Cross-Species Hybridization and the Origin of 55

www.pnas.org/cgi/doi/10.1073/pnas.1817453116 (1) Asia South in Pakistan and East; Peninsula, Middle Arabian the extends the in Africa; clonal–sexual Iran cultivation North and palm Iraq, mixed in Egypt date plan- a to of Morocco or via range from farms propagated traditional habitats The small are arid system. they well-irrigated hot and where or in East tations primarily oases Middle grow desert the palms of including Date regions arid Africa. impor- in North most the crop and fruit-bearing family tant Palmae) (formerly Arecaceae the efforts, in breeding crop change. climate future may of for past light dynam- in the approaches especially in the new occurred Identifying suggest has and Neolithic. expansion also the expansion range since species range crop evolution how cultural species human of of ics nature insights provides the mechanisms into these remain understanding environments multiple but to unclear, adapt to species crop environ- mechanisms allow evolutionary that new thousands and to genetic hundred precise to few The a generations. adapting of of span the ranges, within cultures their and dramatically ments expanded have species rapidly domesticated and many origin, of centers D expansion range introgression palm. date cultivated the of with history hybridization that theophrasti suggest results P. Our Africa. date North of to arrival of late palm a survey supports a which record present archaeobotanical We the in East. Middle the in palm in polymor- fixed single-nucleotide are that for (SNPs) segregating phisms are population this in vari- to African traced North of be genome can the eties of 18% as much as proportions gene that ancestry indicates incorporates of Analysis that populations. these model between population flow a and regions, gressed and African palm North between date subdivision reduced supported admixture, of is tests by hybridization Introgressive Mediterranean. Eastern Eastern Middle dactylifera from popu- its components African with of ancestry North mixed five the has that lation and indicate varieties results Our palm relatives. date closest of genomes the investigated resequencing of We history regions. evolutionary and these the East between Middle shown differentiation the have studies genetic in genetic population widely previous cultivated and Africa, is North species This East. 2018) dle 12, October domesticated review were for that (received regions 2018 ( 6, December palm approved and WI, Date Madison, Wisconsin-Madison, of University D Doebley, F. le John pour by Recherche Edited de Institut (DIADE), Plants of Development and Adaptation Diversity Greece; Crete, Iraq; Chania, 10071, 73100 Baghdad Chania, of Institute Agronomic Mediterranean Emirates; Arab Switzerland; United Zurich, Al-Ain, 8008 University, Emirates Arab United Biotechnology, Emirates; a Fournaraki Christini Perrakis Andreas Flowers M. Jonathan palms North date of African origin the and hybridization Cross-species etrfrGnmc n ytm ilg,NwYr nvriyAuDaiRsac nttt,NwYr nvriyAuDai b hb,Uie Arab United Dhabi, Abu Dhabi, Abu University York New Institute, Research Dhabi Abu University York New Biology, Systems and Genomics for Center h aepl ( palm date The esu pce ntewrd rmgorpial restricted geographically suc- From world. evolutionarily the most in the species cessful among are crops omesticated b etrfrGnmc n ytm ilg,NwYr nvriy e ok Y10003; NY York, New University, York New Biology, Systems and Genomics for Center and | hei dactylifera Phoenix archaeobotany hei theophrasti Phoenix a fcnrliprac ntediversification the in importance central of was .theophrasti P. e g hei dactylifera Phoenix nttt fAcaooy nvriyCleeLno,Lno CH0Y ntdKndm and Kingdom; United 0PY, WC1H London London, College University Archaeology, of Institute yveFerrand Sylvie , .theophrasti P. e e n ihe .Purugganan D. Michael and , eierna ln osrainUi,ItrainlCnr o dacdMdtraenArnmcSuis(CIHEAM) Studies Agronomic Mediterranean Advanced for Centre International Unit, Conservation Plant Mediterranean a,b hldM Hazzouri M. Khaled , | .dactylifera P. domestication .theophrasti P. hrn fhpoye nintro- in haplotypes of sharing , ∼ . samjrfutco farid of crop fruit major a is L.) ,0 g nteNa rMid- or Near the in ago y 7,000 idrltv nei othe to endemic relative wild a , n ag ecnaeo loci of percentage large a and a usmS .Khierallah M. S. Hussam , . sadocosspecies dioecious a is L.) | n t idrltvsby relatives wild its and rpwl relative wild crop n betfo date from absent and Phoenix a,c uilGros-Balthazard Muriel , a,b,2 remains | P. f f,1 aePl eerhUi,Cleeo giutr,Uiest fBaghdad, of University Agriculture, of College Unit, Research Palm Date d oinQ Fuller Q. Dorian , oteAainNeolithic date remains that Arabian Kuwait seed the and from Dhabi, to Abu comes Island, Dalma exploitation from of excavated evidence oldest The 7). (5, African North from dactylifera differentiated P. East- genetically are Middle palms that date indicates ern (6–8) sequences whole-genome and 1 (Fig. 1073/pnas.1817453116/-/DCSupplemental at online information supporting contains article This the 2 in deposited been ( 1 have Repository no. Digital paper Dryad (accession the this in database deposited in Archive reported Read sequences Sequence The deposition: Data BY-NC-ND) (CC 4.0 NoDerivatives License distributedunder is article access open This Submission. Direct PNAS a is article This interest. of conflict paper. no the declare wrote authors M.D.P. The and J.M.F. M.G.-B., and K.M.H., data; J.M.F., K.M.H., analyzed research; J.M.F., D.Q.F. performed and C.F. research; Z.M., and F.A., designed D.Q.F., M.D.P. H.S.M.K., S.F., and A.P., C.F., K.K., D.Q.F., J.M.F., contributions: Author (6). East Middle wild the the in of palms date population cultivated relictual of a progenitor represent may which wild Oman, identified palm in date study domesticated recent of A progenitor 16). wild (15, the palm, be date to sugar unlikely but The (14). East Near sylvestris the in yBP between morphological 2,000 place for took larger- forms Evidence of domesticated selection 13). suggests fruited 12, stones date domesti- 9, archaeological palm (1, in date change region of this origin Arabian in of Upper center cation the ancient and an Crescent supports Fertile Gulf the of edge eastern in the later somewhat appears at cultivation at palm Mesopotamia date of Evidence eateto ytmtcadEouinr oay nvriyo Zurich, of University Botany, Evolutionary and Systematic of Department owo orsodnesol eadesd mi:[email protected]. Email: addressed. be should correspondence whom To 2018. 15, August Deceased fdt palm. date diversification of the with associated Mediterranean—is Eastern palms date between hybridization and of that whole-genome sample show to use regions large relatives We arid a wild of Africa. from crop North data fruit and sequence major East a Middle is the and of world the in species ( palm Date Significance aeplsaeoeo h aletdmsiae recrops. tree domesticated earliest the of one are palms Date ∼ ,0–,0 B 1) h al vdneo utvto at cultivation of evidence early The (12). yBP 5,700–5,500 hei theophrasti Phoenix A vlpeet 49 otele,France Montpellier, 34394 eveloppement, .Pplto eei nlssuigmcoaelts(2–6) microsatellites using analysis genetic Population ). ´ L)Rx. stesse pce of species sister the is Roxb., (L.) a iiMo Ziyi , ihapsil yrdzn nEytadteSudan the and Egypt in zone hybrid possible a with , hei dactylifera Phoenix c hlf etrfrGntcEgneigand Engineering Genetic for Center Khalifa ∼ g ,0–,0 B 1,1)adi h Levant the in and 11) (10, yBP 6,700–6,000 rdrqeAberlenc Frederique , a y osatn Koutroumpa Konstantina , ∼ ree— pce nei othe to endemic species Grueter—a . y y ,0 eoepeet(B)(9). (yBP) present before y 7,000 . Cetv omn Attribution-NonCommercial- Commons Creative y y https://doi.org/10.5061/dryad.tm40gd8 . soeo h lette crop tree oldest the of one is L.) h Unit PRJNA495685 it eRcece(UMR) Recherche de Mixte e ´ www.pnas.org/lookup/suppl/doi:10. .dactylifera P. NSLts Articles Latest PNAS .dactylifera P. h .dactylifera P. , .SPdt aebeen have data SNP ). ∼ ,0 B and yBP 5,000 d y , populations y n its and Phoenix 1 15), (1, | ). f8 of 1 y

