Genetic and Cultural Reconstruction of the Migration of an Ancient ...

Research ArticleGenetic and Cultural Reconstruction of the Migration ofan Ancient Lineage

Desmond D Mascarenhas1 Anupuma Raina2

Christopher E Aston3 and Dharambir K Sanghera3

1Mayflower Organization for Research and Education Sunnyvale CA USA2Department of Forensic Medicine and Toxicology All India Institute of Medical Sciences New Delhi 110029 India3Department of Pediatrics University of Oklahoma Health Sciences Center Oklahoma City OK USA

Correspondence should be addressed to Desmond D Mascarenhas desmondmayflowerworldorg

Received 10 May 2015 Revised 24 August 2015 Accepted 25 August 2015

Academic Editor Peter J Oefner

Copyright copy 2015 Desmond D Mascarenhas et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

A rare R1a1 Y-haplogroup (Y-HG) L657 clade subtype designated as LPKSTR is found in most male members of a clan of ldquofounderrdquofamilies within the Goud Saraswat Brahmin community in Lotli town in Western India TMRCA calculations using pairwisecomparisons to control cohorts suggested a probablemigration history distinct from the canonical narrative formedievalmigrationof orthodox Brahmin families to South India Using Y-HG centroid analysis chi-square analysis of TMRCA distributions andarcheological find-spots and discriminant function analysis we show that the parental Z93 L3422 subclade in which LPKSTRoccurs originated in West Asia and that LPKSTR individuals migrated toward the southeast by a Bolan Pass route distinct fromthe traditionally presumed route of Brahmin ingress into the Indian subcontinent The proposed migration route is supported byarcheological toponymic numismatic linguistic iconographic and literary data Lastly we present cultural metrics demonstratingthat these LPKSTR lineages retained distinct family practices with respect to literacy religious practice and emigration not sharedwith orthodox Brahmins of canonical geographic origin within the same community despite centuries of intermarriage Long-termtransmission of differentiated family practices within a patrilineal endogamous community has rarely been documented

1 Introduction

Increasingly granular subclade typing and the rapid growth ofpublic databases have improved our ability to reconstruct thespread of R1a1 Y-haplogroup (Y-HG) branches along ancienttrade and migration corridors in Asia These new insightshave established the dominant clade of R1a1 in the Iran-Afghanistan migration corridor to South Asia as carrying thedefining SNPs Z93 and L3422 In this group the L657 branchis most abundant [1 2]

In order to further dissect patrilineages within L657unique microsatellite short tandem repeat (STR) profiles canbe helpful In this work we attempt to reconstruct a contextfor the migration of a culturally defined and highly endoga-mous group of priestly families carrying the rare haplotypeldquoLPKSTRrdquo defined as R1a1 Y-HG individuals carrying the

STR profile DYS456 lt 16 DYS458 gt 15 and GATAH4 gt 12LPKSTR individuals listed in public databases also carry theSNPs Z93 and L3422 The LPKSTR type is essentially absentfrom European R1a1 individuals abundant in Afghanistanand quite rare in L657 individuals from South Asia and theArabian Peninsula

Brahmin communities of South Asia have documentedcultural and patrilineal antecedents (such as gotra) sinceancient times [3ndash7] Based in part on the ancient system ofgotra the prevalence of strictly endogamous patrilineageswithin Brahmin communities in South Asia simultaneouslyoffers possibilities for studying horizontal transmission ofdistinct family practices over long intervals while also com-plicating the selection of geographic comparators tradition-ally employed in statistical methods A Brahmin familyY-chromosome lineage is likely to exhibit greater genetic

Hindawi Publishing CorporationBioMed Research InternationalVolume 2015 Article ID 651415 16 pageshttpdxdoiorg1011552015651415

2 BioMed Research International

proximity to a culturally related group situated thousandsof miles away than to any of the dozens of geographicallyproximate communities that might traditionally serve ascomparators Thus in assessing genetic proximity withinhighly differentiated Brahmin milieu the rational choice ofselected comparators for TMRCA computation requires theuse of historical data [8]

Ancient tradition places Saraswat Brahmins with theSarasvata tribe on the banks of the Rig Vedic Sarasvati Riverwhose exact geographic location is not certain but is nowbelieved by some scholars to be the Helmand-Arghandabbasin in Afghanistan [8 9] though this conclusion is stillcontroversialThe region in questionwas settled prior to 4000BCE and its ancient pre-Indo-European namewasHaraxvaiti(later HarahuvatiSarasvati) [10 11] In first millennium BCEpost-Achaemenid Sanskrit literature places Saraswat Brah-mins in the northern Punjab and a substantial Saraswat com-munity remains there to this day The other large Saraswatgroup in South Asia is the Goud Saraswat Brahmin (GSB)community whose arrival in Goarsquos Salcete province duringearly medieval times is memorialized in local oral historiesand legends variously claiming western (Gujarat-Rajasthan)central (Kannauj) or eastern (Tirhut) proximal origins It islikely that todayrsquos Goan GSB community is the product ofmultiple migration events between the 7th and 11th centuriesCE each comprising families of claimed or actual tribalkinship with the original Sarasvata tribe

Within the GSB Brahmin community a high incidence ofLPKSTR was anecdotally observed in male representatives ofPai (surname) families of a single gotra (exogamous lineageKaundinya) who are known founder families in a towncommunidade (Lotli) GSB are a subset of all Saraswatswho are a subset of all Brahmins Table 1(a) describesthese identifiers These so-called Lotli Pai Kaundinya (LPK)families may have retained distinct family practices despitecenturies of intermarriage with other GSB families of neigh-boring towns In this work we explore the hypothesis thatunlike other Brahmins in the community the LPK clanentered Goa via a western migration route from easternPersia and Afghanistan quite distinct from the traditionalnarrative for ingress of Vedic peoples into the subcontinentvia a northern route through Gandhara and the KhyberPass into the PunjabMadhyadesa in the first millenniumBCE and subsequently southward from there in medievaltimes We compare genetic and cultural characteristics ofLPK families with another clan within the same endogamousGSB community with whom they have intermarried Oraland written records of a unique form of town governanceknown as the ldquocommunidaderdquo confirm each clanrsquos identityand privilege in their respective towns for centuries [12ndash14] Samples of Khatri (mercantile Bolan Pass) and SaraswatBrahmin (priestly Khyber Pass) communities are used asgenetic comparators in this study because their cultural andmigration histories represent the two alternatives we wish tocompare

Amongmodernmigrants social success and family prac-tices prior to migration predict relative social success aftermigration [15] Such observations point to portable family-borne attitudes and capabilities including for example

business and educational skills or the pragmatism requiredfor cultural adaptationThe matter has implications for otherdisciplines in the study of ancient history for example theldquopots-are-not-peoplerdquo maxim in archeology [16]

2 Materials and Methods

21 Samples and Subjects The collection under informedconsent of 35 male Saraswat Brahmin and 29 male KhatriDNA samples (listed in Table 2) has been described inprevious publications [32ndash34] 16 GSB samples (15 listedin Table 1(b) one non-R1a1 GSB individual (Q1a3 Y-haplogroup) was omitted from the analyses) were collectedas follows volunteers registered for cheek swab DNA anal-ysis directly with Genebase Inc (Vancouver BC) signedinformed consent forms and followed requisite proceduresas required by the Genebase Testing Laboratory Each volun-teer ordered Y-HG subclade and 67-STR tests Volunteerswere reimbursed for the cost of testing but received noother compensation After testing was complete volunteersprovided investigators with access to their results on theGenebase database Permission to use data for research andpublication was provided in writing

22 Ethics Statement Studies were approved by each ofthe participating organizationsrsquo institutional committees incharge of oversight for ethical conduct in research Thecollection and research use of Khatri DNA samples wereapproved by the Institutional Review Board of the Universityof OklahomaHealth Sciences Center OK (IRB number 2911)and the Ethical Committee of Hero DMC Medical Collegeand Heart Institute Ludhiana Punjab India The GSB studywas approved by the Ethical Committee of the MayflowerOrganization for Research and Education Sunnyvale CAThe collection and research use of Saraswat Brahmin DNAsamples were approved by the Institute Ethical ClearanceBody of the All India Institute of Medical Sciences NewDelhi

23 DNA Analysis Typical tests included the standard 67-STR Y-backbone and subclade testsThe details of these testscan be found on the Genebase website (httpwwwgenebasecom) Assignment of Y-HG and subclade in this study usedthe tests for the markers and single nucleotide polymor-phisms (SNPs) listed in Table S1 (see SupplementaryMaterialavailable online at httpdxdoiorg1011552015651415) Themicrosatellite short tandem repeats (STR) tested are listed inTable S2

24 Validation of Lineages Volunteers were identified andvalidated using oral communications with town social net-works communidade records electoral rolls church bap-tismal and marriage records going back 150ndash400 years forChristian subjects and the Ramnathi temple community forHindu participants Brahmin status for Christian participantswas confirmed from marriage records showing endogamywith other Brahmin families over several generations (Sup-plementary files S3 S4 S5 and S6)

BioMed Research International 3

Table 1 (a)Definition of social terms and identifiers (b) Characteristics ofGSB cohort and controls used in this study See text for descriptionsof LPKSTR haplotype vangor and gotra and Supplementary Files S3ndashS6 and S37 [12]

(a)

Identifier DescriptionTown communidade(eg Lotli Kudtari)

Ancient system of male hereditary town governance in which each of the original founder families ofthe town owns a share (vote) with land ownership rights and tax obligations

Vangor A share (vote) in a communidade that is a patrilineal lineage often using a single surnameBrahmins Members of an ancient hereditary male priestly caste codified in the first millennium BCESaraswat Brahmins The most ancient of all Brahmins associated with the eponymous Sarasvati River and Sarasvata tribeGoud Saraswat Brahmins(GSB)

An endogamous branch of the Saraswat Brahmin community formed by early medieval (8thndash10thcentury CE) migration event(s) to Goa in Western India

