Admixture and gene flow from Russia in the recovering Northern European brown bear (Ursus arctos)

Admix u e and Gene Flow om Russia in he Reco e ing No he n Eu opean B own Bea ( U sus a c os ) Alexande Kopa z 1 *, Hans Gei Eiken 1 , Jouni Aspi 2 , Ilpo Kojola 3 , Camilla Tobiassen 1 , Kons an in F. Ti onen 4 , Pjo I. Danilo 4 , Sno e B. Hagen 1 * 1Bio o sk - No wegian Ins i u e o Ag icul u al and En i onmen al Resea ch, S an ik, No way, 2Depa men o Biology, Uni e si y o Oulu, Oulu, Finland, 3Finnish Game and Fishe ies Resea ch Ins i u e, Oulu, Finland, 4Ins i u e o Biology, Ka elian Resea ch Cen e o he Russian Academy o Science, Pe oza odsk, Russia Abs ac La ge ca ni o es we e pe secu ed o nea ex inc ion du ing he las cen u ies, bu ha e now eco e ed in some coun ies. I has been p oposed ea lie ha he eco e y o he No he n Eu opean b own bea is suppo ed by mig a ion om Russia. We es ed his hypo hesis by ob aining o he i s ime con inuous sampling o he whole Finnish bea popula ion, which is loca ed cen ally be ween he Russian and Scandina ian bea popula ions. The Finnish popula ion is assumed o expe ience high gene low om Russian Ka elia. I so, no o a low deg ee o gene ic di e en ia ion be ween Finnish and Russian bea s could be expec ed. We ha e geno yped bea s ex ensi ely om all o e Finland using 12 alida ed mic osa elli e ma ke s and compa ed hei gene ic composi ion o bea s om Russian Ka elia, Sweden, and No way. Ou ine masked in es iga ion iden i ied wo o e lapping gene ic clus e s s uc u ed by isola ion-by-dis ance in Finland (pai wise F ST = 0.025). One clus e included Russian bea s, and mig a ion analyses showed a high numbe o mig an s om Russia in o Finland, p o iding e idence o eas e n gene low as an impo an d i e du ing eco e y. In compa ison, bo h clus e s excluded bea s om Sweden and No way, and we ound no mig an s om Finland in ei he coun y, indica ing ha eas e n gene low was p obably no impo an o he popula ion eco e y in Scandina ia. Ou analyses on di e en spa ial scales sugges a con inuous bea popula ion in Finland and Russian Ka elia, sepa a ed om Scandina ia. Ci a ion: Kopa z A, Eiken HG, Aspi J, Kojola I, Tobiassen C, e al. (2014) Admix u e and Gene Flow om Russia in he Reco e ing No he n Eu opean B own Bea (U sus a c os). PLoS ONE 9(5): e97558. doi:10.1371/jou nal.pone.0097558 Edi o : Paul Hohenlohe, Uni e si y o Idaho, Uni ed S a es o Ame ica Recei ed Decembe 19, 2013; Accep ed Ap il 21, 2014; Published May 19, 2014 Copy igh : ß2014 Kopa z e al. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed. Funding: The s udy was unded by Bio o sk. The unde had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . Compe ing In e es s: The au ho s ha e decla ed ha no compe ing in e es s exis . * E-mail: alexande .kopa[email p o ec ed] (AK); sno e.hagen@bio o sk.no (SBH) In oduc ion Habi a agmen a ion and an h opogenic dis u bance is a global h ea o wildli e, wi h impac s such as declining popula ion sizes and educed gene low among popula ions. Bo h e ec s a e widely epo ed o p omo e gene ic d i and oppose long- e m popula ion iabili y [1]. Ce ain species a e pa icula ly sensi i e o loss o in e -popula ion connec i i y [2–4]. Examples a e many apex p eda o s, cha ac e ized by small popula ion size, long gene a ion imes, la ge home anges, and high le els o human pe secu ion [2,5–8]. Du ing he las cen u y, la ge e es ial ca ni o es declined bo h in numbe s and geog aphic dis ibu ion (see e.g. [4,9]). E en hough hey we e almos ex i pa ed in mos o Eu ope, la ge ca ni o es ha e now eco e ed in some a eas and popula ions a e expanding [2,4,9,10]. An impo an s ep owa ds unde s anding he unde lying causes o eco e y is o de e mine he cu en deg ee o gene low and gene ic di e en ia ion among la ge ca ni o e popula ions ac oss na ional bo de s [10]. In Finland, he b own bea (U sus a c os) was dis ibu ed h oughou he coun y un il he beginning o he 19 h cen u y [11]. A he end o he 19 h cen u y, bea s seemed o be ex inc om cen al, sou he n, and wes e n Finland, while obse a ions o bea s we e s ill epo ed in he no h and eas [12]. His o ic eco ds indica e ha he b own bea popula ion o Finland wen h ough a demog aphic bo leneck, wi h a leas 9,000 indi iduals killed be ween 1875 and 2000 [12]. I is assumed ha he popula ion size eached i s minimum be ween 1920 and 1950. In 1963, he emaining numbe o bea s was es ima ed o be abou 150 indi iduals [13]. Es ima es based on bea obse a ions sugges ed an inc ease om app oxima ely 300 o 800 indi iduals be ween 1978 and 2003 [14,15]. Mig a ion om Russia in o Finland has been assumed o ha e suppo ed he g ow h o he Finnish popula ion [11,13,16–18]. The mos ecen es ima es based on obse a ions o he numbe o li e s-o - he-yea a e sugges i e o a numbe be ween 1,150 and 1,950 bea s in 2009, wi h highes densi ies in he sou h along he Finnish-Russian bo de [19]. In his a ea, eco ds o killed bea s also indica e a pa icula ly high p opo ion o emale bea s [16,20,21]. The Finnish b own bea popula ion is loca ed cen ally be ween he popula ions o Russia and Scandina ia. The Republic o Ka elia and he Mu mansk Oblas in Russia a e he neighbo ing dis ic s owa ds Finland. Based on hun ing eco ds, obse a