A new Nature study rewrites the history of Papua New Guinea: relevance for Holocene-selection on intelligence

The island of New Guinea constitutes the northern part of the Sahul paleo-continent, just south of the equator.

A new study published in Nature (Mondal et al., 2025) has provided crucial validation for what I've long suspected about Papua New Guinea's outlier patterns in polygenic score analyses. While I've always accounted for the importance of different linkage disequilibrium (LD) patterns in cross-population comparisons, this paper finally quantifies the specific demographic mechanisms behind PNG's unique genetic architecture—mechanisms that explain why PNG populations tend to produce outlier polygenic scores in my analyses. Moreover, it sheds lights on the possible origin of their reported lower average IQ.

Quantifying PNG's Demographic Constraints

The study resolves a long-standing puzzle in PNG genetics through sophisticated demographic modeling. Previous analyses suggested PNG populations diverged from Africans around 100,000 years ago, leading to speculation about early Out-of-Africa populations or heavy Denisovan influence. The Nature paper demonstrates that PNG's apparent genetic "ancientness" actually reflects two key demographic constraints: a severe ancestral bottleneck reducing effective population size to approximately 674 individuals (compared to over 3,500 in Europeans and over 1,700 in East Asians), followed by persistently low population growth rates.

The best-fitting parameters for Model A* (Mondal et al., 2025)

*AFR Africa, EUR European, ASN East Asian, PAP Papua New Guinean, NEAI introgressed Neanderthal, NEAS sequenced Altai Neanderthal, DENI introgressed Denisova, and DENS sequenced Denisova.

The analysis confirms PNG diverged from East Asians around 46,200 years ago, with minimal Neanderthal contributions shared with all OOA populations (4.04%) and later Denisovan introgression (around 3.23% at roughly 31,200 years ago).

Importantly, the authors focus on Papuan Highlanders, which are relatively unadmixed compared to coastal populations, to avoid confounding from more recent admixture. In fact, the latter have received substantial amounts of East Asian-like admixture over the last 3200 years, due to the migrations of Austronesian speakers, as confirmed by direct evidence from ancient DNA (Nägele et al., 2025).

Validating Known LD Architecture Problems

This demographic reconstruction provides the precise quantitative framework for why PNG populations show poor polygenic score portability—something I've consistently observed and adjusted for in my research. Population genetics theory predicts that PNG's severe bottleneck followed by minimal growth would fundamentally reshape LD patterns by drastically reducing haplotype diversity through genetic drift, limiting recombination events over millennia, and preserving large unrecombined blocks including Denisovan segments.

The result is what I've been working around: European-trained polygenic scores perform poorly in PNG because the tagging SNPs capture causal variants within European-specific LD contexts that don't translate to PNG's divergent LD landscape. The same SNPs tag different genomic regions, frequencies of trait-associated alleles were distorted by bottleneck drift, and Denisovan haplotypes introduce unique LD blocks absent from European reference panels.

The Social complexity + farming model of intelligence

The study's demographic parameters also align with archaeological evidence that PNG experienced no Neolithic population explosion comparable to Eurasia. This demographic stasis preserved genetic architectures reflecting pre-agricultural selective regimes, while Eurasian populations underwent rapid expansion that continuously fragmented LD blocks over generations.

This connects to the social complexity + farming model (SCF), which predicts that PNG's late agricultural transition and low population growth produced weaker selective pressure on cognitive traits compared to populations with rapid Neolithic expansion. Ancient genome studies have validated this model by demonstrating post-Neolithic increases in cognitive-related polygenic scores in European and East Asian populations (Piffer & Kirkegaard, 2024; Piffer, 2025).

Estimates for the average IQ score of Papuans are in the 75-85 range, comparable to that of Africans. This is what would be predicted by the SCF model.

References

Mondal, M., André, M., Pathak, A.K. et al. Resolving out of Africa event for Papua New Guinean population using neural network. Nat Commun 16, 6345 (2025). https://doi.org/10.1038/s41467-025-61661-w

Nägele, K., Kinaston, R., Gaffney, D. et al. The impact of human dispersals and local interactions on the genetic diversity of coastal Papua New Guinea over the past 2,500 years. Nat Ecol Evol 9, 908–923 (2025). https://doi.org/10.1038/s41559-025-02710-x

Piffer, D., and E. O. Kirkegaard (2024). Evolutionary trends of polygenic scores in European populations from the Paleolithic to modern times. Twin Research and Human Genetics, 27, 30-49.

Piffer, D. (2025). Directional Selection and Evolution of Polygenic Traits in Eastern Eurasia: Insights from Ancient DNA. Twin Research and Human Genetics, 28, 1-20.