Gotra (eg Kaundinya) Ancient Brahmin system of tracking patrilineage gotra is exogamous that is a member must marryoutside it this system has been strictly observed for millennia

Kuldevclan deity(eg Ramnath)

Ancient tradition in which a GSB clan maintains a primary deity at a single place of worship this is areliable characteristic for tracking family lineages

Lotli Pai Kaundinya (LPK) GSB founder families of Lotli communidade of Kaundinya gotra with the surname of ldquoPairdquo

(b)

ID Cohort Y-HG clade STRa Source location Vangor Gotra Ancient lineageLKR-Vg51 GSB L3422+ L657+ No Lotli (Loutolim) Lotli 5 nalowastlowast PaiLKR-Vg71 GSB L3422+ L657+ Yes Lotli (Loutolim) Lotli 7 nalowastlowast PaiLKR-Vg81lowast GSB L3422+ L657+ Yes Lotli (Loutolim) Lotli 8 nalowastlowast PaiLKR-Vg82lowast GSB L3422+ L657+ Yes Lotli (Loutolim) Lotli 8 nalowastlowast PaiLKR-Vg83lowast GSB L3422+ L657+ Yes Lotli (Loutolim) Lotli 8 nalowastlowast PaiLKR-Vg101 GSB L3422+ L657minus No Lotli (Loutolim) Lotli 10 nalowastlowast PaiLKR-Vg102 GSB L3422+ L657minus No Lotli (Loutolim) Lotli 10 nalowastlowast PaiLKR-Vg121 GSB L3422+ L657+ Yes Lotli (Loutolim) Lotli 12 nalowastlowast PaiLKR-Vg151 GSB L3422+ L657+ Yes Lotli (Loutolim) Lotli 15 nalowastlowast PaiLKR-VgU1 GSB L3422+ L657+ No Lotli (Loutolim) Unknown Kaundinya PaiLKR-VgU2 GSB L3422+ L657+ Yes Lotli (Loutolim) Unknown Kaundinya NayakLKR-VgU3 GSB L3422+ L657+ No Lotli (Loutolim) Unknown Kaundinya PaiPBR11 GSB L3422+ L657minus No Verem Unknown Kaundinya PaiCK11 GSB L3422+ L657+ No Kudtari (Curtorim) Kudtari 8 na KamatCK12 GSB L3422+ L657+ No Kudtari (Curtorim) Kudtari 8 na KamatKH11 Khatri L3422+ L657+ Yes Punjab None na naKH12 Khatri L3422+ L657+ No Punjab None na naKH13 Khatri L3422+ L657+ No Punjab None na nalowastReference individuals for TMRCA computations lowastlowastKaundinya gotra prior to conversion na not applicable

25 TMRCA Computation The largest number of individualsamples (three) available from a single LPK Vangor lineagewas from Lotli Vangor 8 (V8) (Supplementary File S4 andTable 1(b)) Consequently we selected V8 as a reference STRprofile 59 of 63 loci were identical in these three individualsplus another V8 individual not reported in this study Modalvalues for the remaining loci were used in the referenceprofile All V8 pairwise STR comparisons utilized onlyR1a1 Y-HG individuals (in the case of GSB Khatri PunjabiSaraswat Gulf Arab and Indian samples all R1a1 individualscarried Z93 and L3422 SNPs individuals from the publishedcohorts were not tested for these SNPs) At least 40 commonSTRs were used for each comparison listed in Figure 1 andthe dendrogram was constructed using the TMRCA valuesas branch-points From church records between the late

seventeenth century and the mid-twentieth century whenmarriage and child-bearing cultural practices are believed tohave remained stable the mean generation time was 30 plusmn 2years In each pairwise TMRCA computation we input 28and 32 years as generation times and averaged the valuesin each case As this procedure yields proportional errorvalues these are not reported Other variables in the formula(eg mutation rate of 0025) were heuristically determinedusing 5 known branch-points of test subjectsrsquo ancestors fromgenealogical histories going back to ca 1600 CE For Figure 2published data for the indicated cohorts allowed pairwisecomparison of 17 or more STRs Bidirectionally labile STRswere excluded from all comparisons TMRCA computationsused the following settings average mutation rate 00025gt66 probability and average generation of 28 years These


Table 2 Punjabi Khatri and Saraswat cohorts

(a)

ID Haplogroupsubclade Community1139 L3 (M357+) Khatri159 G2a3 (P15+ S126+) Khatri173 R1a1a1h1a (L3422+ L657+) Khatri20034 G2a3 (P15+ S126+) Khatri20062 J2b (M12+) Khatri20150 L3 (M357+) Khatri20180 J2a (M410+) Khatri214 R1a1a1h1a (L3422+ L657+) Khatri216 J1 (M267+) Khatri30282 R1a1a1h1a (L3422+ L657+) Khatri30809 R1a1a1h1 (L3422+ L657minus) Khatri30944 L3 (M357+) Khatri31311 L3 (M357+) Khatri40029 J2b (M12+) Khatri40036 R2 (M479+ M124minus) Khatri40038 J1 (M267+) Khatri5002 R1a1a1h1a (L3422+ L657+) Khatri5003 R1a1a1h1a (L3422+ L657+) Khatri5005 R1a1 Khatri5018 L1 (M76+) Khatri6008 J1 (M267+) Khatri6012 R1a1a1h1a (L3422+ L657+) Khatri6024 R2 (M479+ M124minus) Khatri6036 R1a1a1h1 (L3422+ L657minus) Khatri6040 R2a1a (L294+) Khatri6043 J2a4b (M67+) Khatri6230 R2 (M479+ M124minus) Khatri934 L3 (M357+) Khatri940 L3 (M357+) Khatri

(b)

ID Haplogroupsubclade CommunityHS21 R1a1a1h1a (L3422+ L657+) SaraswatHS26 J2a4 (L26+) SaraswatHS28 L (M11+) SaraswatHS31 R2 (M479+) SaraswatHS32 L (M11+) SaraswatHS33 H1 SaraswatHS34 R1a1 SaraswatHS38 Q1a3 (M346+ M3minus) SaraswatHS39 R1a1 SaraswatHS4 R1a1a1h1 (L3422+ L657minus) SaraswatHS40 L (M11+) SaraswatHS44 R1a1a1h1a (L3422+ L657+) SaraswatHS48 L (M11+) SaraswatHS51 L (M11+) SaraswatHS54 R1b1a2a1 (L150+) SaraswatPS10 Q1a3 (M346+ M3minus) SaraswatPS15 R1a1a1h1a (L3422+ L657+) SaraswatPS18 H1 Saraswat

(b) Continued

ID Haplogroupsubclade CommunityPS22 R2 (M479+) SaraswatPS25 R1a1 SaraswatPS29 R1a1a1h1a (L3422+ L657+) SaraswatPS31 R1a1a1h1a (L3422+ L657+) SaraswatPS32 R2 (M479+) SaraswatPS35 R2 (M479+) SaraswatPS36 L (M11+) SaraswatPS37 R1a1a1h1a (L3422+ L657+) SaraswatPS47 H1 SaraswatPS51 H1 SaraswatPS53 R2 (M479+) SaraswatPS54 J (M304+) SaraswatPS57 R1a1 SaraswatPS60 R1a1 SaraswatPS7 J (M304+) SaraswatPS8 C SaraswatPS9 R1a1 Saraswat

variables were selected based on five pairwise comparisonsbetweenGSB individuals of knownTMRCAs between 50 and400 years based on baptismal and other historical recordsBy restricting comparisons to the same subclade excludingbidirectionally labile STR markers and carefully calibratingformula settings with the aid of attested historical data weexpect our TMRCAs to be more accurate than those derivedin traditional studies thus increasing confidence that thesecomputations may be accurate even over short time spans

26 Public Databases Y-haplogroup (Y-HG) data weredownloaded from public databases at familytreednacom inOctober 2013 (httpswwwfamilytreednacompublicR1aY-Haplogroup andhttpswwwfamilytreednacompublicR1a)The latter link provides a useful diagram of R1a subcladesincluding the major Z93+ branches of Persian Arabian andSouth Asian provenance

27 Data Availability Statement All nonpersonally identifi-able data used in the analyses is included either in the tablesand figures of themain document and in supplementary filescited published references or in the public databases refer-enced Individual STR profiles were deemed to be personallyidentifiable because the families of LPK Vangors are wellknown and every male in each lineage would be expectedto carry a nearly identical STR profile For this reason actualSTR data are not disclosed including those used for TMRCAcalculations

28 Toponymy Numismatics and Iconography Toponymicstems were manually counted for each of the 20 major statesin India using the database at httpswwwfacebookcomwhoyizfref=nf (viewed in September 2012) which lists alltowns and villages for each state Numismatic and icono-graphic data were collected online using the Google search


PBR11

LKR-Vg101LKR-Vg102

LKR-Vg51

KH111213

CK1112

LKR-VgU3

LKR-VgU2LKR-VgU1

LKR-Vg151LKR-Vg121

LKR-Vg71

LKR-Vg83

LKR-Vg82

150012009006003000

1286 BCE375

695

697

792 960

1425

1425

1800CE

840

1765

1880

1635

1530

Location of towns

LPK families TMRCA dendrogram

L3422+ L657minus

L3422+ L657+

LKR-Vg81

C Kudtari (Curtorim)G GuirdolimK KushasthaliL Lotli (Loutolim)

Q Keloshi (Quelossim)R RaiaV Nagoa-Verna

Portuguesefort (1500 s)

Figure 1 Dendrogram of LPK individuals TMRCA branch-points for GSB cohort and controls TMRCAs between pairs of L3422+individuals were computed using the modifications and validation steps for short time intervals described in Sections 1 and 2 Star symboland dark horizontal bar indicate medieval Brahmin migration event CE = common era C = Kudtari G = Guirdolim K = Kushasthali L =Lotli Q = Keloshi R = Rai V = Nagoa-Verna

engine or compiled from the published references cited SeeSupplementary Files S17 S18 S19 and S20

