Extended Data Figure 3: Introgressed archaic sites and putative Denisovan and Neanderthal haplotypes.

a, Distribution of number of putative introgressed sites per individual from archaic humans. The number of Neanderthal-specific introgressed sites per individual increases from Europe to Australia, and then decreases in Amerindians, which is consistent with recurrent Neanderthal (or Neanderthal-related archaic) gene flow during the expansion into Eurasia. Our results are thus indicative of several pulses of Neanderthal gene flow into modern humans, as inferred previously^48,49,50. We note, however, that the apparent high levels of Neanderthal-specific introgressed sites in Australo-Papuans can be explained by the expected number of misclassified Neanderthal introgressed sites resulting from the shared ancestry with Denisovans (see Supplementary Information section S11 for details). b–e, Putative Denisovan (PDH) and Neanderthal haplotypes (PNH). The putative haplotypes correspond to clusters (four or more SNPs spanning at least 4 kb) of heterozygous or homozygous genotypes in complete linkage disequilibrium (‘diplotypes’) that are potentially the result of Neanderthal or Denisovan admixture. Those diplotypes are homozygous ancestral in 10 Africans, homozygous derived in the Denisovan for the PDH (respectively Neanderthal for the PNH), homozygous ancestral in the Neanderthal for the PDH (respectively Denisovan for the PNH), and with the derived allele segregating in all other contemporary non-African humans (see Supplementary Information section S10 for details). We report the average number of PDHs and PNHs (b), the correlation between the estimated amount of Australo-Papuan ancestry (see Fig. 2a, Extended Data Fig. 1, Supplementary Information section S05) and the number of identified PDHs for each Australian sample (c), the sum of the lengths (d) and the average length (e) of the PDHs and PNHs per individual for worldwide populations included in our reference panel (see Supplementary Information section S04).