EVOLUTION A (5). These observations are inconsistent with a bottleneck associated with founder-effect colonization in North Africa. Evidence 40 of an unknown ancestry component in North African cultivated P. canariensis date palm further implies a more complex history in this region 30 (6). The genetic distinctiveness of North African date palms, their absence from early archaeological sites in North Africa, and 20 elevated levels of nucleotide diversity perpetuate the enigma of Latitude P. dactylifera the origins of date palms in the western part of their range. 10 (date palm) P. atlantica To examine the origin of North African date palms, we resequenced a large sample of cultivars from Morocco to Pakistan 0 P. sylvestris and ﬁve wild Phoenix relatives that occur either peripatrically P. reclinata or allopatrically to cultivated date palm (Fig. 1A). Here we −10 present evidence that the North African population is the prod- −30 0 30 60 uct of introgressive hybridization between cultivated date palm and the Cretan date palm Phoenix theophrasti Greuter, a species 38 endemic to Crete and the Eastern Mediterranean. We demon- 37 strate that introgression has been central in shaping patterns of 36 diversity genome-wide, which supports introgression as being an Latitude P. theophrasti 35 important factor that shaped the domestication history of date palm. The growing list of examples of interspeciﬁc hybridiza- 20 25 30 35 tion associated with domestication suggests that hybridization is Longitude a common evolutionary genetic mechanism for the adaptation of B both annual and perennial crops. Reclinata_DP18P. reclinata Results Canariensis_DP6A 100 100 P. canariensis Plant Samples, Nucleotide Diversity, and the Phylogeny of Phoenix Canariensis_93121 Species. We resequenced the genomes of 71 cultivated date palm Theophrasti_B5 varieties and multiple genomes from each of its 5 closest wild 100 100 P. theophrasti

Theophrasti_D1 relatives in the genus Phoenix to address questions about the origin of the domesticated date palm (P. dactylifera). Our sample Sylvestris_RIV_2256_PL_F 100100 100 100 P. sylvestris set included common varieties of dates from its traditional range Sylvestris_RIV_2248_PL_F of cultivation (7) and 2–18 individuals from wild relatives that

100 include P. sylvestris; P. theophrasti; Phoenix atlantica A. Chev.; 100 ZahidiZahidi 100100 Phoenix canariensis hort. ex Chabaud; and a single member of KhisabKhisab a putative outgroup species from sub-Saharan Africa, Phoenix 100100 P. dactylifera Deglet_noorDeglet Noor reclinata Jacq. (SI Appendix, Table S1).

26 Samples were sequenced to moderate to deep coverage (5– 26 Medjool Medjool 56×, mean = 22×; SI Appendix, Table S2) using 2 ×100 bp 65 65 Atlantica_CAP1_POPMAL1 paired-end Illumina sequencing. The close relationships of the 93 P. atlantica

93 Atlantica_CAP50_BOA1 wild relatives to date palm allowed us to align >96% of short reads of each sequenced genome (SI Appendix, Table S2) to 0.0080.008 the draft assembly of the date palm genome (28). We identi- Fig. 1. The geographic distribution and phylogeny of cultivated date palm ﬁed 14,402,469 single-nucleotide polymorphisms (SNPs) across (P. dactylifera) and its wild relatives. (A) Approximate geographic distribu- Phoenix species which we used in population and phylogenomic tion of Phoenix species included in this study (27). (B) Maximum-likelihood analysis. phylogeny of Phoenix species based on a subset of SNPs from whole-genome Reconstruction of the phylogeny of date palm and its wild rela- sequencing data. Support values are the percentage of RAxML bootstrap tives using the whole-genome SNP data supported the close rela- replicates that support each node. The tree is midpoint rooted. tionship of P. sylvestris, P. theophrasti, P. atlantica, and domesticated date palm P. dactylifera, (the “date palm group”) (1). P. canariensis and P. reclinata are more distant relatives (Fig. 1), Alternate origin hypotheses propose a domestication center consistent with previous analyses (1, 15) and the Cape Verde of date palm in North Africa (1, 17, 18). These hypotheses Islands endemic, P. atlantica, is a member of a well-supported are poorly supported by the archaeobotanical record, however, P. dactylifera clade consistent with the samples in our analysis as evidence of date palm cultivation appears ∼3,000 y later in being feral date palms (1, 6). Within the date palm group, P. the records of this region compared with the Middle East. P. sylvestris is the sister species of date palm, and P. theophrasti is a dactylifera remains in the Nile Valley are found as early as the more diverged species (Fig. 1). Estimates of nucleotide diversity predynastic period, more than 5,000 yBP (19, 20), but cultivation in the Phoenix wild relatives suggest P. canariensis (π = 0.0117) > may not have begun until the Middle Kingdom, ∼4,000 yBP (14, P. sylvestris (π = 0.0105) > P. theophrasti (π = 0.0072), while 21). In Libya, cultivation began ∼2,800–2,400 yBP (18, 22). Date estimates in the two populations of date palm indicate higher palm is unknown from elsewhere in the Saharan Maghreb and diversity in North Africa compared with the Middle East (SI the sub-Saharan Sahel until much later (23–25). Appendix, Table S3) as previously reported (6, 7). The archaeobotanical record therefore is consistent with a late arrival of date palm to North Africa and suggests a model where Population Clustering Suggests That North African Date Palms Are date palm was domesticated in the Near or Middle East and Admixed. To identify genetic clusters in P. dactylifera and its wild later expanded to the African continent (26). Paradoxically, how- relatives, we conducted model-based clustering with the program ever, nucleotide diversity in North African date palm is at least STRUCTURE (29). STRUCTURE runs with the independent 20% higher than that in the Middle East (6, 7) and diversity at allele-frequency model produced clusters at K = 2 to K = 4 that microsatellite loci is comparable between the two populations showed evidence of mixed ancestry of North African date palm.