29 Gotra Data on the gotras of Brahmin brides and groomswere collected and tabulated manually from two majormatrimonial sites httpwwwjeevansathicom (viewed inMay 2012) and httpwwwshaadicom (viewed in January2011) The data were quite similar As the sample size wasmuch larger from the first source only those data are reportedhere (Supplementary File S25)

210 Church Records In addition to the references citedchurch records for Lotli town for the period 1914ndash1950were obtained on microfilm from the Church of Latter DaySaints Salt Lake City UT As these records are photographedfrom poorly stored often water- or insect-damaged bookshandwritten by parish priests in Portuguese not all recordscould be satisfactorily recovered or deciphered Neverthelesswe estimate that at least 90 of all records were successfullytabulated (Supplementary File S5)

211 Statistical Methods Probability values (119901 values) werecomputed using Studentrsquos 119905-test and expressed relative toindicated controls except where otherwise noted Group

size was as noted in each case For ldquofellow travellerrdquo dataa discriminant function analysis was run to show thatgroups determined by genetic classification were significantlyseparate geographically based on distances to nearest pointon the northern and southern migration routes These samedistances were then used in an 119865-test analysis of the geo-graphic variation of each genetic group from each migrationroute testing the null hypothesis that it was not different fromeach route

For centroid analysis comparison of standard distances(square root of the average of the squared latitude andlongitude differences from the centroid of the group) betweenpairs of populations used 119865-test analysis equivalent to test-ing for differences in variance between two groups Linearregression was used to test for significant linear reductionin standard distances along the geographic line of migration(based on latitude and longitude of the centroids) Linearregression also was used to determine direction of migrationthrough related centroids The distributions of individu-als across TMRCA determined categories were comparedbetween populations using chi-squared tests of the nullhypothesis that the distributions were equalThe chi-squaredstatistics from each of the pairwise population comparisonswere then compared using 119865-tests to determine if one pair of


50

25

50

25

50

25

50

25

50

25

GSB

KHT GLF

SCT

TMRCA branch-points

Sam

ple (

)

a b c d e f g a b c d e f g

BCTlowast

TJKlowast

PSBlowast

ZABlowast

PSTlowast

GDHlowast

V8 Vangor 8 (reference GP)GSB Goud Saraswat (nonreference GP) KHT Khatri ZAB Zabuli PST PashtunGDH Gandharan BCT BactrianGLF Gulf ArabSCT SC IndianPSB Punjab Saraswat TJK Tajik

a pre-Vedic (lt1500 BCE)

b Vedic (1500-900 BCE)

c classical (900-300 BCE)

d early historic (300 BCE

e N-S migrations (300ndash900 CE)

ndash900 CE)

f late medieval (900ndash1500 CE)

g western colonial (gt1500 CE)

Bolan Pass route Khyber Pass route

V8

KHT

KHT

GSB

GSB

ZAB

ZAB

PST

PST

GDH

GDH

00914 00011

00227

00000 00000 00000

0000000010

00923

00001

0000100007

00526

03433 03356

00314

04931

TJK

00021

00983

02076

01776

Intercommunity relatedness

Figure 2 TMRCA branch-point histogram Each bar represents a 600-year interval For each community pairwise TMRCA computationswere performed between each R1a1 Y-HG individual and a reference modal STR profile of 3 LPK individuals of Vangor 8 (see Section 2)TMRCAs are plotted as a percentage of each population Inset at top right shows 119901 values (chi-square) for histogram differences betweenGSB and communities located near the Bolan and Khyber Passes GSB = Goud Saraswat Brahmins KHT = Khatri ZAB = Zabulistan PST =Pashtun GDH = Gandhara BCT = Bactria GLF = Persian Gulf Countries SCT = South East and Central Peninsular India PSB = PunjabiSaraswat Brahmins TJK = Tajiks asterisks indicate that the comparison was made with a 17-STR profile (see Supplementary Files S3ndashS7)


populations was closer (ie more related) than another pairData analyses were conducted using Microsoft EXCEL 2013and IBM SPSS Statistics (IBM Corp released in 2011 IBMSPSS Statistics forWindows Version 200 Armonk NY IBMCorp) Results corresponding to 119901 values lower than 5 aredescribed as significant and reported

3 Results

31 Validation of LPK Cohort and Comparison with Controls16 male representatives of Brahmin founder families ofancient vangors in the towns of Lotli and Kudtari of Salceteprovince provided cheek swabs under informed consentTheir DNA was analyzed and compared with controls asdescribed in Section 2 and their identifying characteristicsare summarized in Table 1 All GSB samples (save for oneQ1a3 individual who was not studied further) were mappedto the Z93 L3422 subclade and 12 males were additionallyL657 Ten individuals were of the LPKSTR haplotype basedon STR profile Pairwise comparisons of STR profiles using40ndash65 common loci per comparison were used to establishTMRCAs as described in Section 2 The Khatri and Kud-tari control samples were compared as modal profiles Asthe actual branch-points of five pairwise comparisons wereknown or inferred from baptismal and other records goingback four centuries it was possible to calibrate the TMRCAcomputation variables precisely lending confidence to theremaining branch-point assignments A dendrogram of theseTMRCA computationswas shown in Figure 1 (left panel withtown locations shown on the map in the right-hand panel)

40 of all branch-points between males of Lotli vangors(and between them and the Khatri controls) are clusteredbetween 695 CE and 960 CE This interval coincides withthe peak period of medieval north-south Brahmin migra-tions [6] southward expansion of maritime trade [35]Arab invasions of Sindh and Saurashtra in 712ndash738 CE(Supplementary File S43 and references) a documented 8thcenturymigration of Chavda (ldquoChardordquo) Gujjars to Goa [36]and upheavals and opportunities caused by the mid-eighthcentury transition from Chalukya to Rashtrakuta imperialrule provide additional backdrop to this historical periodThus these branch-points would seem to agree with folkloricaccounts placing the initial GSBmigration event to Salcete in740 CE The proximal origin of this migration was probablySaurashtra based on the naming of their first new settlement(Kushasthali) after a famous ancient city in Saurashtra (mod-ern Dwarka)

An additional set of three branch-points maps to theinterval 1435ndash1530 CE immediately preceding and followingthe arrival of Jesuit missionaries and the Portuguese Inqui-sition in 1560 Portuguese documents record the flight ofsome Hindu LPK carrying with them the ancient idol of thetown deity Ramnath [37] LPK council seats (ie vangors)vacated during the exodus may have been purchased andheld by the other LPK of the original council Two L657control individuals (CK11 CK12) of families within the samevangor lineage in the neighboring town of Kudtari share anearly identical STR profile (6263 loci identity) and appearto be only distantly related to the LPK (TMRCA of ca

3200 years going back to early Vedic times) The familiesof Lotli and Kudtari claim a common membership in theGSB community but Kudtari Brahmins may derive from adistinct migration event from Madhyadesa (Tirhut in Biharaccording to tradition)

We next examined the presumed kinship between theLPK samples and two communities in Punjab (see Section 1)Table 2 shows the results obtained from Y-HG and subcladeanalysis of 64 Khatri and Saraswat Brahmin males Kha-tris a mercantile community are the traditional patrons ofSaraswat priests L657 Khatri individuals were the biggestsubclade in this cohort and 3 of 7 L657 individuals testedwere LPKSTR Unfortunately the Saraswat Brahmin DNAsamples available to us were insufficient for STR analysis inthe present study but a report of 17-STR profiles for Punjabiand Himachali Saraswat Brahmins has been published [32]Surprisingly LPKs appear more closely related to secularKhatri individuals than to Saraswat Brahmins Khatris arehistorically associated with the ancient city of Multan insouthwestern Punjab close to the Bolan Pass

These results taken together suggest that the ancientLPK families who form the main subject of this study are aclose-kinship clan of LPKSTRmigrants most likely from thenorthwest and more closely related to Khatris than PunjabiSaraswats

32 Route and Direction of Migration We next examinedgenetic relatedness using Y-HG data from public databasesand published studies

321 TMRCA Histograms Figure 2 shows TMRCA branch-point histograms created using pairwise comparisons of R1a1Y-HG individuals from various communities in the PersianGulf-South Asia region as described in Section 2 Eachhistogram bar represents the percentage of individuals ineach community whose TMRCA with the LPK reference(modal Vangor 8 seeMethods) falls within an interval of 600yearsThismethodology reveals the distribution of kinship tothe reference sample in various geographic regions

The inset table (Figure 2 top right) shows 119901 values fromTMRCA histogram comparisons between the LPK referenceand five communities located in proximity either to the BolanPass (KHT ZAB) or to Khyber Pass (PST GDH) and one(TJK) further away on the Khyber route Using a Bonferronicorrected threshold of 119901 = 00011 for significance shadedcells indicate pairs of communities that are not significantlydifferent As these populations are of similar size the size of119901 values is consistent with the suggestion that the referencesample (V8) is more closely related to KHT and ZAB (BolanPass) than to PST GDH or TJK One explanation of thisresult is that communities near the Bolan Pass are moreclosely related to LPK than Punjabi Sarawats

322 ldquoFellow Travellerrdquo Footprints A previous study showedthat the genetic makeup of communities to the north of theDasht-i-Lut desert in Iran (ldquonorthern routerdquo) is significantlydifferent from the communities to its south (ldquosouthernrouterdquo)Moreover the ingress of Anatolian geneswas deemedto be greater than that fromPakistan suggesting a directional


Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)

Major trade routesPool B (M242 M269 M410 M479)

Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG

KPool A998400 (Pool A minus P15 or M479)

