1 Supplementary Materials for 2 3 4 Genetics and material culture support repeated expansions into 5 Eurasia from a population hub out of Africa 6 7 8 9 Leonardo Vallini1, Giulia Marciani2, Serena Aneli1, Eugenio Bortolini2, 10 Stefano Benazzi2,3, Telmo Pievani1, Luca Pagani1,4 11 12 1 Department of Biology, University of Padova, 13 2 Department of Cultural Heritage, University of Bologna, Italy 14 3 Department of Evolution, Max Planck Institute for Evolutionary Anthropology, 15 Leipzig, 16 4 Institute of Genomics, University of Tartu, Estonia 17 18 19 20 21 Supplementary Materials include: 22 23 - Supplementary Information 24 Section 1: Material Culture Page 2 25 Section 2: Genetic Dataset Page 11 26 Section 3: qpGraph analyses Page 14 27 Section 4: Paleomaps plotting Page 25 28 29 - Supplementary Figures S1 to S7 30 - Supplementary Tables S1 to S4

1

31 Supplementary Section 1: Material Culture 32 33 Many of the DNA samples discussed in this paper come from sites without stratigraphic association with diagnostic archaeological materials. To 34 overcome this issue, we use layers of sites located in proximity of and coeval to DNA sampling sites (Table S1).

35

Broad DNASample/Proxy Techno- chronological Sample/Site name Country Coordinates References complex range (cal kyr BP)

DNASample Vindija M 46.29 N, 16.07 E ~50 *

Leder 2014, Pre-expansion sites in the Ksar_Akil_XXII-XXV IUP 33.91 N, 35.64 E ~45-43 2016

Kuhn 2004; Pre-expansion sites in the Levant Üçagizli_F,Fa,Fb-c,G,H,H1-3,I IUP 35.98 N, 35.96 E ~45-40 Kuhn et al 2009

Boeda, E., Pre-expansion sites in the Levant Umel’Tlel IUP 35.27 N, 38.89 E ~40 Bonilauri, S., 2006

2

Marks, A., Pre-expansion sites in the Levant Boker_Tachtit_layer_4 IUP 30.84 N, 34.78 E ~45-40 Kaufman, D., 1983

DNASample UstIshim IUP* 57.7 N, 71.1 E ~44 *

Zwyns et al. Ust_Ishim Kara_Bom_OH_5_OH6 IUP Russia 50.72 N, 85.57 E >45 2012

Derevianko et al. 2013; Ust_Ishim Tolbor-4_layer_4-5-6 IUP 49.29 N, 102.97 E ~45-35 Zwyns et al. 2019

Zwyns et al. Ust_Ishim Tolbor-16_layer_6 IUP Mongolia 49.23 N, 102.92 E ~45-40 2019

Zwyns, N., 51.44 N, 108.17 Ust_Ishim Kamenka_A IUP Russia ~45-40 Lbova, L. V, E 2019

DNASample Tianyuan IUP* 39.66 N, 115.87 E ~40 *

3

Morgan et al Tianyuan Suindonggou_1 IUP China 38.38 N, 106.51 E <41 2014; Boeda et al. 2013

Li et al. 2019; Tianyuan Suindonggou_2 IUP China 38.38 N, 106.51 E <41 Peng et al.2020

DNASample Kostenki14 UP* Russia 51.23 N, 39.3 E ~37 *

Sinitsyn and Kostenki14 Kostenki_12_Vokov UP Russia 51.39 N, 39.05 E <40 Hoffecker 2006

Hoffecker et Kostenki14 Kostenki_1 UP Russia 51.39 N, 39.05 E <40 al. 2016

DNASample Oase1 IUP-UP* 45.12 N, 21.9 E ~38 *

Tsanova Oase1 Bacho_Kiro_IUP_layer_11 IUP 42.95 N, 25.43 E >45 2009; Hublin et al. 2020

Czech ~41-35 Demidenko Oase1 Ořechov_IV_–_Kabáty IUP ~49.06 N, 16.31 E Republic (underestimate) 2020

4

Czech Richter et al. Oase1 Brno-Bohunice IUP ~49.12 N 16.37 E >45 Republic 2008

Anghelinu Oase1 Românesti-Dumbravita UP Romania 45.49 N, 22.19 E ~45-40 nita 2014; Sitlivy 2012

Anghelinu Oase1 Cosava UP Romania 45.51 N, 22.19 E ~45-40 nita 2014; Sitlivy 2014

Anghelinu Oase1 Tincova UP Romania 45.33 N, 22.9 E ~45-40 nita 2014; Sitlivy 2014

DNASample GoyetQ116-1 UP* 50.45 N, 5.01 E ~35 *

Pirson et al. GoyetQ116-1 Maisières-Canal UP Belgium 50.48 N, 3.98 E <40 2012

Pirson et al. GoyetQ116-1 Spy_Ossiferous_Horizon_2 UP Belgium 50.48 N, 4.67 E <40 2012

5

Dobrovolskay a et al., 2012; DNASample Sunghir UP Russia 56.18 N, 40.50 E ~34 Trinkaus and Buzhilova, 2018

Pitulko et al. DNASample Yana UP Russia 70.72 N, 135.42 E ~31 2017

Khenzykheno va et al., DNASample Mal’ta (MA1) UP Russia 52.9 N, 103.5 E ~24 2019; Lbova, 2019

Tsanova DNASample Bacho_Kiro_IUP_layer_11 IUP Bulgaria 42.95 N, 25.43 E ~45 2009; Hublin et al. 2020

Czech DNASample Zlatý_Kůň S-IUP* 49.37 N, 16.72 E >45 * Republic

Adams 2009; Zlatý Kůň Szeleta S 48.06 N, 20. 37 E ~45-40 Mester 2010

Czech Nejman et al., Zlatý Kůň Pod_Hradem S 49.37 N, 16.72 E ~45-40 Republic 2017

6

Nerudová Czech Zlatý Kůň Moravský_Krumlov_IV S 49.05 N, 16.41 E ~43-42 and Neruda, Republic 2017

Czech Tostevin Zlatý Kůň Stranska_Skala_III-IIIc IUP 49.11 N, 16.40 ~45-40 Republic 2003

Czech Richter et al. Zlatý Kůň Brno-Bohunice IUP ~49.12 N, 16.37 E >45 Republic 2008

36

37 Table S1. Cultural proxies for the ancient genomes analysed in the current work. M: ; S: ; IUP: Initial ; 38 UP: Upper Paleolithic; * indicates an attribution based on nearby coeval sites, which are specified in the subsequent rows. The coordinates of 39 the sites are taken from the reference articles or from the ROCEEH Out of Africa Database (ROAD) (http://www.roceeh.org).