2 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1817453116 Flowers et al. lwr tal. et Flowers and relative; date wild domesticated a of is populations P3 respectively; African palms, North popula- and the assuming Eastern tests by relatives of wild configuration combinations its tion all and gene palm conducted to date We attributed among be species. can between that flow polymorphism derived shared of Eastern or Middle ABBA–BABA, performed to between admixture hybridization dactylifera of interspecific prod- the of tests indeed are uct palms explicit of date caution African conducted suggestive North with whether is establish therefore treated analysis be We STRUCTURE must (30). interpretation the Prod- but Are in hybridization, Palms palms Date date Hybridization. African of Interspecific North of Indicate ucts Admixture of Tests a or relative wild this and date cultivated palm between admixed being population African North African ( North relative of wild S6 ancestry this mixed and the samples supported also that palm to restricted date analyses analysis correspond- The allele-frequency samples. the all independent included to and similar correlated qualitatively ing relative wild results single a yielded from those species and samples palm date runs to limited STRUCTURE Pairwise (31). analysis) (“hierarchical” pairs 30). known to (29, subject to values is K attributable which higher be at model, oversplitting frequencies could correlated or the analysis Africa of our artifacts North in in population variation any of pop- to African higher, trace North and not ( the 5 does in = which appears K ulation component At ancestry However, third above. a as S5 results Table similar tively and S1 ( Fig. model allele-frequency correlated the the repeated with therefore analysis We clusters. of number meaningful the underestimate additional any to samples assign not sources. did and values 2 (Fig. K genomes palm sample date African Eastern North with shared the were of 5–15% Approximately model. allele-frequency independent the using 4 = K to 2 = relative, K for wild diagram a and palm date 2. Fig. North Africa IAppendix SI ial,w odce e faaye etitdt species to restricted analyses of set a conducted we Finally, may model independent the with STRUCTURE Running .SRCUEaaye r hrfr ossetwt the with consistent therefore are analyses STRUCTURE ). ot fia aepl a ie netyfo ideEastern Middle from ancestry mixed has palm date African North n h idCea species, Cretan wild the and i.S1 Fig. , Middle East .theophrasti P. D P,2P,) hr 1adP r Middle are P2 and P1 where (P1,P2,P3,O), .Ti ol ersn nadtoa source additional an represent could This ). .K=2t rdcdqualita- produced 4 = K to 2 = K ). .theophrasti P. Individuals Individuals D .theophrasti P. IAppendix SI et 3,3) ots o excesses for test to 33), (32, tests IAppendix SI n h eane ihMiddle with remainder the and vdneo ie ancestry mixed of Evidence hw saSRCUE(29) STRUCTURE a is Shown . theophrasti .theophrasti P. al S4 Table , i.S n Table and S2 Fig. , lk population. -like .reclinata P. .theophrasti P.

atlantica IAppendix SI sylvestris is,we First, . .Higher ). canariensis sthe is K=3 K=2 K=4 and P. , eaieaddt am hsrsl srbs otecoc of choice the to robust is result replacing This as wild palm. outgroup, this date between and hybridization relative interspecific suggests and palm between alleles derived ( and relative wild the as species. a wild a and and (P2) (P1) palm palm between date flow Eastern gene date indicate Middle tests African negative while North (P3), relative between wild flow gene positive cates A (O). outgroup rec Africa; North pointing Af, edge Abbreviations: 15.7%. migration The populations. from among relatedness in variance ( migration one on based D 3. Fig. CD B A statistics eosreapositive a observe We , .reclinata P. IAppendix SI .theophrasti P. ditr ewe ot fia aepl and palm date African North between Admixture TreeMix ± E ( SE. ; m syl )eet hs oesepan9.%ad9.%o the of 99.9% and 98.6% explain models These event. 1) = B ihboksz f300SP o ( for SNPs 3,000 of size block with , al S7 Table , ) .sylvestris P. oNrhArcndt amhsamxuewih of weight mixture a has palm date African North to f 3 statistics D .reclinata P. .8 (SE 0.58) = D .theophrasti P. ,wihidctsa xeso shared of excess an indicates which ), and ; D ttsi o h etwith test the for statistic ± ttsi nti ofiuainindi- configuration this in statistic E ( SE. the can C , .theophrasti P. with and , ± .canariensis P. n ot fia date African North and 0.02; D NSLts Articles Latest PNAS .canariensis P. aiu-ieiodtree Maximum-likelihood ) C Z eo( zero ) . 72;Fg3 Fig 37.24; = .theophrasti. P. E ideEast; Middle ME, ; .theophrasti P. m )ad( and 0) = produces | f8 of 3 ( D A A ) )

EVOLUTION a similar test outcome (Fig. 3A and SI Appendix, Table S7). proportions (36). A mixture weight of 15.7% on the migra- Tests where samples from the Maghreb countries are replaced tion edge from P. theophrasti to North African date palm with with cultivars from Egypt also show an excess of derived allele m = 1 (Fig. 3D) is comparable to our estimate from the f4-ratio sharing with P. theophrasti although the degree of sharing is approach. These results suggest that 5–18% of the North African less than in cultivars sampled from elsewhere in North Africa date palm genome is derived from the Cretan date palm with (SI Appendix, Table S7). D tests including P. sylvestris and P. varieties from countries west of Egypt sharing greater ancestry canariensis are negative and positive, respectively, possibly sup- with this wild relative. porting additional gene flow with date palm and other wild relatives. However, these are minor contributions compared Introgressed Regions in the North African Date Palm Genome. We with those of P. theophrasti and may represent hybridization in identified introgressed genomic segments in the North African anthropogenic contexts (Fig. 3A and SI Appendix, Table S7). population using a combination of approaches. We traced indi- Second, we tested for admixture between date palm and its vidual alleles in this population to their most likely source pop- wild relatives with f3 tests (34, 35). The f3 statistic tests whether a ulation by characterizing variant sites that are fixed for alternate population is the product of admixture between two reference, or alleles in the Middle Eastern population and P. theophrasti. This source, populations with a significantly negative f3 statistic sup- analysis yielded 1,556,435 nucleotide fixations, of which 90.2% porting admixture and all other outcomes being uninformative are polymorphic in North African P. dactylifera. Of the sites seg- about the admixture history (34). The test with North African regating for a Middle Eastern-like and P. theophrasti-like allele in date palm as the test population and P. theophrasti and Mid- this population, the latter is typically the minor allele with most dle Eastern date palm as reference populations was negative observed at low to moderate frequency. A total of 1,252,999 of and suggests that North African date palms have a mixed ances- 1,404,273, or 89.2% of these polymorphic sites in North Africa, try from these two populations [f3(North Africa; Middle East, have a P. theophrasti-like allele frequency of 30% or less. Only P. theophrasti) = −0.15; |Z | = 19.8; Fig. 3B and SI Appendix, 9.8% of sites that are fixed between Middle Eastern date palm Table S8). All other f3 tests either did not differ significantly and P. theophrasti are fixed for the Middle Eastern-like allele from zero or were positive (Fig. 3B and SI Appendix, Table S8). in North Africa and none are fixed for the P. theophrasti-like The negative f3(North Africa; Middle East, P. theophrasti) test variant. supports admixture between date palm and a wild species, and To characterize regions with introgressed haplotypes that have also suggests that the Middle Eastern population or a Middle risen to moderate-to-high frequency, we estimated the intro- Eastern-like population is a source of alleles segregating in North gression fraction (f d) for the population configuration D(Middle Africa. East, North Africa, P. theophrasti, P. reclinata) (above and SI We modeled the population history of date palm and its rel- Appendix, SI Materials and Methods and ref. 37). A positive f d atives with the maximum-likelihood method TreeMix (36). This implies introgression from P. theophrasti into North African date modeling approach infers the population tree based on the palm. We identified blocks of two or more consecutive 5-kb inter- covariance in allele frequencies among populations and incorpo- vals with an f d in the upper 10th percentile of the genome-wide rates admixture through the addition of migration edges between distribution of this statistic. Applying this criterion, we identi- populations that are poor fits to the strict tree model. To deter- fied 1,281 introgressed segments of 10 kb or larger, which totaled mine whether gene flow between date palm and P. theophrasti is 24.6 Mb of the draft assembly. The median length of tracts with supported by the data, we first inferred the maximum-likelihood outlier f d was 15 kb and the largest region was 105 kb on scaf- tree without migration (Fig. 3C). A model rooted with P. recli- fold NW 008246809.1. This approach yields an underestimate of nata yielded the same relationships among Phoenix species as tract lengths owing, in part, to the current fragmented state of the the maximum-likelihood phylogeny reported above (Fig. 1B) and draft assembly. Moreover, the signature of introgression is not explained 98.6% of the variance in relatedness between popula- limited to these outlier regions as evidenced by an elevated intro- tions. Addition of a single migration event to the model predicted gression fraction and altered patterns of diversity throughout gene flow between P. theophrasti and the lineage leading to the much of the genome (below). Fig. 4A shows f d in sliding windows North African population of date palm (Fig. 3D) and increased across the longest genome scaffolds. Fig. 4B shows an example 3- the percentage of variance in relatedness explained to 99.9%. Mb region with both elevated f d and Tajima’s D in North Africa This migration edge was stable in independent runs of TreeMix and the incidence of high-frequency theophrasti-like alleles in this with both different species included in the model and with dif- region. A gene tree reconstructed from phased genotypes con- ferent block sizes to account for linkage disequilibrium (LD) (SI firmed the existence of shared haplotypes between P. theophrasti Appendix, Table S9). Together, the admixture tests and TreeMix and North African date palm cultivars in this region (Fig. 4C). results suggest that the ancestry of North African date palm can We conducted a phylogenetic analysis of chloroplast (cpDNA) be traced to Middle Eastern date palm and P. theophrasti. and mitochondrial (mtDNA) genomes to examine patterns of introgression in organellar genomes and evaluate whether cul- Mixed Ancestry of North African Date Palm. We estimated the per- tivated date palm P. dactylifera or P. theophrasti served as the centage of ancestry of North African date palms derived from maternal progenitor in the interspecific cross. Both neighbor- P. theophrasti and Middle Eastern date palm using f4 statistics joining and maximum-likelihood analyses of cpDNA genome (35) and TreeMix mixture weights (36). First, the ratio of appro- sequences largely supported the species phylogeny recovered priate f4 statistics can provide an estimate of the proportion of from the whole-genome analysis (SI Appendix, Fig. S3). P. North African ancestry that is derived from the Middle East theophrasti samples formed a distinct clade indicating that and P. theophrasti by assuming the phylogenetic relationships in cpDNA from this species is not introgressed in North Africa. Fig. 1B. We estimated the percentage of ancestry (α) derived Middle Eastern and North African date palms appear as dis- from Middle Eastern date palm at 82% (SE ± 0.01; SI Appendix, tinct and well-supported lineages, although the Middle Eastern Table S10) and the P. theophrasti fraction as 1 – α = 18%. haplotype is also found in some North African samples, a result Repeating the estimation of α by replacing samples from the consistent with a previous report (5). While the North African Maghreb with Egyptian samples as the test population yielded clade is distinct and well supported, it shares a most recent com- a smaller P. theophrasti and larger Middle Eastern component in mon ancestor with P. sylvestris. Single P. theophrasti (Golk¨ oy,¨ these samples (α = 95%) (SE ± 0.007; SI Appendix, Table S10). Turkey) and P. sylvestris (Faisalabad, Pakistan) samples that Second, TreeMix models with gene flow provide mixture weights were admixed with date palm in the STRUCTURE analysis estimates for each migration edge, which approximate ancestry based on nuclear genotypes (Fig. 2) possess P. dactylifera cpDNA

4 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1817453116 Flowers et al. lwr tal. et Flowers to contributed has introgression whether determine To gression. the 5 Fig. in test, diversity signed-rank nucleotide ( Higher population African palm. North date African diversity genetic North of Africa. in patterns North genome-wide in impacted Diversity has Genetic gression on Introgression of Impact interspecific the of sam- direction contributor. that palm the indicate date in and African asymmetry cross North an from suggest absence ples their and lotypes the ( of genome mtDNA distinctness the The appar- of analysis were the relationships in ent phylogenetic Similar samples. hybrid ( haplotypes atlantica within nested (a are hybrids a istan) and probable Crete two Almyros, of from each sample from ancestry haplotypes African Eastern/North Single Middle in (7). mixed East a with Middle (Aseel) the variety Pakistani from haplotype single The the sequences. phased in scaffold of posi- the at origin interval of 5-kb kb a fitted in 1,100 haplotypes were phased tion ( on Curves based (LOESS). SNP. tree regression rule a polynomial majority 50% represents local using circle data Each the to Africa. North in regating NW scaffold for Bottom palm date African ( North assembly. draft the in scaffolds ( fraction gression 4. Fig. BC A .theophrasti P. hw h rqec of frequency the shows nrgesdrgosi ot fia aepl.