ions, and ack coun ing he es ima ed numbe s o bea s in hese dis ic s in 1990 we e abou 3,500 and 500 bea s, espec i ely [22,23]. Towa ds he no h, Finland sha es he bo de wi h No way, whe e nonin asi e gene ic sampling o sca s and hai s has documen ed small b own bea popula ions in he Pas ik Valley, in he Ka asjok-Ana johka egion, and in he a ea o Di idalen in T oms [24,25]. In Sweden, owa ds he wes , e o -co ec ed PLOS ONE | www.plosone.o g 1 May 2014 | Volume 9 | Issue 5 | e97558 moose-hun e obse a ions combined wi h nonin asi e gene ic cap u e- ecap u e s udies ha e been used o es ima e he popula ion o be app oxima ely 3,300 bea s [26]. Recen s udies o b own bea s om No he n Eu ope sugges bo h gene ic s uc u ing due o isola ion-by-dis ance (IBD) and he exis ence o sepa a e gene ic popula ions [24,25,27–31]. P e ious- ly, we ha e de ec ed bi-di ec ional mig a ion a es o abou 30% be ween bea s in Eas e n Finland and bea s u he eas in A khangelsk, Russia [28]. Ano he s udy applying au osomal mic osa elli es o a es ic ed numbe o samples sugges ed ha Finnish bea s a e di ided in o a no he n and a sou he n subpopula ion [30]. A ecen mi ochond ial genome s udy o bea s in No heas e n Eu ope also indica ed a no he n and a sou he n clus e in luenced by mi ogene ic haplog oups om Eu opean Russia [32]. Fu he mo e, we ha e ound indica ions ha he connec i i y be ween he bea popula ions in Eas e n Finland and Scandina ia o be limi ed [25]. Since he eco e y o b own bea s in Finland is assumed o be explained by high gene low om Russian Ka elia, one should expec a low deg ee o gene ic di e en ia ion be ween b own bea s om hese a eas oday, which has no been su icien ly es ed. In addi ion, he esul s o he p e ious s udies sugges ing mo e han one subpopula ion o bea s in Finland [30], may be inaccu a e because o IBD and selec i e sampling. In con as , in his s udy we ha e sampled indi iduals ex ensi ely and con inuously all o e Finland o answe he ques ion whe he o no he e is a no he n and sou he n popula ion o bea s in Finland. We included samples om Russia o sc u inize he in luence o eas e n gene low on he composi ion o he Finnish bea popula ion. In a las s ep, we included ou p e iously published gene ic da a on bea s om Scandina ia (Sweden and No way) o in es iga e he connec i i y u he wes wa ds. Compa ing esul s on he gene ic s uc u e om h ee di e en geog aphic scales allowed us o de e mine mo e p ecisely he unde lying gene ic admix u e and gene low in No hwes e n Eu ope. Ma e ials and Me hods Sampling All samples we e collec ed om dead animals, ha es ed legally in Finland and Russia. Legal ha es o bea s in Finland in he di e en hun ing dis ic s ollows an annual quo a co esponding o he es ima ed abundance and dis ibu ion o b own bea s in hose a eas [19], and he sampling in his s udy ollows his dis ibu ion h oughou Finland. Tissue samples we e ob ained by ou collabo a o s namely he Finnish Game and Fishe ies Resea ch Ins i u e and he Ka elian Resea ch Cen e o he Russian Academy o Science. No e hic pe mi was equi ed, as he sample collec ion did no in ol e li e animals. In ou s udy, we analyzed he da a o a o al o 517 bea s om 2006 o 2010 (Figu e 1, Table S1), including 286 issue samples om indi iduals om Finland, collec ed annually om legally ha es ed bea s (91 emales, 195 males). To in es iga e gene low om Russia in o Finland and wes wa ds o Scandina ia, we included p e iously geno yped indi iduals om No way (N= 97), Sweden (N= 84) and Russia (N= 22); (see Kopa z e al. 2012 and Figu e 1. Sampling loca ions o he b own bea s ( N = 517) in No he n Eu ope. Samples we e collec ed om 2006 o 2010 and each indi idual is ep esen ed by a ed do . doi:10.1371/jou nal.pone.0097558.g001 Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 2 May 2014 | Volume 9 | Issue 5 | e97558 Sch egel e al. 2012) and 28 addi ional issue samples om Russian Ka elia ( o al N= 50 bea s) om he same ime pe iod. Molecula Analysis Immedia ely a e collec ion, issue samples we e s o ed in 95% e hanol un il ex ac ion. Samples we e ex ac ed wi h DNeasy Tissue Ki (Qiagen), ollowing he manu ac u e ’s ins uc ions, and geno yped using 12 di e en dinucleo ide ma ke s (sho - andem- epea s, STRs) de eloped o bea s: G1A, G1D, G10B, G10L [33,34]; Mu05, Mu09, Mu10, Mu15, Mu23, Mu50, Mu51 and Mu59 [35]. We ha e p e iously alida ed hese STRs o hei species sensi i i y, p ecision and p obabili y o iden i y [24,27]. The p o ocol o PCR and agmen analysis can be ound in And eassen e al. [27]. Ou labo a o y p ocedu es ollow he guidelines o he analysis o non-human o ensic DNA ma e ial [36]. We e i ied he uniqueness o all geno ypes by calcula ing hei p obabili y o iden i y using he so wa e Gimle e sion 1.3.3 [37]. Geno ypes we e es ed in Mic o-Checke e sion 2.2.3 o possible allelic d opou , p esence o null alleles, and sco ing e o s caused by s u e peaks [38]. Popula ion S uc u e We es ed o gene ic s uc u e using wo Bayesian assignmen algo i hms (S uc u e and Geneland) and ac o ial co espondence analysis (FCA). Since ea lie s udies ha e indica ed a es ic ed numbe o gene ic clus e s in No hwes e n Eu ope [25,28–30], we se ou Bayesian analyses on gene ic clus e ing o a maximum o K= 10. In