7

40 1.1 IUP-UP definition 41

42 The definition of a techno-complex is based on the association among specific traits identified 43 and recorded in archaeological assemblages (e.g. typology and technology of stone tools). 44 Depending on the degree of magnification used to define a lithic assemblage, each techno- 45 complex could be seen as more/less similar to others. A broad definition of techno-complexes 46 could be compatible with models of population structure/movement, but this requires a high 47 degree of simplification and the loss of finer-grained differences across the examined sites. 48 On the other hand, a strict definition of a techno-complex could focus on the specificities of 49 each assemblage, hampering the possibility of gaining a more general perspective. Strict 50 definitions of lithic assemblages are in fact more compatible with scenarios of independent 51 development 1. In order to develop a model which considers a wide geographical and 52 chronological scale as the one under consideration in this work, it is crucial to describe the 53 techno-complexes choosing the key criteria that are both able to: i) maintain site-specificity; 54 and ii) identify the general tendencies. The question is how the data are collected and 55 compared and what meaning can be given to them concerning the question of the origin of 56 technical innovations. New technologies could be related to demic movements, cultural 57 diffusion/exchange without implying migration, or parallel convergence, i.e. the same 58 response to common adaptive challenges. Distinguishing homoplasy from homology requires 59 testing a hypothesis on the diversity of cultural traits and understanding it through a 60 comprehensive framework of technological evolution 2–6.

61 The terms Initial Upper Paleolithic, Early Upper Paleolithic and Upper Paleolithic, are 62 commonly used in the scientific literature with different meanings. They could have either a 63 chronological connotation or a technological one, or both of them. This paper refers to the 64 term Initial Upper Paleolithic (IUP) to indicate specific techno-complexes characterised by 65 volumetric blade productions and Levallois reduction sequence 5,7 (Table 2). The term Upper 66 Paleolithic (UP) is used to group the lithic industries characterised by the production of several 67 standardised blades and bladelets often together with ornaments and bone tools (Table 2). 68 The non-Mousterian and non-IUP technologies appeared during the Middle to Upper 69 Palaeolithic transition, comprising Uluzzian 8–13, Châtelperronian 14–16, Szeletian 17–19 and 70 Lincombian-Ranisian-Jermanowician (LRJ) 20–22.

71

8

Definition IUP UP

Initial Upper Paleolithic Upper Paleolithic

Lithic The IUP is a blade-based Blade and bladelet based industries technology production. usually used in composite tools.

The reduction sequences use The broad category of UP includes a direct hard hammer percussion, great diversity of techno-complexes platform faceting, and mostly flat- which share the production of faced or semi-tournant cores. The standardised bladelets and blades with blades and some cores resemble a wide range of technical options (e.g. products of Levallois reduction unidirectional debitage by prismatic sequence 5,7,23. Sites in Siberia core, carinated core, burin-cores, and Mongolia are also among others) and percussion characterised by “burin-cores” for techniques 26–33. producing small blades and the exploitation of the narrow face cores 1,24,25.

Emiran 34, Bokerian 35,36, Included 37–39, Bachokirian 40. Protoaurignacian and techno- 31,32,41–45, Spitsynian 46,47, 48– complex 50, Early Upper Paleolithic 51,52, 53–57.

Other In some IUP assemblage is Habitual use (especially in the material documented the use of personal Gravettian) of portable art, graphic evidence ornament and formal bone tools representations, musical instruments, e.g. in Levant 23, Bulgaria, Siberia, and various bone tools 26,61,62. and Mongolia 7,58–60. Impressive burials sites e.g. Sunghir 63.

Chronologically, the IUP is a long Chronology phenomenon comprised between The various stages of the UP make approximately 50 kya and 35 kya their appearances approximately 42 (calibrated). Its stratigraphic kya. Their stratigraphic position follows

9

position follows the Middle the IUP, Uluzzian, Szeletian, Paleolithic assemblages and is Chatelperroian, LRS. before UP assemblage 17–19s 5,7.

72 Table S2. Main criteria here adopted to characterise Initial Upper Paleolithic (IUP) and 73 Upper Paleolithic (UP) sites.

10

74 Supplementary Section 2: Genetic dataset 75 76 We downloaded the aligned sequences for the IUP Bacho Kiro individuals from the European 77 Nucleotide Archive (accession number PRJEB39134), and merged files from the same 78 individual. We clipped 3 bases off the ends of each read to avoid excess of aDNA damage 79 using trimBam (https://genome.sph.umich.edu/wiki/BamUtil:_trimBam) and then generated 80 pseudo-haploid calls with the genotype caller pileupCaller 81 (https://github.com/stschiff/sequenceTools/tree/master/src-pileupCaller) using the 82 majorityCall option that, for each locus, either chooses the allele supported by the highest 83 number of reads or picks one randomly if the number is the same. Only positions with base 84 quality and read quality higher than 30 were considered. 85 The processed Eigenstrat files of the Zlatý Kůň individual were kindly provided by Prof. Cosimo 86 Posth and Dr. He Yu. 87 We merged the files with the 1240K “Allen Ancient DNA Resource” v44.3 database 88 (https://reich.hms.harvard.edu/allen-ancient-dna-resource-aadr-downloadable-genotypes- 89 present-day-and-ancient-dna-data) 64,65, then kept only autosomal SNPs (1,150,639) using 90 Plink 1.9 66 and then converted back to Eigenstrat format using Admixtools convertf 67. 91 92 Version ID Master Publication GroupID qpGraph ID Pop ID

Chimp.REF Chimp Genome Chimp.REF Chimp

GoyetQ116- GoyetQ116-1 FuNature2016 Belgium_UP_GoyetQ116_1_published_all GoyetQ116-1 1_published

Kostenki14 Kostenki14 FuNature2016 Russia_Kostenki14 Kostenki14

MA1.SG MA1 RaghavanNature2013 Russia_MA1_HG.SG MA1 (Mal’ta)

HGDP00449.SDG HGDP00449 BergstromScience2020 Mbuti.SDG Mbuti

HGDP00462.SDG HGDP00462 BergstromScience2020 Mbuti.SDG Mbuti

HGDP00463.SDG HGDP00463 BergstromScience2020 Mbuti.SDG Mbuti

HGDP00467.SDG HGDP00467 BergstromScience2020 Mbuti.SDG Mbuti

HGDP00474.SDG HGDP00474 BergstromScience2020 Mbuti.SDG Mbuti

HGDP00476.SDG HGDP00476 BergstromScience2020 Mbuti.SDG Mbuti

HGDP00478.SDG HGDP00478 BergstromScience2020 Mbuti.SDG Mbuti

11

HGDP00982.SDG HGDP00982 BergstromScience2020 Mbuti.SDG Mbuti

HGDP00984.SDG HGDP00984 BergstromScience2020 Mbuti.SDG Mbuti

HGDP01081.SDG HGDP01081 BergstromScience2020 Mbuti.SDG Mbuti

Oase1_d Oase1 FuNature2015 Romania_Oase Oase1

B_Papuan-15.DG HGDP00546 PrueferNature2013 Papuan.DG Papuan

S_Papuan-1.DG HGDP00550 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-10.DG HGDP00553 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-11.DG HGDP00555 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-12.DG HGDP00556 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-13.DG HGDP00552 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-14.DG HGDP00554 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-2.DG HGDP00540 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-3.DG HGDP00541 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-4.DG HGDP00543 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-5.DG HGDP00545 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-6.DG HGDP00547 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-7.DG HGDP00548 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-8.DG HGDP00549 SkoglundNature2015 Papuan.DG Papuan