( palm. date African North in regions Introgressed samples. IAppendix SI NrhArc ld ern h hrdhpoyei rma from is haplotype shared the bearing clade Africa /North f d n5k ooelpigwnosfo h 0longest 30 the from windows nonoverlapping 5-kb in ) .dactylifera P. B .theophrasti P. .theophrasti P. n e.7 ol eepandb intro- by explained be could 7) ref. and i.S3 Fig. , P B B .theophrasti P. ooe ice ersn h population the represent circles Colored . ouaingntcsmaysaitc of statistics summary genetic Population ) < .sylvestris P. 2.2 ldsa r altpsfo two from haplotypes are as clades ossetwt hi being their with consistent ) × 0261. n5k windows. 5-kb in 008246512.1 10 pN n tN hap- mtDNA and cpDNA a h aenl(pollen) paternal the was apefo aslbd Pak- Faisalabad, from sample − lk lee mi et seg- text) (main alleles -like 16 IAppendix SI w-aldWilcoxon two-tailed ; C Neighbor-joining ) .theophrasti P. A h intro- The ) i.S4 Fig. , Intro- B P. ). , ot fia ouain seeae nteesm regions same these in elevated ( is fraction populations introgression African high North a with and Appendix regions palm in date pronounced African North between by ranked intervals 5 Fig. test, signed-rank from diverged ( less ulation is ( population theophrasti diverged African P. moderately pop- North are regional the palm two However, date the that of observe ulations we Using subdivision, LD. population and differentiation population elevated as introgression Appendix by part in least Tajima’s at driven be to appears ( East Middle Tajima’s effect level, An genome Tajima’s Africa. on North in introgression an diversity of Africa, elevated for North insuffi- account be in to may cient alone diversity hybridization higher that suggests show that observation also average on sion ( Africa S5 North Fig. in diversity nucleotide of genome levels the higher ordered rank on we Africa, based North in diversity elevated ersn h afdcydsac t2. bad3. bfrMdl Eastern Middle for curves kb 30.8 decay populations. and African on kb North 20.5 Symbols and at population. distance ( each half-decay populations. the for palm represent separately date calculated African North LD and of Eastern Middle in East; sity Middle (M.E., Africa; populations North of N.A., pairs between intervals ( nonoverlapping populations. palm date ern 5. Fig. impacted profoundly has population. palm this in date diversity mixed African genome with the North palm that of suggest date observations ancestry African These North population. of distant a admixture recent with is distance sistent half-decay 5 at the maximum (Fig. East its Middle population half the this reaches in Africa rate North slower in a LD at larger decays LD therefore (7), and rium has diversity population African nucleotide North with higher the although population However, a decay. show LD ( equilibrium, size to At population population effective LD. African larger admixture North of the expect evidence we relative, wild to due partially least at is from introgression populations these between observed ( ABC IAppendix SI h mato nrgeso sas paeti esrsof measures in apparent also is introgression of impact The fteepten r h euto eetitorsinfo a from introgression recent of result the are patterns these If .Hwvr ein ihltl rn vdneo introgres- of evidence no or little with regions However, ). atrso eei iest nNrhArcnadMdl East- Middle and African North in diversity genetic of Patterns i.S5 Fig. , S5 Fig. , F D f st d sms rnucdi ein ihhighest with regions in pronounced most is i.S5 Fig. , 0.403; = n on hthgl nrgesdrgosshow regions introgressed highly that found and P ( F the < ). st .Moreover, ). .theophrasti P. , .6)ta steMdl atr pop- Eastern Middle the is than 0.369) = 2.2 .theophrasti P. ,tu ugsigta ouainstructure population that suggesting thus ), f d A eosreta h euto in reduction the that observe we , P × D .We ecompare we When ). A < opo itiuin of distributions Boxplot ) shge nNrhArc hni the in than Africa North in higher is 10 − 2.2 D 16 F .( ). sas paet ttewhole- the At apparent. also is , N st nNrhAfrica. North in × IAppendix SI e B ewe ideEsenand Eastern Middle between hudhv atrrt of rate faster a have should ) itiuino uloiediver- nucleotide of Distribution ) 10 –16 NSLts Articles Latest PNAS w-aldWilcoxon two-tailed , .theophrasti P. ∼ 0k,aptencon- pattern a kb, 20 F st F al S3 Table , st samaueof measure a as ∼ F cosgenomic across 0k,wiein while kb, 30 st N F ae n5-kb on based e IAppendix SI st tequilib- at 0.085). = smost is C .This ). | Decay ) f d f8 of 5 C ( F SI SI ). st ,

EVOLUTION Private Alleles in Date Palm Populations. Our analysis raises ques- colonization that was not supported by previous genome-wide tions about the possibility of an unsampled population that analysis showing elevated genetic diversity in North Africa (6, may have been a source of ancestry in the North African 7). Evidence of hybridization, while appearing to explain much population. We examined the distribution of private polymor- of the elevated diversity in this population, also does not appear phisms among populations and Phoenix wild relatives. When to fully account for differences in nucleotide diversity between measured as a fraction of total SNPs within each group, populations. These observations suggest that a simple expan- we find that the North African population does not have a sion plus hybridization model may be insufficient to account for large class of private polymorphisms (5.7%) relative to Mid- patterns in the genomic data. More complex expansion models dle Eastern date palm (13.4%) (SI Appendix, Table S11). such as those including postexpansion inbreeding (7) or addi- This suggests any additional unsampled source of variation in tional bottlenecking in the Middle East could account for our the North African population is unlikely to be a divergent observations. lineage. Alternatives to the expansion model include those that propose an additional source of variation in addition to P. Archaeobotanical Records Are Consistent with a Late Appearance theophrasti and Middle Eastern date palm in the North African of Date Palm in North Africa. We assembled archaeobotanical gene pool. For example, it is possible that the archaeobotani- records from the published literature to address questions about cal record provides an incomplete picture of Phoenix in North the historical distribution of date palm (SI Appendix, SI Text Africa and that a wild population of P. dactylifera similar to the and Dataset S1). Inspection of these records indicates that date relictual population recently discovered on the Arabian Penin- palm appears in the archaeobotanical record of North Africa sula (6) once existed on the African continent. This population later than in the Near and Middle East (Dataset S1 and SI either could have been domesticated independently or was not Appendix, Figs. S6–S8). A single pair of date stones recovered domesticated, but served as an additional source of variation for from Takarkori, Southwest Libya, might be as old as 9,000 an extant cultivated population with P. theophrasti and Middle yBP (38), but is here rejected as plausibly recent and intrusive. Eastern date palm ancestry. A distinct haplotype found at high The earliest records on the African continent date to predy- frequency in the organellar genomes of the North African popu- nastic Egypt (ca. 5,500 yBP) and Sudan (ca. 4,000 yBP). In lation (SI Appendix, Figs. S3 and S4) (45) provides some support the Maghreb, the oldest reliable records are from Libya (ca. for such