSRs showed a be e g ouping o samples. In con as , some s udies o gene ic di e si y and popula ion s uc u e in maize epo a be e es ima e o popula- ion s uc u e wi h SSRs compa ed o SNPs (Yang e al. 2011; Hamblin e al. 2007). The di e en esul s be ween hose epo s and his s udy migh be due o he di e en numbe s o SNP ma ke s used ( 900 s. 6208) o due o he complexi y o he maize genome. Acco ding o a heo e ical p edic ion by La al e al. (2002), (k-1) imes mo e bi-allelic ma ke s a e needed o achie e a compa able accu acy o he gene ic dis ance as a se o SSRs wi h kalleles. Wi h he a e age o abou 3 alleles pe SSR ma ke in his s udy, he numbe o SNPs needed would be [(3 -1) 948] =96, which a e abou 65 imes less compa ed o he 6208 used. Al oge he he esul s om his s udy show ha he numbe s o SNPs used a e mo e han enough o e ie e a compa able accu acy o gene ic di e si y in ba ley as he se o SSRs used. In addi ion, he SNPs seem o p o ide a highe esolu ion o he gene ic ela edness han ob ained wi h he SSRs. Di e si y and ela ionships wi hin No dic sp ing ba ley The a e age gene ic di e si y o he No dic sp ing ba ley collec ion analysed he e was 0.514 and 0.359 based on he SSRs and SNPs, espec i ely. These a e simila o he di e si y es ima es o No dic b eeding lines and cul i a s eleased a e 1970 (0.601) epo ed in an ea lie s udy based on SSRs (Kolodinska B an es am e al. 2007). In addi ion a simila esul based on SSRs was epo ed in accessions om Eu ope (0.593), E i ea (0.573) and E hiopia (0.620), whe eas he Ho deum ulga e subsp. spon aneum (K. Koch) and H. ulga e accessions om he Wes Asia No h A ica (WANA) egion had a highe di e si y (0.826 and 0.762, espec i ely) (O abi e al. 2007). Table 6 In e al o he es ima ed LD decay (cM) in he o al popula ion and o he di e en g oups Ch omosome No. o SNPs To al popula ion Two- owed lines (n =124) Six- owed lines (n =45) K1 SNP (n =109) K2 SNP (n =47) Admixed SNP (n =13) 1H 437 0–3 6–9 3–6 6–9 3–6 3–6 2H 768 0–3 3–6 3–6 6–9 3–6 0–3 3H 724 7–11 7–11 4–7 7–11 4–7 4–7 4H 578 0–3 3–5 0–3 3–5 0–3 0–3 5H 1025 0–4 0–4 0–4 8–12 4–8 0–4 6H 726 0–3 5–8 0–3 5–8 0–3 0–3 7H 626 0–3 7–10 7–10 7–10 3–7 0–3 Whole genome 4884 0–4 4–8 0–4 8–12 0–4 0–4 Backg ound LD whole genome a 0.20 0.16 0.19 0.10 0.19 0.31 a The 95 h pe cen ile o unlinked (abo e 50 cM) squa e oo ans o med 2 alues Gene Resou C op E ol 123