S_Papuan-9.DG HGDP00542 SkoglundNature2015 Papuan.DG Papuan

baa001.SG baa001 SchlebuschScience2017 South_Africa_1900BP.SG South_Africa_2000BP

12

bab001.SG bab001 SchlebuschScience2017 South_Africa_2000BP.SG South_Africa_2000BP

I9028.SG KhoesanLeip SkoglundCell2017 South_Africa_2200BP.SG South_Africa_2000BP zigHunter

I9133.SG UCT386 SkoglundCell2017 South_Africa_1900BP.SG South_Africa_2000BP

Sunghir1.SG Sunghir1 SikoraScience2017 Russia_Sunghir1.SG Sunghir

Sunghir2.SG Sunghir2 SikoraScience2017 Russia_Sunghir2.SG Sunghir

Sunghir3.SG Sunghir3 SikoraScience2017 Russia_Sunghir3.SG Sunghir

Sunghir4.SG Sunghir4 SikoraScience2017 Russia_Sunghir4.SG Sunghir

Tianyuan Tianyuan YangCurrentBiology2017 China_Tianyuan Tianyuan

Ust_Ishim_published.D Ust_Ishim FuNature2014 Russia_Ust_Ishim_HG_published.DG Ust_Ishim G

Vindija_snpAD.DG Vindija Pruefer2017 Vindija_Neanderthal.DG Vindija_Neanderthal

Yana_old.SG Yana1 SikoraNature2019 Russia_Yana_UP.SG Yana

Yana_old2.SG Yana2 SikoraNature2019 Russia_Yana_UP.SG Yana

ZKU002 ZKU002 Pruefer2021 ZlatyKun ZlatyKun

BB7-240 BB7-240 Hajdinjak2021 Bacho_Kiro Bacho_Kiro

CC7-335 CC7-335 Hajdinjak2021 Bacho_Kiro Bacho_Kiro

F6-620 F6-620 Hajdinjak2021 Bacho_Kiro Bacho_Kiro

93 Table S3: Individuals used in qpGraph modelling. A green background indicates samples 94 present in in the “Allen Ancient DNA Resource” v44.3 database while an orange background 95 indicates samples available from ENA.

13

96 Supplementary Section 3: qpGraph analyses 97 98 We ran Admixtools qpGraph 67, https://github.com/DReichLab/AdmixTools, qpGraph version: 99 7365) with the following parameters: outpop: NULL, blgsize: 0.05, lsqmode: YES, diag: .0001, 100 useallsnps: NO, bigiter: 6, forcezmode: YES, initmix: 10000, precision: .0001, zthresh: 3.0, 101 terse: NO, hires: YES. 102 103 104 3.1 Base graph with Zlatý Kůň and Eurasian trifurcation 105 106 The position of Ust’Ishim with respect to Paleolithic Eastern and Western Eurasian is unclear 107 and described as a near-trifurcation 68,69 with D-statistics in the form of (Tianyuan, Sunghir, 108 Ust’Ishim, Chimp), (Tianyuan, Ust’Ishim, Sunghir, Chimp) and (Sunghir, Ust’Ishim, Tianyuan, 109 Chimp) all not significantly different from zero (Table S4). 110 We started with the simple population tree proposed by Prufer and colleagues 70 (Fig 2c) with 111 Zlatý Kůň as the most basal non African and placed Ust’Ishim either basal to the split between 112 Western and Eastern Eurasians, as a sister of East Eurasians or as a sister of West Eurasians 113 and found that all options are supported with the worst |Z-score| < 3. We used Tianyuan as 114 “East Asian'' and either Sunghir or Kostenki14 as “West Eurasian”. 115 116 117 3.2 Adding Bacho Kiro 118 119 We then tried to add the IUP Bacho Kiro individuals in different positions of these graphs, 120 downstream of the Zlatý Kůň split. We assumed Zlatý Kůň to be more basal than Bacho Kiro 121 because of the following reasons: 122 ● Based on the length of introgressed segments and derived allele sharing 123 with later Eurasians, Zlatý Kůň has been estimated to be older than the ~45 ky old 124 Ust’Ishim individual from central Siberia 70 while the Bacho Kiro individuals (45930- 125 42580 calBP) are about the same age, if not a bit younger, than Ust’Ishim (46880- 126 43210 calBP). 127 ● f4 in the form (X, Zlatý Kůň, Bacho Kiro, Chimp) with X a later Eurasian individual 128 including Ust’Ishim tend to be positive, even if not always significant (Table S4). 129 ● The Bacho Kiro individuals, set aside their higher Neanderthal admixture proportion, 130 have been shown to share more alleles with East Asians compared to Europeans, 131 while Zlatý Kůň is basal to the split between the two 71. 132 133 In order to allow the Neanderthal population that admixed with the ancestors of Bacho Kiro 134 (likely living in Europe) to be different from the Neanderthal population that admixed with the 135 ancestors of all non Africans shortly after the Out of Africa (likely in the Middle East) we 136 modeled it as a different node closer to the Neanderthal from Vindija (Croatia); if it was indeed 137 the same Neanderthal population the inferred drift between the two nodes would be 0. 138 Additionally, in order to avoid our results to be driven by later population interaction between 139 Eurasia and western Africa 72 and/or by the putative admixture of Mbuti pygmies with an 140 archaic ghost hominin 73–75 we used four ancient South African hunter gatherers (Table S4) 141 instead of Mbuti. 142

14

143 As a first attempt we placed Bacho Kiro basal to Ust’Ishim and all other Eurasians except Zlatý 144 Kůň (Figure S1.A): the resulting graph has several f4 with |Z-scores| > 3: Bacho Kiro needs to 145 be closer to Tianyuan (f2 = -3) and further from Sunghir (f2 = 2.6). A similar result is obtained 146 when placing Ust’Ishim basal to Bacho Kiro and Bacho Kiro basal to the split between 147 Europeans and East Asians, here represented by Tianyuan and Sunghir (Figure S1.B). 148 Rather than adding an admixture event from a Bacho Kiro related population to Tianyuan we 149 tried to minimize admixture events and instead placed Bacho Kiro as a sister of Tianyuan 150 hence leaving Ust’Ishim basal to the split Europe/East Asia. This returned a graph with the 151 worst Z-score = -2.8 (Figure S1.C). 152