Figure 3 Genetic footprint of haplotypes Coincident presence of SNPs above an arbitrary threshold of 2 (ldquofellow travellerrdquo) was calculatedfrom published genetic data [17ndash25] See Section 2 for details Major trade and migration routes deduced from archeological data are shown(dotted lines and arrows) as well as known archeological sites of the Kura-Araxes civilization (small circles) [26 27] Ovals indicate the areasadjacent to two major mountain passes used for genetic ingress into the subcontinent (Bolan thick oval Khyber thin oval) K = Khatri S =Saraswat Brahmin G = GSB Pool A (P15 M357 M410 and M479) Pool A1015840 (Pool A alleles minus either P15 or M479) Pool B (M242 M269M410 and M479) Pool C (M17 M69 and M479) data from Supplementary File S11

flow fromWest Asia [38] Based on published data on variouscommunities and our analysis of the Punjabi cohort we askedwhich non-R1a1 SNPs observed in the Punjabi samples arealso found above an arbitrary incidence threshold of 2 ineach of the other populations tested This ldquofellow travellerrdquofootprint (Figure 3) is shown for those communities forwhom published data are available in the Transcaucasia-South Asia geographic corridor and adjacent regions [17ndash25] Clusters of coincident SNPs appear to lie along ancientldquonorthernrdquo and ldquosouthernrdquo Persian migration routes betweenTranscaucasia and the northwestern regions of South Asiaas suggested by the previous study [38] Pools A and A1015840which is a subset of A initially follow a north Persian routeand then spread southeast through Kandahar and the BolanPass into Sindh Gujarat and the Deccan The Khatri cohortcontains Pool A Pool B which is characteristic of Nakhcommunities of Hurro-Urartian lineage such as Chechenand Ingush [39 40] runs along the south Persian route intothe Pakhtun belt and is represented in the Punjabi SaraswatBrahmin cohort Note that these two putative migrationroutes cross at Kandahar the suggested location of the ancientSarasvata tribe Pool C which includes Dravidian languagegroups partially overlaps the Pool AA1015840 footprint beginningin Baluchistan

To establish whether Pool A and Pool B are associatedwith the northern and southern routes respectively we first

used discriminant function analysis to show that groupsdetermined by genetic classification were significantly sepa-rate geographically at which they were 119901 = 00009 119865-testanalysis of the geographic variation of each genetic pool fromeach migration route (Figure 4) showed that Pool A was notsignificantly distant geographically from the northern route(119901 = 050) but was significantly distant from the southernroute (119901 lt 00001) while Pool B was not significantly distantgeographically from the southern route (119901 = 050) but wassignificantly distant from the northern route (119901 = 00051)This analysis confirms the association of Pool A with thenorthern Persian trail and its extension through Zabulistaninto western India

323 Incidence of LPKSTR Haplotype LPKSTR is absentfrom R1a1 clades other than Z93 clades which are abun-dant in the corridor from Transcaucasia to South Asia onboth sides of the Persian Gulf (Supplementary File S9)LPKSTR is almost entirely absent from Indian and ArabL657 individuals in the public databases LPKSTR is foundin eastern Transcaucasia and western Afghanistan Table S7shows incidence of the LPKSTR type within R1a1 Y-HGindividuals for ten communities used in the TMRCA branch-point analysis (Figure 2) In addition easternTranscaucasians(Avars Dargins and Lezghins) show a high incidence ofLPKSTR unlike Nakh and Western Transcaucasian R1a1

BioMed Research International 9D

istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0

Distance from northern route0 2 4 6

Nor

ther

n ro

ute

Southern route

Gene Pool AGene Pool B

N clusterS cluster

Figure 4 Discriminant function analysis 119865-test analysis of thegeographic variation of each fellow traveller genetic pool from eachmigration route showing distance (in arbitrary map units) betweenPool A or Pool B and two major routes of migration northernand southern Persian routes (see text) Pool A was not significantlydistant geographically from the northern route (119901 = 050) but wassignificantly distant from the southern route (119901 lt 00001)while PoolB was not significantly distant geographically from the southernroute (119901 = 050) but was significantly distant from the northernroute (119901 = 00051) This analysis confirms the association of PoolA with the northern Persian trail

cohorts suggesting provenance to the eastern side of Tran-scaucasia (Supplementary File S8) Regional communitieswithin Iran are not well represented in this analysis as theirSTR profiles are not publicly available With that caveatLPKSTR abundance appears to track the distribution of PoolA (Figure 2) Interestingly the average incidence of LPKSTRamong R1a1-HG individuals of the Khatri and Zabuli cohortsclosest to the Bolan Pass (361) is four times that of thePashtun and Gandhara communities (91) that live closerto the Khyber Pass consistent with a history of LPKSTRmigration along the so-called northern Persian trail into thesubcontinent

324 Centroid Analysis Earlier published studies have pri-marily shown data from arbitrarily selected and sampledcommunities To gain a less biased snapshot of haplotypedistribution we queried public databases and expressedrelative incidence after correcting for unequal regional repre-sentation In this way it is possible to map virtual geographiccentroids and standard distances (from the centroids) forobligate chronological allele series such as Z93gtZ93 L3422gtZ93 L3422 L657 and Z283 gt Z283M480 The prediction isthat standard distance (SD) should decrease along a series andthat the average position of centroids will reveal a direction

of migration This prediction was observed with SD for Z93L3422 L657 being significantly smaller than that for Z93(119901 = 0034) and Z93 L3422 (119901 = 00006) The SD forZ283 M480 was less than that for Z283 (119901 lt 00001) TheSD for Z93 L3422 L657 was numerically smaller than forZ93 L3422 but did not reach significance (119901 = 0068)Furthermore the Z93 group was significantly distant from theZ93 L3422 group (lat 119901 = 00041 long 119901 = 0032) andZ93 L3422 from Z93 L3422 L657 (lat 119901 = 00004 long119901 = 00037) and a regression line fit to these groups has adirection just slightly south of southeast Figure 5 shows thisfor R1a1 R2a and J2a4 all found in the Khatri cohort Aplot of the incidence of the series (inset in Figure 5) showsthe more archaic alleles of the series preferentially in WestAsia with more recent alleles abundant in South Asia Thefrequency of newer alleles is higher in the ldquoPersianGulfrdquoregion possibly suggesting a general location for their mostsignificant expansion This analysis supports the conclusionthat Z93 and L3422 expanded in a southeasterly directionfromTranscaucasia into SouthAsia By contrast the series forthe R1a1 Z283 M480 clade abundant in the Baltic and Slaviccountries of Europe [1] proceeds north into Europe

Finally we plotted the mean-minus-mode values of thefive nonconserved (ie ldquodriftingrdquo) STRs from the common setavailable for Transcaucasian (TC) Afghan (AF) and SouthAsian (IE) samples Although this test does not establish thedirection of migration it suggests a gradient that is theAfghan sample shows intermediate values betweenTranscau-casian and Indian samples in all five cases (SupplementaryFile S12) consistent with an intermediate location for geneticdrift

325 Archeological Data A remarkable feature of the abovegenetic reconstructions is their compatibility with the arche-ological records of eastward expansion of West Asian pop-ulations in the 4th millennium BCE culminating in the so-called Kura-Araxes migrations of the post-Uruk IV periodwith the advent of wheeled transport [26 27 41] A ca 3000BCE model of a wheeled cart was recovered at Altyndepe[42] New possibly West Asian body types are reportedfrom the graves of Mehrgarh beginning in the Togau phase(3800 BCE) and this influx may have led to early pre-Harappan settlements of the Indus basin Sothi-Siswal andSorath regions between 3200 and 2800 BCE [43 44] Smallcircles in Figure 3 indicate the locations of known Kura-Araxes archeological sites [26 27]

Several lines of evidence support the existence of along-standing ldquoBolan Pass corridorrdquo for genetic and culturalingress from western Afghanistan into the subcontinent

(a) Harappan Iconography Figure 6 shows the location ofarcheological sites in the pre-Harappan period (black dots)which strongly suggests a Bolan Pass route of entry radiatingsouth to coastal Sindh and Sorath as well as northeast tothe Sothi-Siswal area By contrast the Gandharan (KhyberPass route) appears to be much less significant based onthe number of settlements An analysis of the iconographyon Harappan animal seals (2600ndash1900 BCE) reveals a domi-nant ldquounicornrdquo type widely distributed in the archeological


A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia

Z93+ L3422minusZ93+ L3422+ L657minusZ93+ L3422+ L657+

ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast

R1a1a1h1R1a1a1h1aR1a1a1glowast

R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1

Latitude Longitude SDn

28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108

R-Z93+ L3422minus L657minusR-Z93+ L3422+ L657minusR-Z93+ L3422+ L657+R-Z283+ M458minus

R-Z283+ M458+

R-M124+ L295minus L294minusR-M124+ L295+ L294minus

J-L26+ M67minus

J-L26+ M67+ M92minus

J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)

Figure 5 Net direction of migration deduced from centroid analysis Centroids and average standard deviations were computed for severalobligate allele series as described in Section 2 Each circle in the plot represents an allele Thick-edged circles are alleles found in the PunjabiKhatri cohort The bottom panel lists the series used in the computation Inset relative incidence of alleles (corrected for country samplingvariance) in the public database October 2012 Transcaucasia zone included Armenia Georgia Azerbaijan and Turkey Persian Gulf zoneincluded all nations bordering the Persian Gulf plus Afghanistan South Asia zone included India and Pakistan AU = arbitrary units(Supplementary Table S10)

record [45] The provenance of minor animal totems onthe other hand reveals regional differences The caprid-ibex type (a well-known proto-Elamite Persian icon) andthe bovid type (Sorath region) and 3-headed hybrids oftenseen in early 2nd millennium Dilmun seals from the PersianGulf are more prevalent in the Bolan-Gujarat corridor Sealtotems are believed to be reliable indicators of tribal or

religious cultural affinities The distribution of chi-square 119901values from pairwise comparisons of seal find-spots in fiveHarappan regions (Figure 6 top left inset) suggests that thecore Harappan domain represented by Mohenjodaro (region3) and Harappa (region 2) shows similar cultural affinitiesas represented in animal totems Likewise southern Sindh(region 4) and Gujarat (region 5) show similarity based on