S uc u e e sion 2.3.3 [39,40], we assumed popula ion admix u e and co ela ed allele equencies wi hin he popula ion. Ten independen uns o each K alue be ween one and en we e pe o med. Fo each un, we se a bu n-in pe iod o 100,000 Ma ko Chain Mon e Ca lo (MCMC) i e a ions, ollowed by sampling o 1,000,000 i e a ions. The esul s we e pos -p ocessed wi h he ad-hoc app oach o E anno e al. [41] o es ima e he numbe o gene ic clus e s using S uc u e Ha es e [42]. A membe ship coe icien (q) abo e 0.6 has been conside ed as a easible cu -o membe ship alue o assign indi iduals o a popula ion wi h con idence, since mo e han 50% o he genome is assigned o a g oup and he e o e sugges s in e ed ances y [43,44]. P e ious s udies on bea s ha e used a membe ship coe icien (q) o 0.7 [30,44]. Thus, we ha e applied a h eshold alue o q.0.7 in his s udy. In Geneland [45], we an i e independen uns, whe e he pa ame e s o possible popula ions we e K= 1 o 10, and he Figu e 2. Bayesian clus e ing esul s o Finnish and Russian Ka elian bea s wi h S uc u e (P i cha d e al. 2000). Ba plo s show he assignmen p obabili ies o each bea o one o he iden i ied wo clus e s when only samples om Finland we e analyzed (a) and samples om Finland and Russian Ka elia pooled oge he (d); no he n clus e (g een), sou he n clus e (blue). The y-axis shows he calcula ed membe ship coe icien (q). Indi iduals a e a anged by la i ude om no h (le ) o sou h ( igh ). (b and e) The maps show he geno ypes in acco dance o hei assignmen in S uc u e and geog aphical loca ion. Indi iduals which we e no assigned unambiguously (membe ship coe icien q,0.7) a e shown on a sepa a e map as da k ed do s. (c and ) Maps on he bo om show he assignmen wi h he p og am Geneland (Guillo e al. 2005). doi:10.1371/jou nal.pone.0097558.g002 Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 3 May 2014 | Volume 9 | Issue 5 | e97558 numbe o MCMC i e a ions was 1,000,000, wi h a hinning o 100. The maximum a e o Poisson p ocess was se o 100, and he maximum numbe o nuclei was 300. Geog aphical loca ion o he samples (longi ude, la i ude) was included in o he analysis. FCA was pe o med wi h he p og am Gene ix 4.05.2 [46]. To de e mine he deg ee o di e en ia ion among gene ic clus e s, AMOVA analyses and pai wise F ST alues we e calcula ed wi h he p og am A lequin e sion 3.5.1.2 [47]. Isola ion-by-dis ance We calcula ed IBD among pai s o b own bea s in Finland and Russian Ka elia using he so wa e Spagedi e sion 1.3 [48] wi h he kinship coe icien by Loiselle e al. [49]. Gene Flow To u he es he eas -wes gene low hypo hesis we es ima ed he amoun o mig a ion be ween he bea s in Finland and Russian Ka elia as well as be ween Finland and Scandina ia using wo di e en me hods. Fi s ly, he e ec i e numbe o mig an s (Nm) was es ima ed using he p i a e allele me hod [50] implemen ed in he p og am Genepop [51]. Secondly, o iden i y possible ecen mig an s, we es ima ed he likelihood o a bea o belong o he popula ion i was sampled using he indi idual Bayesian assignmen me hod in he p og am Geneclass 2 [52]. We used he algo i hm by Rannala and Moun ain [53] and esampling as desc ibed in Pae kau e al. [54] o iden i y i s gene a ion mig an s. The simula ion was se o 10,000 indi iduals and he ype I e o (alpha) o 0.05. Gene ic Di e si y We calcula ed numbe o alleles, expec ed and obse ed he e ozygosi y wi h he p og am A lequin e sion 3.5.1.2 [47]. Inb eeding coe icien s and es s o linkage disequilib ium be ween pai s o loci we e pe o med wi h he p og am Gene ix 4.05.2 [46] using he me hod by Black and K a su [55]. De ia ions om Ha dy-Weinbe g equilib ium (HWE) we e es ed wi h Fishe ’s me hod [56] o all loci and popula ions wi h he p og am Genepop e sion 4.0 [51], wi h unbiased P alues by a Ma ko chain me hod o 1000 bu n-in i e a ions, 500 ba ches and 1000 i e a ions pe ba ch. Popula ion Bo lenecks Fo he bea s om Finland and Russia, we es ed o la ge obse ed he e ozygosi y han expec ed o de ec possible gene ic bo lenecks in he ecen his o y o he bea popula ions wi h he p og am Bo leneck 1.2.02 [57]. We applied he wo-phase mu a ion model using 95% single s ep mu a ions o es ima e he expec ed he e ozygosi ies (20,000 i e a ions) and es ed he signi icance o he di e ences be ween obse ed and expec ed he e ozygosi ies using he Wilcoxon es . Fu he , we applied he M a io es o in es iga e i he e a e signs o gene ic bo lenecks u he in he pas (.100 gene a ions) and he e o e we calcula ed he modi ied Ga za-Williamson indices [58] o he clus e s ound implemen ed in he p og am A lequin e sion 3.5.1.2 [47]. Resul s Popula ion S uc u e in Finland and Russian Ka elia The S uc u e clus e ing app oach sugges ed wo gene ic clus e s in Finland wi h a high deg ee o admix u e and geog aphical o e lap (Figu e 2a and b, Figu e S1a, and Figu e S2a). While one clus e was sp ead almos h oughou wo- hi ds o he coun y, he o he one was es ic ed o he sou he n pa o Finland (Figu e 2b). A o al o 60 bea s (21%) we e no assigned unambiguously o he iden i ied clus e s (membe ship coe icien q,0.7), and hose we e mainly ound in he zone whe e he clus e s o e lapped (Figu e 2b). All al e na i e models o popula ion s uc u e using a la ge numbe o clus e s (K=3 o 