a lineage. However, we note that the relatively small 3,000 yBP). West of Libya, there is one record of date palm at numbers of private alleles in North Africa (5.7% of SNPs in Volubilis, Morocco dating to the Classical Period (ca. 2,000 yBP), North Africa vs. 13.4% in the Middle East) suggest that if such but otherwise no archaeobotanical records before the Islamic a population exists, it is unlikely to be a divergent lineage or has Period (ca. 700 yBP) when seed remains appear at the medieval not contributed significantly to the genomic constitution of North sites of Igˆılˆız, Volubilis, and Sijilmasa, Morocco and Essouk and African date palms. Gao Saney, Mali. These observations are consistent with the view Phoenix species are known to hybridize and produce viable that date palm may have only recently expanded to the African offspring (16, 27, 46, 47) and either natural hybridization or continent. horticultural practices could account for the introgression of interspecific alleles into P. dactylifera. For example, the practice Discussion of hand pollination of date palm was widely known through- We have presented evidence that the North African date palm out the ancient world (48) and existing practices of fertiliza- population has a mixed ancestry that can be traced to hybridization with interspecific pollen (49) and seedling propagation tion between date palm and a wild relative P. theophrasti, the (50) suggest a possible mechanism for hybrid origins. Alter- Cretan date palm. P. theophrasti is a distinct species (Figs. 1B natively, putative natural hybrids between P. theophrasti and and 3C and ref. 15) with a present-day distribution that is date palm have been reported (40) and are also apparent in restricted primarily to geographically isolated locations on the our analysis. Although P. theophrasti samples appear as a dis- island of Crete, the East Aegean Islands, and the Anatolian coast tinct cluster in our STRUCTURE results (Fig. 2), individu- of Turkey (Fig. 1A and refs. 39–42). Evidence that the North als with some P. dactylifera ancestry include a sample from a African population has mixed ancestry between this Aegean reported hybrid population in Golk¨ oy,¨ Turkey (40); two sam- Sea endemic and cultivated date palm suggests that P. dactylif- ples from Almyros, Crete; and two samples from Epidaurus, era on the African continent has a hybrid origin and that P. Peloponesse. These samples may represent instances of crop- theophrasti may be the unknown source of variation in North to-wild gene flow (51) or be relicts from an earlier hybridiza- Africa identified by Gros-Balthazard et al. (6). tion event. The geographic context of hybridization between P. theophrasti Crop species frequently retain the capacity to hybridize with and P. dactylifera is obscured by the fact that the present-day their wild relatives and hybrid genotypes are sources of nov- distribution of P. theophrasti does not overlap with the tradi- elty, superior quality, and adaptive traits. Evidence of intro- tional range of date palm cultivation (Fig. 1). One model is that gressive hybridization in annual crops such as rice (52) and hybridization occurred in the Levant or elsewhere in the East- maize (53) and perennial tree crops including citrus (54), ern Mediterranean where the ranges of the two species may have almonds (55), and apricots (56) supports a role for hybridiza- once overlapped. Phoenix stones identified as P. theophrasti have tion in the domestication process. The adaptive significance been found at a site in northern Israel from ∼9,000 yBP, suggest- of hybridization and the traits subject to selection are often ing that this species had a wider distribution in the past and may unknown except in exceptional cases (53). However, it is becom- have been exploited (38, 43). P. dactylifera was also distributed ing increasingly clear that introgressive hybridization is an historically in the Levant including the Jordan River valley and important source of diversity (53) that frequently accompa- the Dead Sea until the Middle Ages when the Judean date palm nies expansion of domesticated species such as apples (57) population went extinct (44). Whether the two species histor- and grapevine (58). Discovery that a regional population of ically occurred sympatrically in the Eastern Mediterranean or domesticated date palm is also the product of hybridization elsewhere is unclear. helps resolve long-standing questions about the origins of date The late appearance of date palm in the archaeobotanical palm and indicates that introgression may be an important record of North Africa suggests a model where Middle East- factor that dramatically alters the genome of domesticated ern date palm expanded its range to the African continent. species, thus providing novel diversity for adaptation during This model predicts a bottleneck associated with founder event domestication.

6 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1817453116 Flowers et al. 17. 16. 15. 14. 13. 12. 11. 10. lwr tal. et Flowers Few remains. crop (C14) dated on dates directly dates are radiocarbon (AMS) these associated spectrometry of by mass cases accelerator direct some or evidence in of archaeological most “ass.”), and reports, as artifactal all associated (indicated on includes based scholar It dated Google articles. are a journal which archaeo- and recent on (70), additional “literature 1981–2004” for of plants search search cultivated database of a regional remains 69), previous logical and on 68 based refs. chapters (e.g., monograph reviews finding on based R reports, the with conducted was analysis (67). Statistical a Language (37). and Programing (66) al. Statistical vcftools et with Martin gen- SNPs from from next or script of (65) analysis (ANGSD) with data alignments sequencing statistics eration sample summary from introgression directly estimated and were Population Popstats. in were implemented proportions ancestry of from Estimates obtained (64). Popstats with conducted was were mixtures using and data splits the population con- to of was fitted model data A genomic (29). STRUCTURE population with of ducted clustering Model-based (62) phangorn (63). packages randomized ape R of and and combination (61) (RAxML) a likelihood using maximum axelerated genomes mtDNA and cpDNA, nuclear, draft palm following accordingly. called date filtered were and the Informa- SNPs (60) (59). Biotechnology to protocols aligner standard for aligned Burrows–Wheeler Center were with National (NCBI) Data the tion at sample. available per one- (28) with lane assembly Sequencer one 2500 HiSeq to Illumina half an on conducted was ornamen- sequencing and ( facilities, gardens research tal collections, germplasm populations, wild Methods and in provided and is Materials methods and SI materials of description detailed A Methods and Materials 9. 8. 7. 6. 5. 4. 3. 2. 