153 154 Figure S1 Placement of Bacho Kiro within the tree proposed in Prufer et al. 2021. When 155 taking into account the excess Neanderthal contribution, Bacho Kiro still yields outlier f4s when 156 placed soon after the split of Zlatý Kůň (A) or after the separation of Ust’Ishim (B). Placing 157 Bacho Kiro as a sister of Tianyuan (C) yields instead no f4 outliers. 158 159 160 The placement of Bacho Kiro that we are proposing with respect to other Eurasians differs 161 from the one proposed by Hajdinjak and colleagues (Hajdinjak et al. 2021) in their Figure S6.2. 162 We speculate this may be due to the presence of Zlatý Kůň in our tree who may inform the 163 genetic drift that characterizes a basal OoA landscape, or to the availability of Neanderthal 164 from the early phases of our qpGraph construction (whose absence, by virtue of the additional 165 Neanderthal ancestry compared to other non Africans, can have the effect of making Bacho 166 Kiro appear more basal if not immediately accounted for)(Table S4). 167 For the sake of completeness we also tried Bacho Kiro as a sister of Sunghir (geographically 168 it would make sense). The graph has several f4 |Z-scores| > 3 and is therefore rejected. 169 170 3.3 Ust’Ishim as an early leaf of the IUP branch 171 To “populate” more the European branch we tried to add Kostenki14, the oldest “genetically 172 European” individual sequenced to date (Figure S2.A – final score 12371). Since we noticed 173 that the drift between nodes B and C was 0 and bearing in mind that D-statistics (Table S4). 174 and previous graphs supported Sunghir/Tianyuan/Ust’Ishim to be a trifurcation, we tried to 175 place Ust-Ishim either as a sister of Tianyuan/Bacho Kiro (Figure S2.B – score 12051) or as 176 a sister of Kostenki14/Sunghir (Figure S2.C – score final 12173). All three graphs produce no 177 f4 outliers and Ust’Ishim shares only one unit of drift with its sisters before branching off (hence

15

178 showing a near-trifurcation). The tree where Ust’Ishim is a sister of Tianyuan/Bacho Kiro 179 (Figure S2.B), however, has the lowest final score and so is the most supported. 180

181 182 Figure S2 Relationship of Ust’Ishim to other Eurasians. After adding Kostenki14 as a 183 crucial reference sample, we tested all possible positions of Ust’Ishim as either basal to East 184 and West Eurasians (A), sister of Bacho Kiro and Tianyuan (B) and sister of Kostenki and 185 Sunghir ( C ). All three graphs produce no f4 outliers and Ust’Ishim shares only one unit of 186 drift with its sisters before branching off (hence showing a near-trifurcation). The tree where 187 Ust’Ishim is a sister of Tianyuan/Bacho Kiro (Figure S2.B), however, has the lowest final score 188 and so is the most supported. 189 190 191 192 3.4 Oase1 could be a descendant of Bacho Kiro, remnant of the IUP movement 193 194 We then tried to place on the emerging tree of Figure S2.B the ~40 ky old Oase1 individual, 195 recovered from a site less than 400 km away from Bacho Kiro which, like and more than 196 the individuals from Bacho Kiro, has been shown to have a Neanderthal relative in its recent 197 genealogy. 198 199 Since it has not been possible to determine a higher affinity of Oase1 for either East Asians 200 or Europeans (retrospectively that might be caused by the combination of low number of SNPs 201 and high Neanderthal admixture proportion), we started by making Oase1 a sister of Zlatý Kůň 202 that later admixed with (Figure S3.A). The resulting graph showed several |Z 203 scores| > 3, a way too high proportion of Neanderthal in Eurasians and a clear tendency for 204 Oase1 and Bacho Kiro to share more drift (f2 = -6.3). We then modeled Oase1 as a sister of 205 Bacho Kiro with an additional Neanderthal pulse and obtained no outlier f4 and a parsimonious 206 placement for this elusive sample (Figure S3.B). 207 208 Our proposed graphs illustrate the claim made by Fu and colleagues 76 that Oase1 209 experienced an additional pulse of Neanderthal admixture between the one shared by all non 210 Africans and the one that occurred 4-6 generations before it lived, and identify said event in 211 the one occurred in the Bacho Kiro population or in a closely related one.

16

212 A similar graph where the node Oase0 splits from BK0 instead of BK1 (Figure S3.C) is also 213 acceptable but does not explain this additional event of Neanderthal admixture. 214 215 Finally, it is worth mentioning that while the low coverage, high contamination and high 216 Neanderthal ancestry of Oase1 prevented the direct assessment of its closer relationship to 217 either Western or Eastern Eurasians, an individual from the same site and with similar age 218 (Oase2) showed a clearly higher affinity for East Asian and Native American populations than 219 with Western Eurasians. The closest sample to Oase2 in outgroup f3 analyses, after Oase1, 220 was reported to be Tianyuan (the individuals from Bacho Kiro cave were not available at the 221 time of those analyses), supporting our claim of its placement in the “genetically East Asian” 222 branch 77. 223

224 225 Figure S3 Placement of Oase along the IUP tree. After accounting for the reported recent 226 pulse of Neanderthal admixture, Oase1 cannot be modelled as a simple basal lineage along 227 the Eurasian tree (A) and is instead best fit as a descendant of the same population Bacho 228 Kiro belonged to (B). This model provides an explanation for the putative three Neanderthal 229 admixtures detected in Oase1 and is consistent with its younger date compared to Bacho Kiro 230 and is therefore to be preferred over an equally fitting model (C) where Bacho Kiro and Oase1 231 have instead parallel evolutionary histories. 232 233 234 235 3.5 Interaction between Zlatý Kůň and Bacho Kiro 236 237 Given the geographical proximity of the two sites we tested whether a contribution to the Bacho 238 Kiro individuals by a population related to Zlatý Kůň is tolerated by our proposed model. 239 240 An admixture event contributing between 2% and 20% from a sister of Zlatý Kůň into the 241 ancestors of Bacho Kiro, depending on the position of Ust’Ishim (Figure S4), yields no outliers. 242 Although not rejected these graphs are all less parsimonious than the graph in Figure S2.B 243 and might pick up a signal of shared drift between Bacho Kiro and Zlatý Kůň that is caused by 244 the reduced time purifying selection had to drive out introgressed Neanderthal segments (Zlatý

17

245 Kůň is the oldest sample and Bacho Kiro has recent Neanderthal introgression), rather than a 246 genuine gene flow. 247

248 249 Figure S4 Interaction between Zlatý Kůň and Bacho Kiro. Given the geographic and partial 250 chronological overlap of Zlatý Kůň and Bacho Kiro samples we allowed for a contribution of 251 the former on the newly arrived IUP population. Such a contribution is acceptable, in varying 252 amounts depending on the position of Ust’Ishim within the tree (A,B,C) but given the viability 253 of the trees without this event we decided to exclude it from further analyses. 254 255 256 3.6 Formation of Paleolithic Siberian ancestry profile 257 258 Paleolithic Siberian populations younger than 40 ky are consistently described as a mix of 259 European and East Asian ancestries 78–80, although a comprehensive cultural or population 260 dynamics to account for these admixture events is still lacking. Building on our dichotomy 261 between Initial Upper Paleolithic (IUP) and Upper Paleolithic waves of expansions across 262 Eurasia, we tried to model ancient Siberian individuals as a mixture of these waves into our 263 best fitting parsimonious graph (Figure S2.B). 264 265 We started by creating a Siberian population that is ancestral to the ~32 ky old Yana 266 individuals 78–80 as a mixture of the E and C nodes in the graph of Figure S2.B. The resulting 267 graph has several f4 outliers and indicates that Yana may fit better as more closely related to 268 Tianyuan (f2 = -6.1) and, to a lesser extent, Bacho Kiro (f2 = -1.9), suggesting that the node 269 of origin of the East Asian component in Siberians needs to be moved further down along that 270 branch. With the East Asian component of Siberians originating from node D the graph still 271 has several outliers and Yana needs to share more drift with Tianyuan (f2 = -3.6), so after 272 taking that into account we obtained a graph that is not rejected (Figure S4.A). 273 To better define the Siberian populations we included the ~24 kya Mal’ta (MA1) individual 76,80, 274 we started by considering it as a sister of Yana (Figure S4.B) and obtained two |Z-scores| ~3,1 275 (Chimp, South_Africa_2000BP, ZlatyKun, Yana and ZlatyKun, Tianyuan, Sunghir, MA1). 276 Since qpGraph performs f-statistic analyses on a subset of SNPs that is shared by all the