5) had lowe likelihoods and showed subs an ially highe numbe s o unassigned indi iduals up o 72% (Figu e S2a). Geneland iden i ied also wo clus e s, i.e. a no he n and a sou he n one (Figu e 2c). FCA suppo ed he esul s by S uc u e, showing wo, o e lapping g oups o bea s (Figu e 3a). Bea s no unambiguously Figu e 3. Fac o ial co espondence plo s o b own bea s sampled 2006–2010 in Finland, Russia, No way and Sweden. Di e en colo s ep esen he clus e s iden i ied by he Bayesian clus e ing app oach. (a) FCA analysis o he Finnish samples only: no he n clus e (g een), no he n clus e wi h a membe ship coe icien (q),0.7 (ligh g een), sou he n clus e (blue), sou he n clus e wi h assignmen membe ship p obabili y ,0.7 (cyan). (b) FCA analysis o Finnish and Russian Ka elian samples: no he n clus e (g een), no he n clus e wi h an assignmen membe ship p obabili y ,0.7 (ligh g een), sou he n clus e (blue), sou he n clus e wi h assignmen membe ship p obabili y ,0.7 (cyan). (c) FCA analysis o Finnish, Russian Ka elian popula ions and bea s sampled in No he n No way and Sweden: no he n clus e (g een), no he n clus e wi h a membe ship coe icien ,0.7 (ligh g een), sou he n clus e (blue), sou he n clus e wi h assignmen membe ship p obabili y ,0.7 (cyan), wes e n (Scandina ian) clus e (o ange), wes e n clus e wi h an assignmen membe ship p obabili y ,0.7 (yellow). doi:10.1371/jou nal.pone.0097558.g003 Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 4 May 2014 | Volume 9 | Issue 5 | e97558 assigned by S uc u e showed highes simila i y and occu ed on he FCA plo be ween he wo clus e s (Figu e 3a). Simila o he esul s in Finland alone, assigning geno ypes om Finland and Russian Ka elia oge he also sugges ed wo gene ic clus e s (Figu e 2d, Figu e S1b and Figu e S2b). One clus e sp ead h oughou he dis ibu ion ange while he o he one was concen a ed mainly o he sou he n pa o he s udy egion (Figu e 2e). Admix u e and geog aphical o e lap as sugges ed by he assignmen p obabili ies (Figu e 2d) could be as well obse ed he e (Figu e 2e). Simila ly, mos o he unambiguously assigned geno ypes, 79 indi iduals in o al (22.8%), we e ound in he geog aphic o e lap zone (Figu e 2e). Geneland showed wo gene ic clus e s: a no he n and sou he n clus e , wi h a dis inc i e bo de in he middle o Finland (Figu e 2 ). FCA analyses suppo ed he esul s by S uc u e o wo, o e lapping gene ic g oups (Figu e 3b). Repea edly, bea s which we e no assigned unambiguously seem o highligh an admix u e g oup be ween he wo iden i ied clus e s (Figu e 3b). Connec i i y wi h Scandina ia A e pooling Finnish and Russian Ka elian bea s oge he wi h bea s om he no he n ans-bo de a ea o Pas ik in No way and Russia, T oms in No way, and Va¨s e bo en in Sweden, he Bayesian clus e ing app oaches (S uc u e and Geneland) sugges - ed h ee gene ic clus e s: a wes e n one, including mainly geno ypes om Scandina ia, namely Va¨s e bo en and T oms, a no he n one, including geno ypes om Pas ik and no he n Finland and Mu mansk, and a sou he n clus e con aining geno ypes om middle and sou he n Finland as well as Russian Ka elia (Figu e 4a and b, Figu e S1c and Figu e S2c). In compa ison o he analyses o popula ion s uc u e using samples om Finland and Russian Ka elia only, he bo de be ween he wo g oups in Finland and Ka elia was loca ed a bi u he no h (Figu es 2 and 4). He e, we ound app oxima ely 20 indi iduals in he no h assigned o he sou he n clus e , compa ed o he esul s when using solely Finnish bea samples, whe e only one indi idual in he no h has been assigned o he sou he n clus e . Fu he mo e, only 18 (3.5%) o he indi iduals could no be assigned unambiguously, sugges ing ha mos o he unassigned geno ypes ound a smalle spa ial scales we e in e media e geno ypes om he admix u e zone a he han indi iduals om an unknown popula ion. Geno ype assignmen wi h S uc u e in acco dance o he sampling loca ion is shown in Figu e 4b. FCA Figu e 4. Assignmen o bea s sampled 2006–2010 in No hwes e n Eu ope wi h he p og am S uc u e (P i cha d e al. 2000). Ba plo s show he assignmen p obabili ies o each bea o one o he h ee iden i ied clus e s (a). Geno ypes a e so ed ‘‘clockwise’’ in acco dance o hei loca ion om sou h-wes in Sweden o sou h eas in Finland and Russia acco ding o om le o igh : in o ange Va ¨s e bo en (sou h-no h) and T oms (wes -eas ), in g een No he n Finland (wes -eas ) and Pas ik (no h-sou h), in blue Sou he n Finland and Russian Ka elia (no h-sou h). They- axis shows he calcula ed membe ship coe icien (q). (b) The maps show he geno ypes in acco dance o hei assignmen in S uc u e and geog aphical loca ion: wes e n clus e in Scandina ia, namely Va ¨s e bo en and T oms, shown in o ange; no he n clus e in g een and sou he n clus e in blue. (c) Maps on he bo om show he assignmen wi h he p og am Geneland (Guillo e al. 2005). doi:10.1371/jou nal.pone.0097558.g004 Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 5 May 2014 | Volume 9 | Issue 5 | e97558 analysis was in line wi h he esul s by he Bayesian assignmen s and isualized h ee dis inc i e g oups o bea s (Figu e 3c). Popula ion Di e en ia ion in Finland and Russian Ka elia The pai wise F ST alues be ween he wo subpopula ions we e signi ican (P,0.001), wi h F ST = 0.025 be ween he no he n and sou he n clus e ound in Finland and F ST = 0.026 be ween he wo clus e s iden i ied when Finnish and Russian samples we e analyzed oge he . AMOVA analysis wi hin Finland showed ha 2.54% o he a ia ion was be ween he clus e s and 97.46% wi hin hem (P,0.01). Fo he no h-sou h di ision in Finland and Russian Ka elia, 2.59% a ia ion was be ween he clus e s and 97.41% a ia ion wi hin he hem, espec i ely (P,0.01). Isola ion-by-dis ance We de ec ed a signi ican , nega i e ela ionship (P,0.001) be ween kinship and spa ial dis ance be ween pai s o indi iduals sampled con inuously in Finland and Russian Ka elia, p o iding e idence o an in luence o IBD on he deg ee o gene ic s uc u ing. All dis ance classes showed signi ican de ia ion o kinship om he popula ion mean (Figu e 5). Gene Flow The es ima ed e ec i e numbe o mig an s was much highe be ween Finland and Russian Ka elia han be ween Finland and Scandina ia (Figu e 6, Table S2). Simila esul s we e ound using he so wa e Geneclass 2, which de ec ed 18 mig an s be ween Finland and Russian Ka elia. Ou o hese, 15 bea s we e iden i ied as i s gene a ion mig an s om Russia in o Finland. In compa ison only h ee indi iduals we e iden i ied as mig an s om Finland in o Russia. Be ween Scandina ia and Finland 8 indi iduals we e de ec ed as mig an s. All o hem we e sampled in Finland and o igina ed om Scandina ia. Gene ic Di e si y Mean obse ed and expec ed he e ozygosi y in Finland we e highe in he no he n clus e han in he sou he n one. The no he n clus e also showed highe numbe o alleles pe locus han he sou he n one (Table 1). One locus (G10B) showed de ia ion om HWE wi hin he sou he n popula ion, due o excess o he e ozygo es. This locus as well as locus Mu05 showed signi ican , nega i e alues o F IS (Table 1). When all geno ypes om Finland we e pooled, i esul ed in he whole popula ion de ia ing om HWE. One locus (Mu09) de ia ed as well and showed an ele a ed, albei low, posi i e alue o F IS due o excess o homozygo es (Table 1). This o e all de ia ion om HWE may be mos p obably caused by he Wahlund e ec , by pooling samples om wo di e en gene ic clus e s in o one. F IS a locus Mu23 was ele a ed and signi ican (Table 1). We ound signi ican linkage disequilib ium (P,0.01) a e sequen ial Bon e oni co ec ion in 40 ou o 66 ma ke pai s. No able is ha ou o hese, 29 pai s we e solely ound in he Scandina ian clus e (Va¨s e bo en and T oms). Signi ican LD ound was no consis en ac oss all samples and all gene ic clus e s iden i ied. Popula ion Bo lenecks We de ec ed a gene ic bo leneck (P= 0.034) o he gene ic clus e iden i ied in Scandina ia (Va¨s e bo en and T oms). This clus e showed also he lowes alue o he Ga za-Williamson index wi h M= 0.64, which is jus below M c i o M,0.68 p oposed by Ga za and Williamson [58] o sugges he occu ence o a gene ic bo leneck in he pas . Discussion We ha e es ed he hypo hesis ha he bea popula ion o Russian Ka elia has ac ed as a sou ce popula ion du ing he eco e y o he Finnish and Scandina ian bea popula ions. We Figu e 5. Co ela ion be ween geog aphical dis ance and kinship o he b own bea s in Finland and Russian Ka elia. Samples we e collec ed om 2006 o 2010 (N= 346) and IBD was analyzed wi h he p og am Spagedi 1.3 (Ha dy and Vekemans 2002). All nine dis ance classes di e signi ican ly (P,0.001) om he mean kinship o he popula ion. doi:10.1371/jou nal.pone.0097558.g005 Figu e 6. The es ima ed numbe o e ec i e mig an s pe gene a ion ( Nm ). Es ima ed mig an s be ween Finland and Russian Ka elia (black squa es) as well as be ween Finland and Scandina ia (g ey iangles) plo ed agains sample size (see Table S2). doi:10.1371/jou nal.pone.0097558.g006 Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 6 May 2014 | Volume 9 | Issue 5 | e97558 Table 1. Gene ic di e si y o he Finnish b own bea popula ion. B own bea s in Finland 2006–2010 - Gene ic di e si y No h (N = 164) Sou h (N = 122) No h and sou h ( N = 286) Locus AH E H O F IS AH E H O F IS AH E H O F IS MU05 10 0.82 0.82 20.001 9 0.76 0.82 20.086* 10 0.81 0.82 20.017 MU09 13 0.89 0.90 20.014 9 0.86 0.84 0.014 13 0.88 0.88 0.006 MU10 10 0.82 0.84 20.021 9 0.75 0.78 20.037 10 0.81 0.81 20.007 MU15 7 0.78 0.76 0.019 7 0.81 0.85 20.047 7 0.80 0.80 0.001 MU23 13 0.89 0.84 0.050 11 0.81 0.80 0.013 13 0.87 0.83 0.056* MU50 10 0.76 0.76 0.008 9 0.58 0.56 0.046 10 0.70 0.67 0.040 MU51 9 0.82 0.84 20.026 7 0.75 0.77 20.032 9 0.81 0.81 20.002 MU59 16 0.90 0.91 20.014 12 0.87 0.83 0.048 16 0.90 0.88 0.021 G1A 10 0.83 0.81 0.031 9 0.82 0.82 0.004 10 0.83 0.81 0.027 G1D 9 0.86 0.87 20.019 8 0.82 0.79 0.038 9 0.85 0.84 0.012 G10B 12 0.87 0.88 20.023 8 0.81 0.81 20.076* 12 0.84 0.88 20.040 G10L 11 0.77 0.76 0.004 8 0.73 0.75 20.027 11 0.75 0.76 20.006 Mean 10.8 0.83 0.83 20.001 8.8 0.78 0.79 20.012 10.8 0.82 0.82 0.008 *P,0.05 Expec ed, (H E ) and obse ed (H O ) he e ozygosi ies, numbe o di e en alleles (A) and inb eeding alues (F IS ) calcula ed o he 12 sho andem epea s in he Finnish b own bea popula ion o 286 indi iduals, sampled om 2006 o 2010. Signi ican F IS alues a e ma ked wi h *. Loci de ia ing signi ican ly om Ha dy-Weinbe g equilib ium a e Bon e oni co ec ion a e highligh ed by bold F IS alues. doi:10.1371/jou nal.pone.0097558. 