1. Mne 18)Oii fdt amcliainadpoaaini fia Historical Africa: in propagation and cultivation palm date A of India: Origin Northwestern (1981) P in Munier sylvestris P. and dactylifera Phoenix (2013) al. et C, Newton based (Arecaceae) Phoenix genus the in delimitation Species of (2010) al. domestication et the JC, in Pintaud trends Comparative attenuated: in and cultivation Long garden (2018) palm DQ date Fuller of history early and Beginnings (2012) M Tengberg (2012) E Weiss M, Hopf D, el’Oueil. Zohary Tell Iraq: lowland in sites archaeological from remains Plant (1991) R Neef (1981) S Lloyd MA, Mustafa F, Safar BehM(03 rhebtncleiec o al aecnupini the in consumption date early for evidence Archaeobotanical two (2003) supports cultivars M palm date Beech of survey genome-wide A (2015) insights al. yields et palms LS, date Mathew of re-sequencing genome Whole (2015) al. et a K, reveals Hazzouri Oman in palms date wild of discovery The (2017) dactylif- al. et (Phoenix M, palm Gros-Balthazard date the of structure Genetic (2015) an al. of et evidence S, genetic Zehdi-Azouzi provide markers DNA Male-specific (2013) al. et E, char- Cherif Development, (2012) M Latifian H, Mohammadi Reza M, Bahar H, Arabnejad palm date in relationships genetic and Coryphoideae). polymorphism Molecular (Palmae: (2012) al. L. et S, Phoenix Zehdi of monograph A (1998) SC Barrow rhebtnclrcrso aepl eegtee rmpublished from gathered were palm date of records Archaeobotanical for separately SNPs of subset a on conducted was analysis Phylogenomic oe ntepicplArcnpl rwn areas. growing palm African principal the on notes relationships. complex their of 267–286. glimpse pp Denmark), Aarhus, Press, Univ (Aarhus JI Davis dactylifera). AS, (Phoenix palm Monocotyledons Barfod the date in the Evolution of and identity Phylogeny, the Diversity, on emphasis with markers, SSR on fruits. tree the and Europe, East. Asia, Middle the South-west in Plants Domesticated of Basin Mediterranean Spread and Origin 321–329. pp Paris), 1985 Civilisations, de Travaux Oueili: Baghdad). Heritage, and Antiquities of for 11–32. Organization pp Center Emirates), Arab Emirates United (The Dhabi, Abu Research Research, and and Studies Studies Strategic Strategic for Center Emirates The ed Gulf. Arabian dactylifera. Phoenix in subpopulations major crop. tree fruit a of diversification into Africa. and East Middle the in centers Biol involving history domestication complex western and eastern between differentiation strong a reveals populations. World Old the paternal in of era) tracing the allow and palm. arrest date recombination in lineages a system, chromosome palm XY date in analysis germplasm for L.). markers dactylifera microsatellite (Phoenix of use and acterization markers. microsatellite 263. nuclear of utility The L.): dactylifera (Phoenix 53:513–575. Phoenix 27:2211–2218.e8. e itArchaeobot Hist Veg idrltvsfo aiu ore,icuigcletosfrom collections including sources, various from relatives wild IAppendix SI n Bot Ann TreeMix h aePl–rmTaiinlRsuc oGenWealth Green to Resource Traditional Palm–From Date The rdEnviron Arid J Ofr nvPes e ok,4hEd. 4th York), New Press, Univ (Oxford 116:101–112. e Phytol New TreeMix c Hort Sci itr egt n ihthe with and weights mixture Wrso 95,e utJ ( JL Huot ed 1985], of [Works al S1 Table , rey eotie aite of varieties obtained we Briefly, . Phoenix 86:139–147. 134:150–156. 27:165–176. 3) ditr et with tests Admixture (36). 197:409–415. Eridu oetcto fPat nteOdWrd The World: Old the in Plants of Domestication .Woegnm 2 Whole-genome ). a Commun Nat tnstesle,bttoeavailable those but themselves, stones Palms Mnsr fClueadIfrain State Information, and Culture of (Ministry G3 57:37–50. 5:1429–1438. Fruits 6:8824. dtosRcece u les sur Recherches Editions ´ 36:437–450. d eegO eesnG, Petersen O, Seberg eds , f 4 × rtoapoc as approach -ratio f 3 0 paired-end 100 and c Hort Sci .dactylifera P. IAppendix SI D statistics e Bull Kew 148:255– Curr , , 38. 37. 36. 35. 34. 33. 32. 31. 30. 29. 28. 27. 26. 25. 24. 23. 22. 21. 20. 19. 18. etakmmeso h ...lbrtr n esc oiaAdl for Abdala Molina Jessica and laboratory manuscript. M.D.P. the the improve of helped members who thank Ross- reviewers We Jeffrey anonymous thank two Abdul also We and Institute, samples. Ibarra Research providing Palm for (Date Pakistan) University, Markhand Palm Latif S. Date Ghulam University Youssef State and Tunisia), (Arizona Collection), Kebili, Thirkhill Agriculture, Deborah Kadri of (NYUAD), El Ministry Idaghdour Nabila Agri- Dates, Syria), of Biosaline of Development Commission Center for (Technical Energy and Center (Atomic Haider Research (International Zaid Nadia Masmoudi Palm culture), Abdelouahhab Khaled (UAEU)], [Date University UAEU) Visser Emirates (DPRU, Col- Arab J. Medical United of Hendrik (DPRU), Cornell Unit (Weill (Institute Qatar), Malek Naqvi in A. Abbas lege Joel Summar - Pakistan), Syed Sciences, Palme Kingdom), Horticultural Baker le J. United William per Gardens, France), Montpellier, de Ricerche (Kew Plantes des e (Jardin Department Spick Studi Emmanuel Jos (US [Centro Italy], Krueger (CSRP), Littardi R. Sanremo Robert Claudio thank Biol- Agriculture), We Systems of and assistance. Genomics technical for for Center ogy (NYUAD) Dhabi Abu University Dryad York the in ACKNOWLEDGMENTS. deposited been have data (72). SNP Repository Digital (71). Archive our Read in Sequence flagged are these in provided (38), are details context Additional 8,000-yBP a ( database to of antiquity attributed true the median Libya, over a uncarbonized concerns as and are such D.Q.F.) there recent remains, Where of from range. associated expertise taken this of been the from case taken has on the range (based In sam- age literature calculated. the period archaeological been of standard has phase a the median evidence represent a dating to which taken a been from summed has ple, been range have 1-sigma dates the radiocarbon and Calibrated “AMS*.” as indicated are ad hsrsac a uddi atb nNUDIsiuegat as Genome grant, Plant Institute M.D.P.). Foundation (to NYUAD Science Foundation National Family an Zegar US by the the and part Pin- from Program Research in grants Jean-Christophe by funded by as possible was well made research was This work taud. This discussions. helpful Rvr ,e l 21)Crooia nlsso hei Aeaee:Cnrbtosto Contributions (Arecaceae): Phoenix of analysis Carpological (2014) al. et to statistics D, from ABBA-BABA Rivera of mixtures use the and Evaluating (2015) splits CD Jiggins population JW, Davey of SH, Martin Inference (2012) JK Pritchard history. JK, human in Pickrell Indian admixture Reconstructing Ancient (2009) (2012) L al. et Singh N, AL, admixture Patterson Price ancient N, for Patterson Testing K, (2011) Thangaraj M D, Slatkin Reich genome. D, Neandertal Reich the N, of Patterson sequence EY, draft Durand over-interpret A to (2010) not al. conundrum. how et 2 = on R, K tutorial The Green (2017) A al. et (2018) JK, D Janes Falush L, Dorp van using structure DJ, population Lawson of Inference (2000) P Donnelly dactylifera M, Phoenix Stephens palm JK, date Pritchard the of sequence Genome (2013) al. et IS, Al-Mssallem review. A Phoenix: genus the in Hybridization (2013) M Gros-Balthazard (1973) P Munier Sta Van A, Fili AS, Ettahiri M, town Tengberg merchant Islamic MP, early Ruas an at use Plant (2011) DQ Fuller MA, Murray S, Nixon Fewet the from charcoal and fruits Seeds, (2013) L F temples. Buldrini two G, of Bosi deposits elite AM, foundation an Mercuri in from macroremains Fruits Plant (1983) (2008) W I Zeist El-Garf van M, Fadl S, Khodary AG, Fahmy region. Heirakonpolis the of flora predynastic evidence. the The of (1982) review MN A El-Hadidi Libya: in agriculture Ancient (1995) M Veen der Van N eunedt e oti td aebe eoie nthe in deposited been have study this to new data sequence DNA h aooyadeouinr itr ftegenus. the of history evolutionary and taxonomy the loci. introgressed locate data. frequency allele genome-wide 1093. history. population populations. related closely between 710–722. plots. bar ADMIXTURE and STRUCTURE data. genotype multilocus L. Agric tree. particu- argan with the Morocco) of (Anti-Atlas, exploitation Igiliz the 433. of to site reference fortified lar medieval the at research (Mali). Essouk-Tadmakka of archaeobotany The Sahel: Archaeobot African West the in simulations. climate and data Pollen 7:1351–1362. Chad: northern Yoa, Lake at evolution compound. 