18

277 individuals in a model, this might sometimes severely limit the number of SNPs used in a test, 278 especially when ancient individuals are included. We tested whether the two outliers remain 279 significantly different from zero when including all SNPs available to each quadruple and found 280 they do not (Table S4), deeming this outcome as more robust. 281 Finally, in order to allow MA1 and Yana to have different proportions of East and West 282 Eurasian components, we added an EUR_Sib node and parallelized the admixture events of 283 Yana and MA1. The resulting graph (Figure S5.C) has two Dstat Z-scores (Chimp, 284 South_Africa_2000BP, ZlatyKun, Yana and ZlatyKun, Tianyuan, Sunghir, MA1) higher than 3 285 in absolute value (the same two that were present earlier, with Z = 3.1 and -3.0 respectively), 286 however both of them result not significantly different from zero when computing the D-statistic 287 using the all the available SNPs (Table S4) . 288

289 290 Figure S5 Adding Paleolithic Siberians to the qpGraph of Figure S2.B. We first placed 291 Yana1 on the graph as a putative mixture of West Eurasian (EUR) and East Asian (EAS) 292 components (A) and, after having obtained no outliers we proceeded with adding Mal’ta as a 293 sister leaf (B). To allow for the two Siberian samples to derive independent fractions of each 294 ancestry we then relaxed the proposed model to obtain (C). 295 296 297 3.7 Adding GoyetQ116-1 to the emerging picture of IUP and UP interaction 298 299 The ~35000 years old GoyetQ116-1 individual from Belgium, although closer to Europeans 300 than to East Asians, shares more alleles with the latter compared to other contemporary 301 Europeans 81 and it has recently been described as a recipient of gene flow from a population 302 related to Bacho Kiro 71. 303 304 We started from Figure S5.C and modeled Goyet as mixture between the Kostenki14/Sunghir 305 branch and the Bacho Kiro branch (Figure S6.A); however doing so shows that Goyet needs 306 to share more drift with Tianyuan (f2 = -2.4, plus several f4 outliers) highlighting how its East 307 Asian genetic component is more variegated than the one found in Bacho Kiro or Tianyuan 308 alone.

19

309 When adding to Goyet a contribution from both the Bacho Kiro and Tianyuan branches we 310 obtained only two minor outliers (Chimp, South_Africa_2000BP, ZlatyKun, Yana = 3.4 and 311 South_Africa_2000BP, ZlatyKun, ZlatyKun, Sunghir = -3.3), the first of which results non 312 significant when computed using all SNPs (Table S4). The results are comparable whether 313 the two genetically East Asian components admix with each other first, and then with the 314 European one (Figure S6.B), or vice versa (Figure S6.C). 315

316 317 Figure S6 Goyet Q116-1 as a mixture of UP and IUP lineages. Modelling the East Asian 318 component found in Goyet Q116-1 as a simple interaction between UP and the Bacho Kiro 319 branch (A)yields an unexplained attraction between Goyet Q116-1 and Tianyuan. Modelling 320 the IUP component as a mixture of Bacho Kiro- and Tianiyuan-related lineages before (B) or 321 after ( C) the arrival of the UP lineages provides a better fit to the data. 322 323 324 3.8 Peopling of Oceania (Graph 7) 325 326 The position of Oceanian populations with respect to the East and West Eurasian in still 327 unclear, with Oceanians often seen as either an earlier split 82,83 or a sister population to East 328 Asians 84,85. 329 Given the relatively simple population tree needed to explain the post OoA Eurasian 330 population history using aDNA samples available to date, and benefiting from the basal 331 position of Zlatý Kůň, we tried to model Oceanian populations (modern Papuans) within our 332 hypothesized model. We avoided including the sampled Denisova aDNA 86 within the 333 population tree to eliminate attractions from yet uncharacterized, deep splits along the hominin 334 branch, and opted for a surrogate ghost split along the archaic human lineage to account for 335 the documented Denisova presence in Papuans 82,83,87,88. 336 337 We started from Graph S2.B and tried to have Papuans split either before or after the split of 338 the Zlatý Kůň lineage; in both cases the graph was rejected as papuans needed to share more 339 drift with Tianyuan(f2: -11 and -8). For this reason we then modeled them as a sister group of 340 Tianyuan and obtained a graph whose only two outliers (South_Africa_2000BP, Bacho Kiro,

20

341 Zlatý Kůň, Papuan = -3.4 and Zlatý Kůň, Papuan, Bacho Kiro, Sunghir = 3.0) resulted non 342 significant when taking all SNPs into account (Figure 1.B, Figure S7.A). 343 Subsequently we tried to model Papuans as a mixture of a Tianyuan sister population and a 344 more basal lineage; the contribution from said lineage decreases the more that split is moved 345 backward: it is 96% when splitting just before the split Tianyuan/Bacho Kiro (Figure S7.B), 346 53% when splitting before Ust’Ishim (Figure S7.C), 40% when splitting after Zlatý Kůň but 347 before any other Eurasian branching (Figure S7.D) and only 2% when splitting before Zlatý 348 Kůň (Figure S7.E). Finally, we decided to test whether a small contribution to Papuans from a 349 population of AMH that left Africa before the 70-60 kya Out of Africa (xOoA) [Pagani 2016] is 350 rejected by the model (Figure S7.F). A 2% contribution from said population to modern 351 papuans is not rejected since the only f4 outlier(Zlatý Kůň, Papuans, Bacho Kiro, Sunghir = 352 3.0) is ~0 when including all available SNPs (Table S4). 353 All the acceptable solutions for the placement of Papuans within the broader OoA tree confirm 354 Zlatý Kůň as the most basal human genome among the ones ever found Out of Africa. It is 355 therefore reasonable to describe Oceanians as either an almost even mixture between East 356 Asians and a lineage basal to West and East Asians occurred sometimes between 45 and 357 37kya, or as a sister lineage of East Asians with or without a minor basal OoA or xOoA 358 contribution.