001 Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 7 May 2014 | Volume 9 | Issue 5 | e97558 applied con inuous sampling co esponding o he es ima ed dis ibu ion o bea s in he a ea and co e ed all possible mig a ion ou es. Ou esul s showed ha he b own bea popula ion in Finland and Russian Ka elia consis s o wo clus e s, a no he n and a sou he n one. The clus e s showed subs an ial geog aphical o e lap and he gene ic di e en ia ion be ween hem was modes , sugges ing a high deg ee o admix u e. Mig a ion analyses suppo ed hese indings and showed ha gene low be ween Finland and Russian Ka elia was high, especially in he eas -wes di ec ion. In compa ison, gene low be ween Finland and Scandina ia appea ed o be es ic ed, and was ound o be absen om he eas owa ds Scandina ia. The s uc u ing o he Finnish bea popula ion [30] was no as expec ed, showing a subs an ial deg ee o o e lap be ween he clus e s. This is mos p obably caused by he s ong in luence o IBD on his popula ion. In he p e ious s udy by Tammeleh e al. [30], 70 Finnish bea samples we e epo ed o show a pai wise F ST be ween wo clus e s o F ST = 0.067. This esul is conside ably highe han in ou s udy and may be an e ec o non-con inuous sampling [59]. Howe e , i may also be explained in pa by in oking he his o y o he b own bea s in he coun y. Wi h he ecen demog aphic bo leneck o he Finnish bea popula ion in mind, i may be possible ha he wo clus e s we e once a single popula ion o bo h popula ions may ha e been connec ed be e in he pas and migh ha e been di ided du ing ime o pe secu ion. Acco ding o his o ical eco ds, he b own bea was i ually ex inc om mos pa s o Finland, wi h excep ion o he a ea sou h-eas and in he no h, nea he bo de o Russia [11]. Du ing he ime o ex ensi e pe secu ion he numbe o indi iduals plumme ed and he popula ion may ha e become subdi ided. The demog aphic eco e y migh ha e connec ed he clus e s again. I ha was he case, he di e en ia ion be ween he wo popula ions was p obably ne e high. The signi ican , albei low F ST alues could be explained by such a scena io [3]. Pe haps he no he n and he sou he n clus e iden i ied ep esen wo lineages o ecoloniza ion o Finland du ing he las decades: one lineage om he sou h-eas and ano he om no h-eas . The occu ence o his o ic mig a ion e en s has been s ongly indica ed by esul s o analyses using he mi ochond ial genome [32]. Howe e , ecen in o ma ion and da a om a eas u he eas emain ague and e i ica ion he e o e is no possible a his poin . No iceable is he high numbe o bea s which could no be assigned unambiguously o any o he iden i ied clus e s when gene ic s uc u e was analyzed on an in e media e scale (Finland and Russian Ka elia). The ambiguously assigned indi iduals we e mainly om he admix u e zone in he sou he n pa o Finland along he bo de o Russia (and in Russian Ka elia) and aise he ques ion on hei o igin. Howe e , no e ha o he clus e assignmen analysis on he la ge scale (incl. Scandina ia), he numbe o ambiguously assigned indi iduals was a he low (3.5%). Hence, bea s ha could no be assigned unambiguously on he in e media e scale we e assigned almos comple ely when compa ed o a mo e dis an popula ion (i.e. Scandina ia). We belie e ha hese esul s sugges he exis ence o a con inuous bea popula ion s uc u ed by IBD in Finland and Russian Ka elia. Consequen ly, we in e p e he low membe ship coe icien s as a likely esul o admix u e be ween subpopula ions. This migh ha e led o di icul ies in clea ly assigning indi idual geno ypes o one o he iden i ied clus e s du ing he analyses. Howe e , in luence o o he bea popula ions o he eas and sou h, e.g. owa ds S . Pe e sbu g and u he sou h o Es onia [30] may also be possible. I he bea popula ions u he eas o sou h sha e indeed he same his o y o pe secu ion, hei eco e y and expansion may explain he g adual inc ease o immig a ing indi iduals om o he popula ions in o Finland. No gene ic bo leneck was de ec ed o he bea s in Finland and Russian Ka elia, as p e iously indica ed o a small pa o he popula ion [28], leading o he assump ion, ha a su icien numbe o indi iduals may ha e su i ed du ing he ime o he demog aphic bo leneck and/o he bo leneck was e y sho in ime; oo sho o lead o a subs an ial loss in gene ic ma e ial. I has been epo ed ea lie , ha Russian bo de ences loca ed along he Finnish-Russian bo de may p e en o a ec wol es oaming in eas -wes di ec ion [60]. Ou esul s showed ha he gene low ac oss he Finnish-Russian bo de has been su icien and ha hose ences may no cons i u e a se ious obs acle o b own bea s. Connec i i y o he Finnish and Russian Ka elian b own bea s wi h popula ions in he wes owa ds Scandina ia seemed mo e es ic ed, as ou p e ious s udy has indica ed [25]. All de ec ed i s gene a ion mig an s in he no h we e iden i ied as indi iduals o igina ing om he Scandina ian popula ion, which mig a ed owa ds eas , in o Finland, poin ing o unidi ec ional gene low. The Scandina ian bea popula ion has i s main dis ibu ion in Sweden wi h ou lie s in o No way. App oxima ely 30 yea s a e he eco e y s a ed, he bea s in Sweden we e di ided in o h ee gene ic clus e s, which co esponded o a