Khyber

THARdesert

50 100

5

4

3

2

1

Animal motif seals found in each indicated region ()

Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001

Figure 6 Distribution of find-spots for Harappan animal motif seals Pre-Harappan archeological settlements black dots (map adapted from[28]) are divided into 5 zones (ovals) for the purpose of scoring the incidence of animal seal icons (Supplementary Files S18 S22) Dottedovals show the regions adjacent to the Bolan and Khyber Passes Left top inset 119901 values (chi-square) from pairwise comparisons of regionsLeft bottom inset Elamite-Persian caprid iconography and indigenous (all other animal totems) and hybrid seals are scored as a percentageof all animal seals found in each of the five domains

seal finds But regions 4-5 are clearly distinct from 2 to 3the core Harappan domain These data may indicate thepresence of a nonindigenous proto-Elamite-Persian culturalinfluence along the region 4-5 corridor running throughcoastal Sindh into Gujarat Additional support for a long-standing northwestern Indian cultural corridor is shown inSupplementary Files S41 S42

(b) 2nd Millennium Archeological Footprint The post-Harappan 2nd millennium BCE archeological record of thesubcontinent (Figure 7) is dominated by twomajor traditionsthe Ochre-Colored Pottery horizon which is contiguouswith the Copper Hoard tradition (OCP-CH) of the northand northeast [46 47] and the Black-and-Red Ware (BRW)horizon which originates in Gujarat and covers much of thesubcontinent Based on provenance of the pottery the BRWintrogression throughGujarat is easiest to explain as originat-ing in the northwest and entering the subcontinent throughSindh (not northern Punjab) in the secondmillennium BCEThis is especially seen in a subset of theBRWtradition knownas the White Painted BRW (WP-BRW) whose find-spotsradiate from Gujarat in three directions in a chronological

series (square symbols in Figure 7) Coincident with thesetracks is a separate (and slightly younger) archeological trailof ldquochannel-spouted bowlsrdquo [28] believed to have Persiancultural origins (triangular symbols in Figure 7 see alsoSupplementary Files S15 S16)

326 Medieval Migration In themid-eighth century CE theheartland of Brahmin orthodoxy was located in the northeastof the subcontinent [5] Qualitative evidence in support of thethesis that LPK families migrated to Salcete from a westernrather than eastern proximal origin includes the following

(1) The Geographic Footprint of Clan Deities TMRCA datasupport a mid-eighth century migration of LPK families toSalcete Goa (above) Such an event would be consistent withLPK devotional preferences temples dedicated to their clandeity Ramnath are rare and still retain a distinctly westerncoastal footprint concentrated in the Saurashtra peninsula ofGujarat (Supplementary File S43) Another rare male clandeity brought to Salcete by Brahmin families other thanLPK in the same migration event according to tradition isManganath Ancient locations for Manganath worship are


Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)

2000ndash1700 BCEndash1600 BCE

ndash1200 BCEndash1000 BCE

Figure 7 Second millennium BCE archeological footprint Hypo-thetical migration (arrows) from Khyber and Bolan areas is shownHorizontal hatched zone is the approximate extent of the Harappancivilization (2500ndash1900 BCE) Grey-shaded zones are the OCP-CHhorizons A B C and D (Supplementary File S14) [29 30] Hatchedrectangles are the major areas of known copper ore depositsassociated with the OCP-CH cultures Squares (WP-BRW find-spots Supplementary File S13) and triangles (Persian-style channel-spouted bowl find-spots Supplementary File S15) show temporaland spatial radiation of BRW migration tracks from the deducedGujarat entry area [31]

almost exclusively located in the Saurashtra peninsula Suchstriking colocation of minor religious traditions is unlikelyto be coincidental given the rarity of the epithets RamnathandManganath (rather than Rameshwar andMangesh) Sup-plementary File S31 lists the male deity devotions prevalentin Salcete during the 16th century based on an inventory ofHindu idols in the gt300 temples destroyed by the PortugueseInquisition in 1567 [29] This listing shows that Ramnathwas worshipped as a primary deity only in Lotli town Fora summary of the emergent religious environment in earlymedieval Saurashtra see Supplementary files S27 S28 andS29

(2) Arab Invasion We examined the possible impetus forsouthward migration in 740 CE Supplementary File S43shows the likely route for invading Arab Muslim armies intoGujarat while Al-Junaydwas Emir of SindhThese adventuresinclude a major Arab raid in 725 CE (defeated by a Gurjarking in Malwa) and a final decisive defeat of Arab armies byPulakesi at Navsari in 738-739 CE (Supplementary File S43and references therein) Despite these results Arab raids arereported to have devastatedVallabhi theMaitraka capital the

most likely place of employment for literate Brahmins wholived in Saurashtra at the time

Taken together the above data support the noncanonicalroute of ingress proposed by our hypothesis It should also benoted that genetic differences between LPK and Kudtari con-trol individuals listed in Table 1 are consistent with migrationfrom two different locations (See also Supplementary FilesS33ndash35 S41 and S42)

33 Persistence of Family-Centric Practices within LPK Fam-ilies We investigated LPK family practices over recent cen-turies by tracking metrics that represent secular or pragmaticchoices versus more orthodox alternatives The underlyingpremise here is that the proposed LPK migration througha ldquonon-Brahmanicrdquo geographical zone influenced by Iraniccultures since ancient times and their genetic proximity tothe financially successful (Supplementary File S40) Khatricommunity both suggest secular possibly mercantile tradi-tions for these families by contrast their counterparts inKudtari may have migrated from orthodox Brahmin countryin the northeast (see Section 1) If this assessment is accurateone might expect a more secular tradition within LPKfamilies comparedwith their Kudtari Brahmin neighborsWetherefore examined several metrics of orthodoxy

(a) Rate of Emigration Orthodox Brahmanic tradition dis-courages all forms of transoceanic travel Figure 8 showsextinction rates for LPK and Kudtari GSB (control) vangorsover time presumably because males of founder families lefttown permanently In Figure 8(a) the cumulative extinctionrates for all Brahmin communidade towns of Salcete areshown as of 1848 There is a clear and expected relation-ship between extinction rate and location relative to theancient ingress-egress point for Salcete province the townof Kushasthali (Portuguese Cortalim) Lotli and Kudtari areequidistant and located furthest away from this exit pointThus the extinction rates for these two towns should not beaffected by their location Figure 8(b) (right panel) shows thatthe extinction rate for LPK families has been significantlyhigher than that for Kudtari GSB (119901 = 0037) The simplestexplanation for these data is that the LPK community washistorically more open to emigration suggesting in turn lessorthodox family attitudes

(b) Career Choices We used family data from the eighteenththrough early twentieth centuries to examine the rate atwhichmale offspring joined the seminary [30] versus alterna-tive more secular way of achieving literacy that was availablethrough the professions [48] The data in Figure 8(b) (leftpanel) show that the LPK and Kudtari families emphasizeddifferent routes to literacy with LPKs being more secular(higher ratio of medical school graduates versus ordainedpriests 119901 lt 0008)

(c) Reluctance to Convert to Christianity in the SixteenthCentury Figure 8(b) (middle panel) shows data from the his-tory of early conversion to Christianity Although both Lotliand Kudtari towns were located roughly equidistant fromthe Jesuit mission at Rachol during its first three decades


10 20Distance from geographic egress (kms)

Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)

Figure 8 Emigration and orthodox practices (a) Emigration rate as deduced fromvangor extinction (data fromSupplementary File S39) Seetext for description Each circle represents one Brahmin-founded town White and black circles are Lotli and Kudtari towns respectively (b)Orthodox versus secular community practices (data from Supplementary Files S36 S38) See text for description LPK vangors are comparedwith KK control vangors in the neighboring town Kudtari

(between 1560 and 1586) virtually all Brahmin converts toChristianity originated from LPK families with none fromKudtari families (119901 lt 0021) The greater reluctance toconvert on the part of GSB families of Kudtari may be readas stronger (pre-Christian) religious orthodoxy

(d) Choice of First Names An additional supporting metricis shown in Supplemental File S32 in a census of first namesof male children in the 16th century Kudtari families were 3times as likely as LPK families to select names associated withShiva who was by then of an older religious tradition threecenturies having elapsed since the conversion of most GSBfrom Shaivism to Vaishnavism

(e) Choice of Godparents at Conversion By emergent customthe first Brahmin convert from each vangor in the sixteenthcentury took on the first and last names of the godparentgenerally the baptizing Jesuit missionary The names of allJesuits missionaries posted in Goa during the sixteenthcentury are known from records kept at Society of Jesusheadquarters in Rome From this and other 16th centuryrecords we tentatively identified the godfather after whomeach of the initial Brahmin converts was named The choiceof godfathers for early converts of the LPK families isinformative two of the first five converts whose names arememorialized in a document dated May 31 1586 [12] appearto have broken with tradition by taking godfathers who werenot Jesuitmissionaries but wealthy seamerchants instead (seeSupplementary File 37)This extraordinary behaviormay alsoindicate a strong secular attitude within the LPK families

Taken together the above data suggest a consistentlymore secular culture for the families of the LPK lineagecompared to their more orthodox Kudtari neighbors over aspan of centuries Nevertheless both groups appear to haveintermarried during this time and maintained their status as

social elites in the same GSB priestly community for over amillennium

4 Discussion

Wehave used statistical analysis and cultural data to constructa meta-narrative for the migration history of an endogamousLPK Brahmin lineage largely comprising individuals of therare type LPKSTR This geographically restricted haplotypewithin the R1a1 Y-HG is found primarily in Eastern Per-siaWestern Afghanistan along an ancient migration corri-dor extending from greater Khurasan to the west coast ofthe Indian subcontinent Multiple methodologies includingcentroid analysis ldquofellow travellerrdquo analysis and TMRCAsegmentation were implemented to examine the migrationhistory of the LPK patrilineage As L657 individuals inthe Indian subcontinent rarely carry the LPKSTR signature(in early 2014 the largest public database did not containa single example) it is possible that LPKSTR expansionin Afghanistan is linked to the late introgression of SakaParthian (Pahlava) or Hephthalite (Huna) tribes in a post-Iron-Age time frame Such a hypothesis might also explainthe high incidence of LPKSTR we found in some tribes ofeastern Transcaucasia The largest of these tribes the Avarsis believed to have migrated there from Khurasan in the 6thcentury CE [49] Perhaps this hypothesis can be reexaminedwhen STR data for additional northern Persian communitiesis published