21

359 360 Figure S7 The place of Papuans within the out of Africa graph. Modern Papuans can be 361 invariably described as a sister population of Tianyuan (A) with or without the addition of a 362 more basal, H. sapiens lineage. The position of this deeper lineage along the population model 363 influences the genetic contribution it gives to Papuans: 94% if placed within the IUP branch 364 after Ust’Ishim (B), 53% before Ust’Ishim (C), 4% before the separation of IUP and UP 365 branches (D) and 2% either as the most basal OoA branch (E) or even as an extinct, extra 366 OoA (xOoA) (F). 367 368 369 370 371

22

372 373 374 375 Pop1 Pop2 Pop3 Pop4 D Z SNPs Supp. Section

Ust_Ishim Kostenki14 Tianyuan Chimp 0.0006 0.087 817238 3.1

Ust_Ishim Sunghir Tianyuan Chimp -0.0023 -0.34 847217 3.1

Tianyuan Kostenki14 Ust_Ishim Chimp -0.0037 -0.523 817238 3.1

Tianyuan Sunghir Ust_Ishim Chimp 0.0024 0.386 847217 3.1

Ust_Ishim Tianyuan Kostenki14 Chimp 0.0043 0.588 817238 3.1

Ust_Ishim Tianyuan Sunghir Chimp -0.0047 -0.758 847217 3.1

Bacho_Kiro Kostenki14 Ust_Ishim Chimp -0.0263 -4.088 997593 3.2

Bacho_Kiro Sunghir Ust_Ishim Chimp -0.0178 -3.261 1052955 3.2

Bacho_Kiro Tianyuan Ust_Ishim Chimp -0.0207 -3.223 841030 3.2

Tianyuan ZlatyKun Bacho_Kiro Chimp 0.0364 5.142 554809 3.2

Kostenki14 ZlatyKun Bacho_Kiro Chimp 0.0131 1.878 594315 3.2

Sunghir ZlatyKun Bacho_Kiro Chimp 0.0138 2.199 620949 3.2

Ust_Ishim ZlatyKun Bacho_Kiro Chimp 0.0192 2.762 619549 3.2

ZlatyKun Tianyuan Sunghir MA1 0.0085 1.112 422908 3.6

Chimp South_Africa_2000BP ZlatyKun Yana 0.0122 2.864 625628 3.6, 3.7

South_Africa_2000BP Bacho_Kiro ZlatyKun Papuan 0.0115 1.975 636119 3.8

ZlatyKun Papuan Bacho_Kiro Sunghir 0.0003 0.046 636130 3.8

23

Kostenki14 ZlatyKun Papuan Chimp 0.029 4.028 587236 3.8

Sunghir ZlatyKun Papuan Chimp 0.027 4.452 614406 3.8

Ust_Ishim ZlatyKun Papuan Chimp 0.0363 5.213 613032 3.8

Tianyuan ZlatyKun Papuan Chimp 0.0736 10.746 547447 3.8

376 377 Table S4: Relevant f4 tests. The first seven columns report population and values as 378 outputted by Admixtools. The eight column reports the Supplementary Section where a given 379 f4 is mentioned. 380

24

381 Supplementary Section 4: Paleomaps plotting 382 383 To plot the paleomaps, we used R (version 4.0.5 - “Shake and Throw”), the GUI RStudio 384 (version 1.4.1103 - “Wax Begonia”) and the package “oce” (https://dankelley.github.io/oce/). 385 We downloaded the alti-bathymetric maps from the ETOPO1 dataset 89 and used the inferred 386 sea level values at given times in the past from the “Global 1Ma Temperature, Sea Level, and 387 Ice Volume Reconstructions” dataset 90 available from the NOAA Paleoclimatology Program. 388 We plotted the maps applying Lambert conformal conic projections ("+proj=lcc +lat_1=25 389 +lat_2=50 +lon_0=97" in PROJ.4 format). 390 391 392 393 Supplementary References

394 1. Zwyns, N. & Lbova, L. V. The Initial Upper Paleolithic of Kamenka site, Zabaikal region

395 (Siberia): A closer look at the blade technology. Archaeol. Res. Asia 17, 24–49 (2019).

396 2. Boëda, E., Hou, Y. M., Forestier, H., Sarel, J. & Wang, H. M. Levallois and non-Levallois

397 blade production at Shuidonggou in Ningxia, North China. Quat. Int. 295, 191–203

398 (2013).

399 3. Morgan, C. et al. Redating Shuidonggou Locality 1 and Implications for the Initial Upper

400 Paleolithic in East Asia. Radiocarbon vol. 56 165–179 (2014).

401 4. Peng, F. et al. A chronological model for the Late Paleolithic at Shuidonggou Locality 2,

402 North China. PLoS One 15, e0232682 (2020).

403 5. Kuhn, S. L. Initial Upper Paleolithic: A (near) global problem and a global opportunity.

404 Archaeological Research in Asia 17, 2–8 (2019).

405 6. Boyd, R. & Richerson, P. J. Culture and the Evolutionary Process. (University of

406 Chicago Press, 1988).

407 7. Kuhn, S. L. & Zwyns, N. Rethinking the initial Upper Paleolithic. Quaternary International

408 vol. 347 29–38 (2014).

409 8. Benazzi, S. et al. Early dispersal of modern in Europe and implications for

410 Neanderthal behaviour. Nature 479, 525–528 (2011).

411 9. Moroni, A. et al. (Apulia-Southern Italy). The Uluzzian in the mirror. J.

412 Anthropol. Sci. 96, 125–160 (2018).

25

413 10. Peresani, M., Bertola, S., Delpiano, D., Benazzi, S. & Romandini, M. The Uluzzian in the

414 north of Italy: insights around the new evidence at Riparo Broion. Archaeological and

415 Anthropological Sciences vol. 11 3503–3536 (2019).

416 11. Collina, C. et al. Refining the Uluzzian through a new lithic assemblage from Roccia San

417 Sebastiano (Mondragone, southern Italy). Quaternary International vol. 551 150–168

418 (2020).

419 12. Marciani, G. et al. Lithic techno-complexes in Italy from 50 to 39 thousand years BP: An

420 overview of lithic technological changes across the Middle-Upper Palaeolithic boundary.

421 Quaternary International vol. 551 123–149 (2020).

422 13. Riel-Salvatore, J. What Is a ‘Transitional’ Industry? The Uluzzian of Southern Italy as a

423 Case Study. Sourcebook of Paleolithic Transitions 377–396 (2009) doi:10.1007/978-0-

424 387-76487-0_25.

425 14. Roussel, M. Méthodes et rythmes du débitage laminaire au Châtelperronien :

426 comparaison avec le Protoaurignacien. C. R. Palevol 12, 233–241 (2013).

427 15. Ruebens, K., McPherron, S. J. P. & Hublin, J.-J. On the local Mousterian origin of the

428 Châtelperronian: Integrating typo-technological, chronostratigraphic and contextual

429 data. J. Hum. Evol. 86, 55–91 (2015).

430 16. Roussel, M., Soressi, M. & Hublin, J.-J. The Châtelperronian conundrum: Blade and

431 bladelet lithic technologies from Quinçay, . J. Hum. Evol. 95, 13–32 (2016).

432 17. Nerudová, Z. & Neruda, P. Technology of Moravian Early Szeletian leaf point shaping: A

433 case study of refittings from Moravský Krumlov IV open-air site ().

434 Quaternary International vol. 428 91–108 (2017).

435 18. Mester, Z. The problems of the Szeletian as seen from Hungary. Recherches

436 Archéologique Nouvelle Serie vol. 9 19–48 (2018).

437 19. Neruda, P. & Nerudová, Z. Technology of Early Szeletian leaf point shaping: a refitting

438 approach. Archaeological and Anthropological Sciences vol. 11 4515–4538 (2019).