10:351–354. Egypt. Upper Hierakonpolis, predynastic at cemetery 102–115. pp Egypt), Giza, Science, Report of Faculty Interim An Hierakonpolis: of nastic Palaeobot zn M hn ,BaontP rne 21)Lt ooeepatadclimate and plant Holocene Late (2011) G Krinner P, Braconnot W, Zheng AM, ezine ´ a Commun Nat 25:831–842. aae S1 Dataset 35:85–98. rdZn rhelMonogr Archaeol Zone Arid 20:223–239. 4:2274. ePalmier-Dattier Le Nature n xlddfo h ytei fdt amspread. palm date of synthesis the from excluded and ) etakMr roxadNzrDo nteNew the in Drou Nizar and Arnoux Marc thank We o ilEvol Biol Mol Genetics 461:489–494. lmd(oaia adno aeca Spain), Valencia, of Garden (Botanical Plumed e ´ Phoenix TeDt am Mionuee aoe Paris). Larose, et (Maisonneuve Palm] Date [The 155:945–959. IAppendix SI LSGenet PLoS o ilEvol Biol Mol 32:244–257. a Commun Nat 6:177–190. dHfmnM CioUi Herbarium, Univ (Cairo MA Hoffman ed , tnso ag iefo Takrakori, from size large of stones o Ecol Mol 8:e1002967. , 28:2239–2252. vlJ 21)Archaeobotanical (2011) JP evel IText SI ¨ NSLts Articles Latest PNAS n Bot J Int o inSoc Linn J Bot 9:3258. 26:3594–3602. e itArchaeobot Hist Veg . 4:205–212. Genetics 175:74–122. rhelSci Archaeol J Science mrJFood J Emir h Predy- The 192:1065– lmPast Clim | e Hist Veg 20:419– f8 of 7 Acta 328:

EVOLUTION 39. Boydak M (1985) The distribution of Phoenix theophrasti in the Datça Peninsula, 56. Zhang Q, et al. (2018) The genetic architecture of floral traits in the woody plant Turkey. Biol Conserv 32:129–135. Prunus mume. Nat Commun 9:1702. 40. Boydak M, Barrow S (1995) A new locality for Phoenix in Turkey: Golk¨ oy-B¨ odrum.¨ 57. Cornille A, et al. (2012) New insight into the history of domesticated apple: Secondary Principes 39:117–122. contribution of the European wild apple to the genome of cultivated varieties. PLoS 41. Greuter W (1995) Phoenix theophrasti. The Red Book of Rare and Threatened Plants Genet 8:e1002703. of Greece, eds Phitos D, Strid A, Snogerup S, Greuter W (World Wide Fund for Nature 58. Myles S, et al. (2011) Genetic structure and domestication history of the grape. Proc Greece: K. Michalas S.A., Athens, Greece), pp 412–413. Natl Acad Sci USA 108:3530–3535. 42. Tsakiri M, Kougioumoutzis K, Iatrou G (2016) Contribution to the vascular flora of 59. Li H (2013) Aligning sequence reads, clone sequences and assembly contigs with BWA- Chalki Islands (East Aegean, Greece) and bio-monitoring of a local endemic taxon. MEM. arXiv:1303.3997v1 [q-bio.GN]. Preprint, posted May 26, 2013. Willdenowia 46:175–190. 60. Poplin R, et al. (2017) Scaling accurate genetic variant discovery to tens of thousands 43. Kislev M, Hartmann A, Galili E (2004) Archaeobotanical and archaeoentomological of samples. bioRxiv:10.1101/201178. Preprint, posted July 24, 2018. evidence from a well at Atlit-Yam indicates colder, more humid climate on the Israeli 61. Stamatakis A (2014) RAxML version 8: A tool for phylogenetic analysis and coast during the PPNC period. J Archaeol Sci 31:1301–1310. post-analysis of large phylogenies. Bioinformatics 30:1312–1313. 44. Goor A (1967) The history of the date through the ages in the Holy Land. Econ Bot 62. Schliep KP (2011) phangorn: Phylogenetic analysis in R. Bioinformatics 27:592–593. 21:320–340. 63. Paradis E, Claude J, Strimmer K (2004) APE: Analyses of Phylogenetics and Evolution 45. Zehdi-Azouzi S, et al. (2016) Endemic insular and coastal Tunisian date palm genetic in R language. Bioinformatics 20:289–290. diversity. Genetica 144:181–190. 64. Skoglund P et al. (2015) Genetic evidence for two founding populations of the 46. Wrigley G (1995) Date palm, (Phoenix dactylifera). Evolution of Crop Plants, eds Americas. Nature 525:104–108. Smartt J, Simmonds NW (Longman Scientific and Technical, Harlow, UK), pp 399– 65. Korneliussen TS, Albrechtsen A, Nielsen R (2014) ANGSD: Analysis of next generation 403. sequencing data. BMC Bioinformatics 15:356. 47. Gonzalez-P´ erez´ MA, Sosa PA (2009) Hybridization and introgression between the 66. Danecek P, et al. (2011) The variant call format and VCFtools. Bioinformatics 27:2156– endemic Phoenix canariensis and the introduced P. dactylifera in the Canary Islands. 2158. Open For Sci J 2:78–85. 67. R Core Team (2015) R: A Language and Environment for Statistical Computing (R 48. Zohary D, Spiegel-Roy P (1975) Beginnings of fruit growing in the Old World. Science Foundation for Statistical Computing, Vienna), Version 3.2.2. 187:319–327. 68. Pelling R (2014) Patterns in the archaeobotany of Africa: Developing a database for 49. Popenoe PB (1913) Date Growing in the Old World and the New (West India Gardens, North Africa, the Sahara, and the Sahel. The Archaeology of African Plant Use, eds Altadena, CA). Stevens CJ, Nixon S, Murray MA, Fuller DQ (Left Coast Press, Walnut Creek, CA), pp 50. Johnson DV, Al-Khayri JM, Jain SM (2013) Seedling date palms (Phoenix dactylifera 205–224. L.) as genetic resources. Emir J Food Agric 25:809–830. 69. Fuller DQ (2015) The economic basis of the Oustul splinter state: Cash crops, subsis- 51. Ellstrand NC, Prentice HC, Hancock JF (1999) Gene flow and introgression tence shifts and labour demands in the Post-Meroitic transition. The Kushite World. from domesticated plants into their wild relatives. Annu Rev Ecol Syst 30:539– Beitrage¨ zur Sudanforschung Beiheft 9 [Contributions to Sudan Research Supple- 563. ment]. Proceedings of the 11th International Conference for Meroitic Studies, Vienna, 52. Choi JY, et al. (2017) The rice paradox: Multiple origins but single domestication in 1-4 September 2008, ed Zach MH (Verein der Forderer¨ der Sudanforschung, Vienna), Asian rice. Mol Biol Evol 34:969–979. pp 33–60. 53. Hufford MB, et al. (2013) The genomic signature of crop-wild introgression in maize. 70. Kroll H (2005) Literature on archaeological remains of cultivated plants 1981-2004. PLoS Genet 9:e1003477. Available at archaeobotany.de/database.html. Accessed January 2, 2016. 54. Wu GA, et al. (2014) Sequencing of diverse Mandarin, pummelo and orange genomes 71. Leinonen R, Sugawara H, Shumway M (2010) The Sequence Read Archive. Nucleic reveals complex history of admixture during citrus domestication. Nat Biotech 32: Acids Res 39:D19–D21. 656–662. 72. Flowers JM, et al. (2019) Data from “Cross-species hybridization and the origin of 55. Delplancke M, et al. (2012) Gene flow among wild and domesticated almond species: North African date palms.” Dryad Digital Repository. Available at datadryad.org. Insights from chloroplast and nuclear markers. Evol Appl 5:317–329. Deposited December 26, 2018.

8 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1817453116 Flowers et al.