eas wi h high concen- a ion o emales [29]. These a eas a e assumed o ep esen his o ic elic a eas, in which a ew bea s ha e su i ed he ime o in ensi e hun [29,61,62]. Despi e o a gene ic bo leneck, he Scandina ian bea s showed ela i ely high le els o he e ozygosi y (H O = 0.66) [31], al hough conside able smalle han ound in Finland and Russia in his s udy. This ema kable misma ch may be he esul o he ex eme di e ences in gene low om Russia. The Finnish bea popula ion is he only connec ion o he Scandina ian bea popula ion o he Russian one. Al hough he Finnish and he Scandina ian popula ions bo h s a ed o hei eco e y om being hun ed down o nea ex inc ion in mos pa s, hei mechanisms o eco e y mus ha e been qui e di e en and his is e lec ed in oday’s gene ic composi ion. Ou esul s show ha he Finnish popula ion p obably has always expe ienced gene low om Russia in compa ison o he Scandina ian bea popula ion, which eco e ed wi hou subs an ial suppo om o he popula ions. We p opose ha u u e s udies should analyze his o ical samples o elucida e he his o y o he b own bea s in Finland, Russian Ka elia and Scandina ia du ing he ime o pe secu ion and ini ial phases o eco e y. Fu he , analyses on ecen mig a ion should be moni o ed and ocus mo e in ensi ely on bea s in di e en egions by applying nonin asi e gene ic sampling and es ima ion o cap u e-ma k- ecap u e p obabili ies. This can esul in easible es ima ions on possible demog aphical changes, such as ep o- duc ion and u n-o e a es as well as he a io be ween e ec i e and census popula ion sizes. Suppo ing In o ma ion Figu e S1 Bayesian clus e ing esul s o No he n Eu opean bea s wi h he p og am S uc u e (P i cha d e al. 2000). Samples we e collec ed om 2006 o 2010. (a) Finnish samples only; (b) Finnish and Russian Ka elian b own bea samples oge he as well as (c) b own bea samples om all o e he sampling ange o no he n Scandina ia, Finland and no h wes e n Russia o he. P esen ed a e he mean likelihoods L(K) and s anda d de ia ions o K= 1 o 10 clus e s o e 10 independen uns (1,000,000 i e a ions and 100,000 bu n-in) and he es ima e o DKusing he app oach desc ibed by E anno e al. Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 8 May 2014 | Volume 9 | Issue 5 | e97558 (2006). G aphs we e plo ed using he web based analysis S uc u e Ha es e (Ea l and onHold 2012). (TIF) Figu e S2 Bayesian clus e ing esul s o he No he n Eu opean bea s wi h S uc u e (P i cha d e al. 2000). Ba plo s showing he assignmen p obabili ies o each bea o he iden i ied clus e s om K= 2 o 5 when only samples om Finland we e analyzed (a) and samples om Finland and Russian Ka elia pooled oge he (b). Indi iduals a e a anged by la i ude om no h (le ) o sou h ( igh ). Ba plo s o K= 2 o 5 when all da a (Finland, Russian Ka elia and Scandina ia) is analyzed oge he (c). No able is he inc ease o unassigned indi iduals (q,0.7) wi h inc easing K. Fo Finland, he numbe o K= 2 was 60 (20.98%), while o K= 3 o 5 he numbe inc eased om 103 (36.01%) o 206 (72.03%). (ZIP) Table S1 B own bea geno ypes (N= 517) used in his s udy. (XLSX) Table S2 The es ima ed numbe o e ec i e mig an s pe gene a ion (Nm). Numbe o mig an s be ween Finland and Russian Ka elia as well as be ween Finland and Scandina ia. Nm esul s a e shown a e co ec ion o di e en sample sizes o 10, 25 and 50 as well as o he mean sample size. (XLSX) Acknowledgmen s In memo y o he la e Minna Ruokonen who deceased p io o publica ion o his s udy. We would like o hank Julia Sch egel and h ee anonymous e iewe s o help ul commen s on his manusc ip . Fu he , we would like o hank he Finnish Hun e s and Finnish Hun e s’ Associa ion, he Swedish En i onmen al P o ec ion Agency and he No wegian S a e Na u e Inspec o a e o collec ing he bea samples used in his s udy. Financial suppo o his p ojec has been p o ided by he No wegian Minis y o En i onmen . Au ho Con ibu ions Concei ed and designed he expe imen s: AK HGE JA IK SBH. Pe o med he expe imen s: AK HGE SBH. Analyzed he da a: AK HGE SBH. Con ibu ed eagen s/ma e ials/analysis ools: IK KFT PID CT SBH. W o e he pape : AK HGE SBH. Commen ed on he manusc ip d a s: JA IK KFT PID. Re e ences 1. Mills LS, Schwa z MK, Tallmon DA, Lai KP (2003) Measu ing and in e p e ing connec i i y o mammals in coni e ous o es s. In: Zabel CJ, An hony RG, edi o s. Mammal Communi y Dynamics: Managemen and Conse a ion in he Coni e ous Fo es s o Wes e n No h Ame ica. Camb idge: Camb idge Uni e si y P ess. 587–613. 2. C ooks KR (2002) Rela i e sensi i i ies o mammalian ca ni o es o habi a agmen a ion. Conse Biol 16: 488–502. 3. Schwa z MK, Mills LS, McKel ey KS, Ruggie o LF, Allendo FW (2002) DNA e eals high dispe sal synch onizing he popula ion dynamics o Canada lynx. Na u e 415: 520–522. 4. T aill LW, B ook BW, F ankham RR, B adshaw CJA (2010) P agma ic popula ion iabili y a ge s in a apidly changing wo ld. Bio Cons 143: 28–34. 5. Kendall KC, S e z JB, Boulange J, Macleod AC, Pae kau D, e al. (2009) Demog aphy and Gene ic S uc u e o a Reco e ing G izzly Bea Popula ion. J Wildl Manage 73: 3–17. 6. Libe g O, Chap on G, Wabakken P, Pede sen HC, Hobbs NT, e al. (2012) Shoo , sho el and shu up: c yp ic poaching slows es o a ion o a la ge ca ni o e in Eu ope. P oc Biol Sci 279: 910–915. 