439 20. Flas, D. The Middle to Upper Paleolithic transition in Northern Europe: the Lincombian-

440 Ranisian-Jerzmanowician and the issue of acculturation of the last Neanderthals. World

26

441 Archaeology vol. 43 605–627 (2011).

442 21. Kot, M. et al. Chronostratigraphy of Jerzmanowician. New data from Koziarnia Cave,

443 . (2020) doi:10.1101/2020.04.29.067967.

444 22. Krajcarz, M. T., Krajcarz, M., Ginter, B., Goslar, T. & Wojtal, P. Towards a chronology of

445 the jerzmanowician-a new series of radiocarbon dates from nietoperzowa cave

446 (Poland). Archaeometry 60, 383–401 (2018).

447 23. Kuhn, S. L. et al. The early Upper Paleolithic occupations at Uçağizli Cave (Hatay,

448 Turkey). J. Hum. Evol. 56, 87–113 (2009).

449 24. Slavinsky, V. S., Rybin, E. P., Khatsenovich, A. M. & Belousova, N. E. Intentional

450 fragmentation of blades in the initial upper Paleolithic industries of the Kara-Bom site

451 (Altai, Russia). Archaeological Research in Asia vol. 17 50–61 (2019).

452 25. Zwyns, N., Rybin, E. P., Hublin, J.-J. & Derevianko, A. P. Burin-core technology and

453 laminar reduction sequences in the initial Upper Paleolithic from Kara-Bom (Gorny-Altai,

454 Siberia). Quaternary International vol. 259 33–47 (2012).

455 26. Zilhão, J. et al. Analysis of Aurignacian interstratification at the Chatelperronian-type site

456 and implications for the of Neandertals. Proc. Natl. Acad. Sci. U.

457 S. A. 103, 12643–12648 (2006).

458 27. Le Brun-Ricalens, F. L., Bordes, J.-G. & Eizenberg, L. A crossed-glance between

459 southern European and Middle-Near Eastern early Upper Palaeolithic lithic

460 technocomplexes. Existing models, new perspectives. in The Mediterranean from

461 50,000 to 25,000 bp: Turning Points and New Directions. vol. 66 11–33 (2009).

462 28. Teyssandier, N., Bon, F. & Bordes, J.-G. WITHIN PROJECTILE RANGE: Some

463 Thoughts on the Appearance of the Aurignacian in Europe. Journal of Anthropological

464 Research vol. 66 209–229 (2010).

465 29. Moreau, L. Le Gravettien ancien d’Europe centrale revisité : mise au point et

466 perspectives. L’Anthropologie vol. 116 609–638 (2012).

467 30. Kozłowski, J. K. The origin of the Gravettian. Quat. Int. 359-360, 3–18 (2015).

468 31. Bataille, G. Extracting the ‘Proto’ from the Aurignacian. Distinct production sequences of

27

469 blades and bladelets in the lower Aurignacian phase of Siuren I, units H and G

470 (Crimea). Mitteilungen der Gesellschaft für Urgeschichte 25, e85 (2016).

471 32. Falcucci, A. Towards a renewed definition of the Protoaurignacian. Mitteilungen der

472 Gesellschaft für Urgeschichte 27, 87–130 (2018).

473 33. Kadowaki, S., Suga, E. & Henry, D. O. Frequency and production technology of

474 bladelets in Late , Initial Upper Paleolithic, and Early Upper Paleolithic

475 (Ahmarian) assemblages in Jebel Qalkha, Southern . Quat. Int. (2021)

476 doi:10.1016/j.quaint.2021.03.012.

477 34. Kuhn, S. L., Stiner, M. C. & Güleç, E. Initial Upper Palaeolithic in south-central Turkey

478 and its regional context: a preliminary report. Antiquity 73, 505–517 (1999).

479 35. Leder, D. Technological and typological change at the Middle to Upper Palaeolithic

480 Boundary in Lebanon. (2014).

481 36. Leder, D. Core reduction strategies at the Initial Upper Palaeolithic sites of and

482 Abou Halka in Lebanon. Lithics, The Journal of the Lithic Studies Society. (2017).

483 37. Richter, D., Tostevin, G. & Skrdla, P. Bohunician technology and thermoluminescence

484 dating of the type locality of Brno-Bohunice (Czech Republic). J. Hum. Evol. 55, 871–

485 885 (2008).

486 38. Škrdla, P. Middle to Upper Paleolithic transition in Moravia: New sites, new dates, new

487 ideas. Quaternary International vol. 450 116–125 (2017).

488 39. Demidenko, Y. E., Škrdla, P. & Rychtaříková, T. Initial Upper Paleolithic bladelet

489 production: Bladelets in Moravian Bohunician. Přehled výzkumů 21–29 (2020)

490 doi:10.47382/pv0611-02.

491 40. Hublin, J.-J. et al. Initial Upper Palaeolithic Homo sapiens from Bacho Kiro Cave,

492 Bulgaria. Nature 581, 299–302 (2020).

493 41. Teyssandier, N., Bon, F. & Bordes, J.-G. WITHIN PROJECTILE RANGE: Some

494 Thoughts on the Appearance of the Aurignacian in Europe. Journal of Anthropological

495 Research vol. 66 209–229 (2010).

496 42. Sitlivy, V. et al. The earliest Aurignacian in Romania: New investigations at the open air

28

497 site of Româneşti-Dumbrăviţa I (Banat). (2012) doi:10.7485/QU59_4.

498 43. Demidenko, Y. E. & Škrdla, P. Aurignacian in Moravia New geochronological, lithic and

499 settlement data. Památky archeologické (2017).

500 44. Tafelmaier, Y. Technological Variability at the Beginning of the Aurignacian in Northern

501 : Implications for the Proto- and Early Aurignacian Distinction. (2017).

502 45. Riel-Salvatore, J. & Negrino, F. Proto-aurignacian lithic technology, mobility, and human

503 niche construction: A case study from riparo bombrini, Italy. in Studies in Human

504 Ecology and Adaptation 163–187 (Springer International Publishing, 2018).

505 46. Vishnyatsky, L. B. & Nehoroshev, P. E. The beginning of the Upper Paleolithic on the

506 Russian Plain. The Early Upper Paleolithic Beyond Western Europe 80–96 (2004).

507 47. Usik, V. I., Monigal, K. & Kulakovskaya, L. New perspectives on the Transcarpathian

508 Middle to Upper Paleolithic boundary. (na, 2006).

509 48. Kadowaki, S., Omori, T. & Nishiaki, Y. Variability in Early Ahmarian lithic technology and

510 its implications for the model of a Levantine origin of the Protoaurignacian. J. Hum. Evol.

511 82, 67–87 (2015).

512 49. Barzilai, O., Hershkovitz, I. & Marder, O. The Early Upper Paleolithic Period at Manot

513 Cave, Western Galilee, Israel. Hum. Evol. 31, 85–10 (2016).

514 50. Goring-Morris, N. & Belfer-Cohen, A. The Ahmarian in the Context of the Earlier Upper

515 Palaeolithic in the . in The Middle and Upper Paleolithic Archeology of the

516 Levant and Beyond (eds. Nishiaki, Y. & Akazawa, T.) 87–104 (Springer Singapore,

517 2018).

