Abstract

New Guinea stands out in studies of human diversity because of its linguistic richness, unique transition from hunting and gathering to food production and, in genetic analyses, for its lack of large-scale data. We have generated high-coverage whole-genome sequences from 39 individuals from Papua New Guinea (PNG) and genomewide SNP genotypes from 381, representing 85 language groups from both highland and lowland populations. Using these data, together with available datasets from other areas, we have investigated the genetic history of the region.

We find little evidence of admixture with external populations in the highlands, but substantial East Asian admixture in the lowlands, reflecting population expansions from Southeast Asia in the last few millennia, and excluded these components from subsequent analyses where appropriate. Genetic divergence between Asia and PNG dates back to around 50,000 years, indicating a long history of independent evolution. Within PNG , genetic differentiation is high, both between highlands and lowlands, and within each of these areas, exceeding levels typically seen between major populations within entire continents and mirroring the great linguistic and cultural diversity of the country. Yet the within-PNG differentiation appears to have a much more recent origin 20,000 years ago or less. Incorporating both genetic and non-genetic data, we propose a model where the present-day structure is dominated by a population expansion in the highlands from a single genetic source, starting in the early Holocene and potentially accompanied by the spread of the Trans-New Guinea language family and food production.