Supplementary File S26 suggests that at least three differ-ent (large) pools of the R1a1 Y-HG entered the subcontinentseparately over a span of at least fourmillenniaThe ancestorsof orthodox Brahmins may have entered the subcontinent inthe Iron Age while LPK individuals may entered with themost recent of those introgressions (first millennium CE)


We supported a putative migration history for these lineageswithin archeological horizons backed by radiocarbon dating(Kura-Araxes Quetta Ware Harappan OCP-CH and WP-BRW) We further constructed a possible cultural context forLPK migration based on toponymic iconographic numis-matic epigraphic architectural and ancient literary data

The earliest recognizable cultural zone along the LPKSTRmigration path formed by the end of the fourth millenniumBCE with the advent of wheeled transport and a rapid influxof agricultural and metallurgical communities from Tran-scaucasia (the so-called Kura-Araxes migrations Figure 3[26 27]) The archeological context for this cultural zone isthe horizon comprising Tureng Tepe TepeHissar AltyndepeMundigak Shahr-i-Sokhta andMehrgarh archeological sitesfrom about 3000 BCE onward This zone is associatedwith Burnished Grey Ware Quetta Ware and Faiz Moham-mad Grey Ware and partially overlaps the Damb-Sadaatarcheological horizon [50] The mid-3rd millennium BCEHaraxvaiti urban complex (including Mundigak and Shahr-i-Sokhta near present-day Kandahar and Seistan) rivaledMesopotamian city-states in both size and sophistication [5152] Southern reaches of this horizon included pre-Harappansettlements in Cholistan and Sothi-Siswal If domesticatedspecies were brought by thesemigrant agricultural communi-ties as is likely founder effects may help explain the curiousobservation that house mice from Transcaucasia and NorthIndia are closely related by isozyme typing [53]

We show data in support of the proposition that coastalSindh and Sorath (Gujarat) in Harappan times formed adistinct cultural corridor for Elamo-Persian caprid iconog-raphy a proto-Iranic motif (Figure 6) After the collapse ofthe Harappan civilization a large influx of settlers enteringthe subcontinent through Gujarat is memorialized in theBRWarcheological tradition (Figure 7) Some of these settlers(marked by the so-called WP-BRW and ldquochannel-spoutedbowlrdquo artifacts) may have been Iranic tribes Mitra andVaruna toponyms are more prevalent near Gujarat and maydate back to this period of settlement (Supplementary FileS41) suggesting a geographic entry point for Vedic traditionsthat has not previously been suggested Moreover due to itsgeographic location the Bolan Pass may have served as animportant conduit for such migrations

Some scholars have suggested that the Rig Veda was com-posed on the banks of a river inHaraxvaiti province in south-ern Afghanistan (Persian Harahvati Sanskrit Sarasvati pos-sibly the Helmand or Arghandab) though their conclusionis considered controversial [8 9] The origin of the nameldquoRigrdquo is unclear It is noteworthy that the Kandahar region isadjacent to a large (nowmostly desert) domain known locallyas Rigestan (stan = place) The oldest ldquofamily booksrdquo of theRig Veda fall into two categories one group emphasizing textterms relating to the sacrificial tradition (Indra Agni andSoma books 2 3 4 and 6) and the other group emphasizingthe sun gods Mitra Surya and Vishnu (books 5 and 7)suggesting the apposition of at least two major distinct tribalaffiliations in the early Vedic community (SupplementaryFile S21) Our findings are consistent with the speculativepossibility thatmainstream orthodox Brahmins (Indra Agniand Soma) and LPK lineages (Mitra) correspond to these

two coextant traditions within the ancient Sarasvata tribe inHarahvaiti

It is notable that Rig Vedic Book 7 is ascribed to thesage Vasishta whose affiliation with the mainly Iranic sun-worshipping Mitra-Varuna tradition is nevertheless codifiedin the Hindu gotra system and became especially popular inthe medieval assimilation of sun-worshipping Iranic peoples(Huna Maitraka and Gurjar) in Gujarat and the Westernstates [8] LPK Brahmins follow the Kaundinya gotra of theVasishta group Moreover Brahmins of the Vasishta gotraoccur in significantly higher numbers along the westernside of the subcontinent (Supplementary File S42) Takentogether these observations support the possibility of Mitra-Varuna antecedents for LPKSTR LPK families

In early medieval times LPKSTR LPK families probablylived in Saurashtra Gujarat at a time of great social mobil-ity assimilation of Iranic tribes and devotional synthesesbetween Mitra sun worship and Pashupata Shaivism (suchas Zun worship) (Supplementary Files S27ndashS30) One mayspeculate therefore that LPKSTR LPK families were exposedto several distinct religious traditions in different locationsover the centuries a particularly significant matter for anypriestly lineage and possibly related to their pragmatic moresecular outlook

Within the GSB community members of the oldestSmartha (Shaivite) sect have frequently accused the Madhva(worshippers of Vishnu a sun god) of ldquoallowing westernBrahmins into the foldrdquo Based on our work there may besome truth to their assertion It is worth noting that the eventsin question would be about 1300 years old which speaks tothe astounding longevity of community memory

One salient weakness of our work (as with most studiesof ancient culture) is that only tentative cultural associationsmay be made with qualitative data types especially forperiods preceding historical times On the other hand wehave here attempted to bring to the current analysis a cross-disciplinary approach rarely employed by other investigatorsof Y-HG Some of the methodological innovations we haveemployed in this study may be of general use to futureinvestigators of long genealogies

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors wish to thank the following people who helpedmake this project successful Layla Mascarenhas Ryan Mas-carenhas and Fatima Noronha for helping with archivaldocuments and recruitment of individual participants tothe study Vinayak Angle for many helpful tutorials aboutdeity worship and related matters and Pedro Carmo Costafor invaluable advice about Kudtari families and societyAuthors thank all the participants of Khatri Sikhs as partof Asian Indian Diabetic Heart StudySikh Diabetes Study


for their contribution in this study This work was sup-ported by grants funded by the National Institute of Health(NIH) R01DK082766 (NIDDK) NOT-HG-11-009 fundedby NHGRI and a VPR Bridge Grant from University ofOklahoma Health Sciences Center

References

[1] P A Underhill G D Poznik S Rootsi et al ldquoThe phylogeneticand geographic structure of Y-chromosome haplogroup R1ardquoEuropean Journal of Human Genetics vol 23 no 1 pp 124ndash1312015

[2] H Pamjav T Feher E Nemeth and Z Padar ldquoBrief communi-cation New Y-chromosome binary markers improve phyloge-netic resolution within haplogroup R1a1rdquoTheAmerican Journalof Physical Anthropology vol 149 no 4 pp 611ndash615 2012

[3] U Singh A History of Ancient and Early Medieval India Fromthe Stone Age to the 12th Century Pearson Education Press NewYork NY USA 2009

[4] T P Mahadevan ldquoThe Rsi index of the Vedic Anukramanisystem and the Pravara lists toward a pre-history of theBrahmansrdquo Electronic Journal of Vedic Studies vol 18 no 2 pp1ndash140 2011

[5] R Thapar Ancient Indian Social History Some InterpretationsOrient Blackswan Publications 1978

[6] SDattaMigrant Brahmanas inNorthern IndiaTheir SettlementandGeneral Impact Motilal Banarsidass Publishers NewDelhiIndia 1989

[7] U Singh Kings Brahmanas and Temples in Orissa An Epi-graphic Sudy AD 300-1147 Munshiram Manoharlal PublishersNew Delhi India 1994

[8] R ThaparThe Penguin History of Early India From the Originsto AD 1300 Penguin Books London UK 2002

[9] R Kochhar The Vedic People Their History and GeographyOrient Longman Hyderabad India 2000

[10] MWitzel ldquoEarly Indian historymdashlinguistic and textual param-etersrdquo in The Indo-Aryans of Ancient South Asia LanguageMaterial Culture and Ethnicity G Erdosy Ed p 106 DeGruyter Berlin Germany 1995

[11] M Witzel ldquoAutochthonous aryans The evidence from oldIndian and Iranian textsrdquo Electronic Journal of Vedic Studies vol7 no 3 pp 1ndash115 2001

[12] J B A Gracias ldquoOs primeiros cristaos em Salseterdquo inOOrientePortugues vol 6 pp 306ndash337 1934

[13] P N Xavier Bosquejo Historico das Communidades TipografiaRangel Bastora Goa India 1907

[14] B H Baden-Powell ldquoThe villages of goa in the early sixteenthcenturyrdquo Journal of the Royal Asiatic Society vol 32 no 02 pp261ndash291 1900

[15] R Ballard ldquoMigration and kinship the differential effects ofmarriage rules on the processes of Punjabimigration to Britainrdquoin South Asians Overseas Migration and Ethnicity C Clarke CPeach and S Vertovec Eds Cambridge University Press 1990

[16] C C Lamberg-Karlovsky ldquoArchaeology and language TheIndo-Iraniansrdquo Current Anthropology vol 43 no 1 pp 63ndash882002

[17] V Grugni V Battaglia B H Kashani et al ldquoAncient migratoryevents in the Middle East new clues from the Y-chromosomevariation of modern Iraniansrdquo PLoS ONE vol 7 no 7 ArticleID e41252 2012

[18] S Sengupta L A Zhivotovsky R King et al ldquoPolarity andtemporality of high-resolution Y-chromosome distributions inIndia identify both indigenous and exogenous expansions andreveal minor genetic influence of Central Asian pastoralistsrdquoTheAmerican Journal of HumanGenetics vol 78 no 2 pp 202ndash221 2006