518 51. Hoffecker, J. F. The early upper Paleolithic of eastern Europe reconsidered. Evol.

519 Anthropol. 20, 24–39 (2011).

520 52. Kadowaki, S., Suga, E. & Henry, D. O. Frequency and production technology of

521 bladelets in Late Middle Paleolithic, Initial Upper Paleolithic, and Early Upper Paleolithic

522 (Ahmarian) assemblages in Jebel Qalkha, Southern Jordan. Quaternary International

523 (2021) doi:10.1016/j.quaint.2021.03.012.

524 53. Sinitsyn, A. A. Variabilité du Gravettien de Kostienki (Bassin moyen du Don) et des

29

525 territoires associés. Paléo 181–201 (2007).

526 54. Dobrovolskaya, M., Richards, M.-P. & Trinkaus, E. Direct radiocarbon dates for the Mid

527 Upper Paleolithic (eastern Gravettian) burials from Sunghir, Russia. Bull. Mem. Soc.

528 Anthropol. Paris 24, 96–102 (2012).

529 55. Moreau, L. Le Gravettien ancien d’Europe centrale revisité : mise au point et

530 perspectives. Anthropologie 116, 609–638 (2012).

531 56. Kozłowski, J. K. The origin of the gravettian. Quat. Int. 359-360, 3–18 (2015).

532 57. Klaric, L. Faciès lithiques et chronologie du Gravettien du sud du Bassin parisien et de

533 sa marge sud-occidentale. Pierre Bodu, Lucy Chehmana, Laurent Klaric, Ludovic

534 Mevel, Sylvain Soriano éds 56, 61–88 (2013).

535 58. Rybin, E. P. Tools, beads, and migrations: Specific cultural traits in the Initial Upper

536 Paleolithic of Southern Siberia and Central Asia. Quat. Int. 347, 39–52 (2014).

537 59. Kaifu, Y., Izuho, M., Goebel, T., Sato, H. & Ono, A. Emergence and Diversity of Modern

538 Human Behavior in Paleolithic Asia. (Texas A&M University Press, 2014).

539 60. Derevianko, A. P. & Rybin, E. P. The earliest representations of symbolic behavior by

540 Paleolithic humans in the Altai Mountains. Archaeology, Ethnology and Anthropology of

541 Eurasia 27–50 (2003).

542 61. Conard, N. J. Palaeolithic ivory sculptures from southwestern Germany and the origins

543 of figurative art. Nature 426, 830–832 (2003).

544 62. Conard, N. J., Malina, M. & Münzel, S. C. New flutes document the earliest musical

545 tradition in southwestern Germany. Nature 460, 737–740 (2009).

546 63. Trinkaus, E., Buzhilova, A. P., Mednikova, M. B. & Dobrovolʹskai︠a︡, M. V. The People of

547 Sunghir: Burials, Bodies, and Behavior in the Earlier Upper Paleolithic. (Oxford

548 University Press, 2014).

549 64. Skoglund, P. et al. Genetic evidence for two founding populations of the Americas.

550 Nature 525, 104–108 (2015).

551 65. Bergström, A. et al. Insights into human genetic variation and population history from

552 929 diverse genomes. Science 367, (2020).

30

553 66. Chang, C. C. et al. Second-generation PLINK: rising to the challenge of larger and

554 richer datasets. Gigascience 4, 7 (2015).

555 67. Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065–1093

556 (2012).

557 68. Yang, M. A. et al. 40,000-Year-Old Individual from Asia Provides Insight into Early

558 Population Structure in Eurasia. Curr. Biol. 27, 3202–3208.e9 (2017).

559 69. Lipson, M. & Reich, D. A Working Model of the Deep Relationships of Diverse Modern

560 Human Genetic Lineages Outside of Africa. Mol. Biol. Evol. 34, 889–902 (2017).

561 70. Prüfer, K. et al. A genome sequence from a modern human skull over 45,000 years old

562 from Zlatý kůň in Czechia. Nat Ecol Evol (2021) doi:10.1038/s41559-021-01443-x.

563 71. Hajdinjak, M. et al. Initial Upper Palaeolithic humans in Europe had recent Neanderthal

564 ancestry. Nature 592, 253–257 (2021).

565 72. Chen, L., Wolf, A. B., Fu, W., Li, L. & Akey, J. M. Identifying and Interpreting Apparent

566 Neanderthal Ancestry in African Individuals. Cell 180, 677–687.e16 (2020).

567 73. Hsieh, P. et al. Model-based analyses of whole-genome data reveal a complex

568 evolutionary history involving archaic introgression in Central African Pygmies. Genome

569 Res. 26, 291–300 (2016).

570 74. Hammer, M. F., Woerner, A. E., Mendez, F. L., Watkins, J. C. & Wall, J. D. Genetic

571 evidence for archaic admixture in Africa. Proc. Natl. Acad. Sci. U. S. A. 108, 15123–

572 15128 (2011).

573 75. Lipson, M. et al. Ancient West African foragers in the context of African population

574 history. Nature 577, 665–670 (2020).

575 76. Fu, Q. et al. An early modern human from Romania with a recent Neanderthal ancestor.

576 Nature 524, 216–219 (2015).

577 77. Siska, V. Human population history and its interplay with natural selection. (2019)

578 doi:10.17863/CAM.31536.

579 78. Massilani, D. et al. Denisovan ancestry and population history of early East Asians.

580 Science 370, 579–583 (2020).

31

581 79. Sikora, M. et al. The population history of northeastern Siberia since the Pleistocene.

582 Nature 570, 182–188 (2019).

583 80. Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry of Native

584 Americans. Nature 505, 87–91 (2014).

585 81. Fu, Q. et al. The genetic history of Ice Age Europe. Nature 534, 200–205 (2016).

586 82. Malaspinas, A.-S. et al. A genomic history of Aboriginal . Nature 538, 207–214

587 (2016).

588 83. Choin, J. et al. Genomic insights into population history and biological adaptation in

589 Oceania. Nature 592, 583–589 (2021).

590 84. Wall, J. D. Inferring Human Demographic Histories of Non-African Populations from

591 Patterns of Allele Sharing. Am. J. Hum. Genet. 100, 766–772 (2017).

592 85. Mallick, S. et al. The Simons Genome Diversity Project: 300 genomes from 142 diverse

593 populations. Nature 538, 201–206 (2016).

594 86. Meyer, M. et al. A high-coverage genome sequence from an archaic Denisovan

595 individual. Science 338, 222–226 (2012).

596 87. Reich, D. et al. Denisova admixture and the first modern human dispersals into

597 Southeast Asia and Oceania. Am. J. Hum. Genet. 89, 516–528 (2011).

598 88. Meyer, M. et al. A high-coverage genome sequence from an archaic Denisovan

599 individual. Science 338, 222–226 (2012).

600 89. Amante, C. & Eakins, B. W. ETOPO1 Global Relief Model converted to PanMap layer

601 format. (2009) doi:10.1594/PANGAEA.769615.

602 90. Bintanja, R., van de Wal, R. S. W. & Oerlemans, J. Modelled atmospheric temperatures

603 and global sea levels over the past million years. Nature 437, 125–128 (2005).

604

32