7. Mille CR, Wai s LP (2003) The his o y o e ec i e popula ion size and gene ic di e si y in he Yellows one g izzly (U sus a c os): implica ions o conse a ion. P oc Na l Acad Sci USA 100: 4334–4339. 8. P oc o MF, McLellan BN, S obeck C, Ba clay RM (2005) Gene ic analysis e eals demog aphic agmen a ion o g izzly bea s yielding ulne ably small popula ions. P oc Biol Sci 272: 2409–2416. 9. Ca dillo M, Mace GM, Jones KE, Bielby J, Bininda-Emonds ORP, e al. (2005) Mul iple Causes o High Ex inc ion Risk in La ge Mammal Species. Science 309: 1239–1241. 10. Dale um F, Came on EZ, Kunkel K, Some s MJ (2009) Di e si y and deple ions in con inen al ca ni o e guilds: implica ions o p io i izing global ca ni o e conse a ion. Biol Le 5: 35–38. 11. Nyholm ES, Nyholm KE (1999) S a us and managemen o he b own bea in Finland. In: Se heen C, He e o S, Pey on B, edi o s. Bea s - s a us su ey and conse a ion ac ion plan. Camb idge: IUCN. 63–67. 12. E mala A (2003) A su ey o la ge p eda o s in Finland du ing he 19 h –20 h cen u ies. Ac a Zool Li 13: 15–20. 13. Pulliainen E (1983) Beha io o an expanding popula ion o he b own bea (U sus a c os) in No he n Eu ope. Mamm Biol 48: 290–297. 14. Kojola I, Ma¨a¨ a¨ E (2004) Suu pe ojen lukuma¨a¨ a¨ ja lisa¨a¨n uminen uonna 2003. Helsinki. 1–7 p. 15. Nyholm ES (1995) Suomen suu pe okanna uosina 1991–1993. Ahma, il es ja ka hu unsas unee . Helsinki. 1–9 p. 16. Kojola I, Danilo PI, Lai ala HM, Belkin V, Yakimo A (2003) B own bea popula ion s uc u e in co e and pe iphe y: analysis o hun ing s a is ics om Russian Ka elia and Finland. U sus 14: 17–20. 17. Pulliainen E (1990) Recoloniza ion o Finland by he b own bea in he 1970s and 1980s. Aquilo Zool 27: 21–25. 18. Pulliainen E (1997) The expansion o b own bea s om eas in o Finland. In e na ional Bea News 6: 10–11. 19. Wikman M (2010) Riis akanna 2010. Riis aseu an ojen ulokse . Helsinki: Riis a- ja kala alouden u kimuslai os. 1–47 p. 20. Kojola I, Heikkinen S (2006) The s uc u e o he expanded b own bea popula ion a he edge o he Finnish ange. Ann Zool Fennici 43: 258–262. 21. Kojola I, Lai ala HM (2000) Changes in he s uc u e o an inc easing b own bea popula ion wi h dis ance om co e a eas: ano he example o p esa u a ion emale dispe sal? Ann Zool Fennici 37: 59–64. 22. Ches in I (1999) S a us and managemen o he b own bea in Russia. In: Se heen C, He e o S, Pey on B, edi o s. Bea s - s a us su ey and conse a ion ac ion plan. Camb idge: IUCN. 23. Danilo PI (1994) The b own bea o No hwes Russia. In Con Bea Res Manage 9: 199–203. 24. Eiken HG, And eassen RJ, Kopa z A, Bje amoen SG, Wa iainen I, e al. (2009) Popula ion da a o 12 STR loci in No he n Eu opean b own bea (U sus a c os) and applica ion o DNA p o iles o o ensic casewo k. Fo ensic Sci In Gene ics Supplemen Se ies 2: 273–274. 25. Sch egel J, Kopa z A, Hagen SB, B ose h H, Smi h ME, e al. (2012) Limi ed gene low among b own bea popula ions in a No he n Eu ope? Gene ic analysis o he eas -wes bo de popula ion in he Pas ik Valley. Mol Ecol 21: 3474–3488. 26. Kindbe g J, Swenson JE, E icsson G, Bellemain E, Miquel C, e al. (2011) Es ima ing popula ion size and ends o he Swedish b own bea U sus a c os popula ion. Wildl Biol 17: 114–123. 27. And eassen R, Sch egel J, Kopa z A, Tobiassen C, Knappskog PM, e al. (2012) A o ensic DNA p o iling sys em o No he n Eu opean b own bea s (U sus a c os). Fo ensic Sci In Gene 6: 798–809. 28. Kopa z A, Eiken HG, Hagen S, Ruokonen M, Espa za-Salas R, e al. (2012) Connec i i y and popula ion subdi ision a he inge o a la ge b own bea (U sus a c os) popula ion in No h Wes e n Eu ope. Conse Gene 13: 681–692. 29. Manel S, Bellemain E, Swenson JE, F ancois O (2004) Assumed and in e ed spa ial s uc u e o popula ions: he Scandina ian b own bea s e isi ed. Mol Ecol 13: 1327–1331. 30. Tammeleh E, Remm J, Ko s en M, Da ison J, Tumano I, e al. (2010) Gene ic s uc u e in la ge, con inuous mammal popula ions: he example o b own bea s in no hwes e n Eu asia. Mol Ecol 19: 5359–5370. 31. Wai s L, Tabe le P, Swenson JE, Sandeg en F, F anzen R (2000) Nuclea DNA mic osa elli e analysis o gene ic di e si y and gene low in he Scandina ian b own bea (U sus a c os). Mol Ecol 9: 421–431. 32. Keis M, Remm J, Ho SYW, Da ison J, Tammeleh E, e al. (2012) Comple e mi ochond ial genomes and a no el spa ial gene ic me hod e eal c yp ic phylogeog aphical s uc u e and mig a ion pa e ns among b own bea s in no h-wes e n Eu asia. J Biogeog 40: 915–927. 33. Pae kau D, Cal e W, S i ling I, S obeck C (1995) Mic osa elli e analysis o popula ion s uc u e in Canadian pola bea s. Mol Ecol 4: 347–354. 34. Pae kau D, S obeck C (1994) Mic osa elli e analysis o gene ic a ia ion in black bea popula ions. Mol Ecol 3: 489–495. 35. Tabe le P, Cama a JJ, G i in S, Uh es E, Hano e O, e al. (1997) Nonin asi e gene ic acking o he endange ed Py enean b own bea popula ion. Mol Ecol 6: 869–876. 36. Linac e A, Gusmao L, Hech W, Hellmann AP, May WR, e al. (2011) ISFG: Recommenda ions ega ding he use o non-human (animal) DNA in o ensic gene ic in es iga ions. Fo ensic Sci In Gene 5: 501–505. Reco e ing No he n Eu opean B own Bea PLOS ONE | www.plosone.o g 9 May 2014 | Volume 9 | Issue 5 | e97558