[19] K K Abu-Amero A Hellani A M Gonzalez J M Larruga VM Cabrera and P AUnderhill ldquoSaudi ArabianY-chromosomediversity and its relationship with nearby regionsrdquo BMC Genet-ics vol 10 article 59 2009

[20] T Ghosh D Kalpana S Mukerjee et al ldquoGenetic diversity of 17Y-short tandem repeats in Indian populationrdquo Forensic ScienceInternational Genetics vol 5 no 4 pp 363ndash367 2011

[21] H Lacau A Bukhari T Gayden et al ldquoY-STR profiling in twoAfghanistan populationsrdquo LegalMedicine vol 13 no 2 pp 103ndash108 2011

[22] H Lacau T Gayden M Regueiro et al ldquoAfghanistan froma Y-chromosome perspectiverdquo European Journal of HumanGenetics vol 20 no 10 pp 1063ndash1070 2012

[23] M Haber D E Platt M Ashrafian Bonab et al ldquoAfghanistanrsquosethnic groups share a y-chromosomal heritage structured byhistorical eventsrdquo PLoS ONE vol 7 no 3 Article ID e342882012

[24] O Balanovsky K Dibirova A Dybo et al ldquoParallel evolution ofgenes and languages in the Caucasus regionrdquoMolecular Biologyand Evolution vol 28 no 10 pp 2905ndash2920 2011

[25] I Nasidze E Y S Ling D Quinque et al ldquoMitochondrial DNAandY-chromosome variation in the caucasusrdquoAnnals ofHumanGenetics vol 68 no 3 pp 205ndash221 2004

[26] P L Kohl ldquoOrigins homelands and migrations situating thekura-araxes early ttranscaucasian ldquoCulturerdquo within the historyof bronze age rurasiardquo Tel Aviv vol 36 no 2 pp 241ndash265 2009

[27] P L Kohl The Making of Bronze Age Eurasia CambridgeUniversity Press Cambridge UK 2007

[28] H D Sankalia ldquoNew light on the Indo-Iranian or WesternAsiatic relations between 1700 BC-1200 BCrdquo Artibus Asiaevol 26 no 3-4 pp 312ndash332 1963

[29] P Pissurlencar ldquoTombo das Rendas que SuaMagestade tem nasterras de Salceterdquo Boletim do Instituto Vasco da Gama vol 68pp 19ndash79 1952

[30] F X VazMonumenta Goana Eclesiastica Tipografia BragancaNova Goa Portugal 1925

[31] W Henkelman ldquoThe achaemenid heartland an archeological-historical perspectiverdquo in A Companion to the Archeology of theAncient Near East D Potts Ed p 933 Blackwell PublishingSussex UK 2012

[32] B YadavA Raina andTDDogra ldquoHaplotype diversity of 17Y-chromosomal STRs in Saraswat Brahmin Community of NorthIndiardquo Forensic Science International Genetics vol 5 no 3 ppe63ndashe70 2011

[33] B Yadav A Raina and T D Dogra ldquoGenetic polymorphismsfor 17 Y-chromosomal STR haplotypes in Jammu and KashmirSaraswat Brahmin populationrdquo LegalMedicine vol 12 no 5 pp249ndash255 2010

[34] D K Sanghera S K Nath L Ortega et al ldquoTCF7L2 poly-morphisms are associated with type 2 diabetes in Khatri Sikhsfrom North India genetic variation affects lipid levelsrdquo Annalsof Human Genetics vol 72 no 4 pp 499ndash509 2008

[35] P Malekandathil Maritime India Trade Religion and Polity inthe Indian Ocean Primus Books New Delhi India 2010

[36] T De Souza Essays in Goan History Concept Publishing 1989


[37] F Sousa Oriente Conquistado a Jesu Christo Officina deValentim da Costa Deslandes Lisboa Portugal 1710

[38] M Regueiro A M Cadenas T Gayden P A Underhill and RJ Herrera ldquoIran tricontinental nexus for Y-chromosomedrivenmigrationrdquo Human Heredity vol 61 no 3 pp 132ndash143 2006

[39] G Bucellati and M Kelly-Bucellati ldquoUrkesh and the questionof the Hurrian homelandrdquo Bulletin of the Georgian NationalAcademy of Sciences vol 175 no 2 pp 142ndash150 2007

[40] J A Greppin and I M Diakonoff ldquoSome effects of thehurro-urartian people and their languages upon the earliestArmeniansrdquo Journal of the American Oriental Society vol 111no 4 pp 720ndash730 1991

[41] G AlgazeThe Uruk World System The Dynamics of Expansionof EarlyMesopotamianCivilization University of ChicagoPressChicago Ill USA 2005

[42] L B Kirtcho ldquoThe earliest wheeled transport in SouthwesternCentral Asia new finds fromAltyn-DeperdquoArchaeology Ethnol-ogy and Anthropology of Eurasia vol 37 no 1 pp 25ndash33 2009

[43] J R McIntosh The Ancient Indus Valley New PerspectivesABC-CLIO 2007

[44] G L Possehl The Indus Civilization A Contemporary Perspec-tive Alta Mira Press 2002

[45] J M Kenoyer ldquoMaster of animals and animal masters in theiconography of the Indus traditionrdquo in The Master of Animalsin Old World Iconography D B Counts and B Arnold EdsArcheolingua Foundation Budapest Hungary 2010

[46] P Yule Metalwork of the Bronze Age in India Oscar BeckMunich Germany 1985

[47] P Yule ldquoThe copper hoards of the Indian subcontinentrdquoJahrbuch des Romisch-Germanischen Zentralmuseums Mainzvol 36 pp 193ndash275 1992

[48] P J P CostaMedicos que Concluiram o Curso na EscolaMedico-Cirurgica de Goa Desde a Sua Fundacao Ate o Ano Lectivo de1955-1956 Tipografia Rangel Bastora India 1957

[49] S P TolstovAncient Khwaresm IzdanieMGUMoscow Russia1948

[50] VMMasson and I V SarianidiCentral Asia Turkmenia beforethe Achaemenids Thames amp Hudson London UK 1972

[51] M Tosi S M Shahmirzadi and M Joyenda ldquoThe bronze agein Iran and Afghanistanrdquo in History of Civilizations of CentralAsia A H Dani and V M Masson Eds vol 1 chapter 9 pp191ndash224 1992

[52] S Ratnagar ldquoA bronze age frontier problems of interpretationrdquoin Recording the Progress of Indian History S Jafri Ed pp 297ndash324 Primus Books Delhi India 2012

[53] A N Milishnikov L A Lavrenchenko and V S LebedevldquoOrigin of the housemice (superspecies complexMusmusculussensu lato) from the Transcaucasia region a new look atdispersal routes and evolutionrdquo Russian Journal of Genetics vol40 no 9 pp 1011ndash1026 2004

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Anatomy Research International

PeptidesInternational Journal of


Hindawi Publishing Corporation httpwwwhindawicom

International Journal of

Volume 2014

Zoology


Molecular Biology International

GenomicsInternational Journal of


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


BioinformaticsAdvances in

Marine BiologyJournal of



Signal TransductionJournal of


BioMed Research International

Evolutionary BiologyInternational Journal of



Biochemistry Research International

ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Genetics Research International


Advances in

Virolog y

Hindawi Publishing Corporationhttpwwwhindawicom

Nucleic AcidsJournal of

Volume 2014

Stem CellsInternational



Enzyme Research



Microbiology


proximity to a culturally related group situated thousandsof miles away than to any of the dozens of geographicallyproximate communities that might traditionally serve ascomparators Thus in assessing genetic proximity withinhighly differentiated Brahmin milieu the rational choice ofselected comparators for TMRCA computation requires theuse of historical data [8]

Ancient tradition places Saraswat Brahmins with theSarasvata tribe on the banks of the Rig Vedic Sarasvati Riverwhose exact geographic location is not certain but is nowbelieved by some scholars to be the Helmand-Arghandabbasin in Afghanistan [8 9] though this conclusion is stillcontroversialThe region in questionwas settled prior to 4000BCE and its ancient pre-Indo-European namewasHaraxvaiti(later HarahuvatiSarasvati) [10 11] In first millennium BCEpost-Achaemenid Sanskrit literature places Saraswat Brah-mins in the northern Punjab and a substantial Saraswat com-munity remains there to this day The other large Saraswatgroup in South Asia is the Goud Saraswat Brahmin (GSB)community whose arrival in Goarsquos Salcete province duringearly medieval times is memorialized in local oral historiesand legends variously claiming western (Gujarat-Rajasthan)central (Kannauj) or eastern (Tirhut) proximal origins It islikely that todayrsquos Goan GSB community is the product ofmultiple migration events between the 7th and 11th centuriesCE each comprising families of claimed or actual tribalkinship with the original Sarasvata tribe

Within the GSB Brahmin community a high incidence ofLPKSTR was anecdotally observed in male representatives ofPai (surname) families of a single gotra (exogamous lineageKaundinya) who are known founder families in a towncommunidade (Lotli) GSB are a subset of all Saraswatswho are a subset of all Brahmins Table 1(a) describesthese identifiers These so-called Lotli Pai Kaundinya (LPK)families may have retained distinct family practices despitecenturies of intermarriage with other GSB families of neigh-boring towns In this work we explore the hypothesis thatunlike other Brahmins in the community the LPK clanentered Goa via a western migration route from easternPersia and Afghanistan quite distinct from the traditionalnarrative for ingress of Vedic peoples into the subcontinentvia a northern route through Gandhara and the KhyberPass into the PunjabMadhyadesa in the first millenniumBCE and subsequently southward from there in medievaltimes We compare genetic and cultural characteristics ofLPK families with another clan within the same endogamousGSB community with whom they have intermarried Oraland written records of a unique form of town governanceknown as the ldquocommunidaderdquo confirm each clanrsquos identityand privilege in their respective towns for centuries [12ndash14] Samples of Khatri (mercantile Bolan Pass) and SaraswatBrahmin (priestly Khyber Pass) communities are used asgenetic comparators in this study because their cultural andmigration histories represent the two alternatives we wish tocompare

Amongmodernmigrants social success and family prac-tices prior to migration predict relative social success aftermigration [15] Such observations point to portable family-borne attitudes and capabilities including for example

business and educational skills or the pragmatism requiredfor cultural adaptationThe matter has implications for otherdisciplines in the study of ancient history for example theldquopots-are-not-peoplerdquo maxim in archeology [16]

2 Materials and Methods

21 Samples and Subjects The collection under informedconsent of 35 male Saraswat Brahmin and 29 male KhatriDNA samples (listed in Table 2) has been described inprevious publications [32ndash34] 16 GSB samples (15 listedin Table 1(b) one non-R1a1 GSB individual (Q1a3 Y-haplogroup) was omitted from the analyses) were collectedas follows volunteers registered for cheek swab DNA anal-ysis directly with Genebase Inc (Vancouver BC) signedinformed consent forms and followed requisite proceduresas required by the Genebase Testing Laboratory Each volun-teer ordered Y-HG subclade and 67-STR tests Volunteerswere reimbursed for the cost of testing but received noother compensation After testing was complete volunteersprovided investigators with access to their results on theGenebase database Permission to use data for research andpublication was provided in writing

22 Ethics Statement Studies were approved by each ofthe participating organizationsrsquo institutional committees incharge of oversight for ethical conduct in research Thecollection and research use of Khatri DNA samples wereapproved by the Institutional Review Board of the Universityof OklahomaHealth Sciences Center OK (IRB number 2911)and the Ethical Committee of Hero DMC Medical Collegeand Heart Institute Ludhiana Punjab India The GSB studywas approved by the Ethical Committee of the MayflowerOrganization for Research and Education Sunnyvale CAThe collection and research use of Saraswat Brahmin DNAsamples were approved by the Institute Ethical ClearanceBody of the All India Institute of Medical Sciences NewDelhi

23 DNA Analysis Typical tests included the standard 67-STR Y-backbone and subclade testsThe details of these testscan be found on the Genebase website (httpwwwgenebasecom) Assignment of Y-HG and subclade in this study usedthe tests for the markers and single nucleotide polymor-phisms (SNPs) listed in Table S1 (see SupplementaryMaterialavailable online at httpdxdoiorg1011552015651415) Themicrosatellite short tandem repeats (STR) tested are listed inTable S2

24 Validation of Lineages Volunteers were identified andvalidated using oral communications with town social net-works communidade records electoral rolls church bap-tismal and marriage records going back 150ndash400 years forChristian subjects and the Ramnathi temple community forHindu participants Brahmin status for Christian participantswas confirmed from marriage records showing endogamywith other Brahmin families over several generations (Sup-plementary files S3 S4 S5 and S6)



(a)









(b)






(a)


(b)


(b) Continued







PBR11

LKR-Vg101LKR-Vg102

LKR-Vg51

KH111213

CK1112

LKR-VgU3

LKR-VgU2LKR-VgU1

LKR-Vg151LKR-Vg121

LKR-Vg71

LKR-Vg83

LKR-Vg82

150012009006003000

1286 BCE375

695

697

792 960

1425

1425

1800CE

840

1765

1880

1635

1530

Location of towns


L3422+ L657minus

L3422+ L657+

LKR-Vg81












50

25

50

25

50

25

50

25

50

25

GSB

KHT GLF

SCT

TMRCA branch-points

Sam

ple (

)


BCTlowast

TJKlowast

PSBlowast

ZABlowast

PSTlowast

GDHlowast







ndash900 CE)




V8

KHT

KHT

GSB

GSB

ZAB

ZAB

PST

PST

GDH

GDH

00914 00011

00227

00000 00000 00000

0000000010

00923

00001

0000100007

00526

03433 03356

00314

04931

TJK

00021

00983

02076

01776





3 Results












Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)


Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG








istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



(a)









(b)






(a)


(b)


(b) Continued







PBR11

LKR-Vg101LKR-Vg102

LKR-Vg51

KH111213

CK1112

LKR-VgU3

LKR-VgU2LKR-VgU1

LKR-Vg151LKR-Vg121

LKR-Vg71

LKR-Vg83

LKR-Vg82

150012009006003000

1286 BCE375

695

697

792 960

1425

1425

1800CE

840

1765

1880

1635

1530

Location of towns


L3422+ L657minus

L3422+ L657+

LKR-Vg81












50

25

50

25

50

25

50

25

50

25

GSB

KHT GLF

SCT

TMRCA branch-points

Sam

ple (

)


BCTlowast

TJKlowast

PSBlowast

ZABlowast

PSTlowast

GDHlowast







ndash900 CE)




V8

KHT

KHT

GSB

GSB

ZAB

ZAB

PST

PST

GDH

GDH

00914 00011

00227

00000 00000 00000

0000000010

00923

00001

0000100007

00526

03433 03356

00314

04931

TJK

00021

00983

02076

01776





3 Results












Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)


Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG








istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



(a)


(b)


(b) Continued







PBR11

LKR-Vg101LKR-Vg102

LKR-Vg51

KH111213

CK1112

LKR-VgU3

LKR-VgU2LKR-VgU1

LKR-Vg151LKR-Vg121

LKR-Vg71

LKR-Vg83

LKR-Vg82

150012009006003000

1286 BCE375

695

697

792 960

1425

1425

1800CE

840

1765

1880

1635

1530

Location of towns


L3422+ L657minus

L3422+ L657+

LKR-Vg81












50

25

50

25

50

25

50

25

50

25

GSB

KHT GLF

SCT

TMRCA branch-points

Sam

ple (

)


BCTlowast

TJKlowast

PSBlowast

ZABlowast

PSTlowast

GDHlowast







ndash900 CE)




V8

KHT

KHT

GSB

GSB

ZAB

ZAB

PST

PST

GDH

GDH

00914 00011

00227

00000 00000 00000

0000000010

00923

00001

0000100007

00526

03433 03356

00314

04931

TJK

00021

00983

02076

01776





3 Results












Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)


Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG








istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


PBR11

LKR-Vg101LKR-Vg102

LKR-Vg51

KH111213

CK1112

LKR-VgU3

LKR-VgU2LKR-VgU1

LKR-Vg151LKR-Vg121

LKR-Vg71

LKR-Vg83

LKR-Vg82

150012009006003000

1286 BCE375

695

697

792 960

1425

1425

1800CE

840

1765

1880

1635

1530

Location of towns


L3422+ L657minus

L3422+ L657+

LKR-Vg81












50

25

50

25

50

25

50

25

50

25

GSB

KHT GLF

SCT

TMRCA branch-points

Sam

ple (

)


BCTlowast

TJKlowast

PSBlowast

ZABlowast

PSTlowast

GDHlowast







ndash900 CE)




V8

KHT

KHT

GSB

GSB

ZAB

ZAB

PST

PST

GDH

GDH

00914 00011

00227

00000 00000 00000

0000000010

00923

00001

0000100007

00526

03433 03356

00314

04931

TJK

00021

00983

02076

01776





3 Results












Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)


Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG








istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


50

25

50

25

50

25

50

25

50

25

GSB

KHT GLF

SCT

TMRCA branch-points

Sam

ple (

)


BCTlowast

TJKlowast

PSBlowast

ZABlowast

PSTlowast

GDHlowast







ndash900 CE)




V8

KHT

KHT

GSB

GSB

ZAB

ZAB

PST

PST

GDH

GDH

00914 00011

00227

00000 00000 00000

0000000010

00923

00001

0000100007

00526

03433 03356

00314

04931

TJK

00021

00983

02076

01776





3 Results












Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)


Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG








istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



3 Results












Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)


Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG








istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


Kura-Araxes sites

K

Pool A (P15 M357 M410 M479)


Pool C (M17 M69 M479)

G

S

S

PSBKhatri

GSBG








istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


istan

ce fr

om so

uthe

rn ro

ute

6

4

2

0


Nor

ther

n ro

ute

Southern route


N clusterS cluster










A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


A1

B1

A2

A3C2

C1

B2

D1

D2

D3South Asia

Persian Gulf

Transcaucasia


ID ISOGG 2011

A1

A2A3B1

B2

C1C2

D1D2

D3

R1a1a1lowast


R1a1a1g1bR2alowast

R2a1aJ2a4lowast

J2a4blowast

J2a4b1


28

61

33

17

32

65

40

51

43

60

377

327

263

440

517

398

238

384

435

463

349

461

593

220

207

424

680

309

235

218

160

134

105

146

36

175

128

166

114

108


R-Z283+ M458+


J-L26+ M67minus


J-L26+ M67+ M92+

0

1

2

3

4

5

6

7

8

Relat

ive i

ncid

ence

of h

aplo

grou

p (A

U)





Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


Khyber

THARdesert

50 100

5

4

3

2

1


Bolan

Indigenous

Caprid

Hybrid

0 5

4

3

1

2

Region12

2

3

3

4

4 500144 00409

03986 00008

00722

01607 02096

000000001000001








Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


Khyber

Bolan

1900 1200

1600

yx

z

32

30

28

26

24

22

20

18

16

14

12

10

8

32

30

28

26

24

22

20

18

16

14

12

10

80 400

(km)














Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



Van

gor e

xtin

ctio

n (

)

60

40

20

R2 = 0955

(a)

04

02

02

01Rate

per

van

gor

Ratio ofdoctors

to priests

Conversionrate

LPK

70

35

Extinction ()

p lt 0037

p lt 0021

p lt 0008

(1560ndash1586)

KKLPK KK LPK KK

(b)







4 Discussion















Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology













Acknowledgments




References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



References






























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


























Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

Genetic and Cultural Reconstruction of the Migration of an Ancient ...

Documents

Transcript of Genetic and Cultural Reconstruction of the Migration of an Ancient ...