David Reich – Bronze Age shock, the Neanderthal puzzle, & the sudden spread of farming

David Reich, a Harvard geneticist specializing in ancient DNA, discusses a new preprint revealing that natural selection in humans has been far more intense and "rampant" than previously believed, particularly over the last 5,000 years during the Bronze Age.

• The study challenges the long-standing view that directional natural selection has been "quiescent" for hundreds of thousands of years, a view derived from the lack of 100% frequency differences between major populations like Europeans and East Asians.
• Prior to this work, small sample sizes and the confounding effects of massive population migrations (genetic drift) obscured signals of selection, which accounted for only ~2% of frequency changes compared to ~98% caused by migration and drift.

Methodological Breakthrough

• The study leveraged a dataset of approximately 16,000 ancient individuals from Europe and the Middle East, a 14-fold increase in data volume compared to previous efforts, allowing for the detection of subtle frequency changes over time.
• Researchers developed a novel statistical method using a "genetic relatedness matrix" to model population history (drift, admixture, bottlenecks) and isolate the specific signal of directional natural selection.
• Validation was achieved by cross-referencing selection signals with modern Genome-Wide Association Studies (GWAS); mutations with the highest selection statistics showed a 5-fold enrichment for affecting known biological traits, confirming the signals were real and not statistical artifacts.

Key Findings on Natural Selection

• The study identified roughly 479 positions with 99% confidence and ~3,800 positions with >50% confidence of having undergone directional selection in the last 10,000 years.
• Immune traits showed the strongest enrichment for selection signals (4-5 fold), followed by metabolic traits (obesity, fat distribution, type 2 diabetes risk).
• Behavioral and psychiatric traits (e.g., intelligence, years of schooling) showed no detectable enrichment in the strongest signals, though Reich notes this is likely due to the polygenic nature of these traits (many weak-effect genes) rather than a lack of selection.
• A specific genetic variant linked to intelligence and years of schooling shows strong selection peaks between 2,000 and 4,000 years ago, with the selection strength nearly doubling compared to other periods.

The Bronze Age Intensification

• Contrary to the assumption that the Neolithic transition to agriculture (11,000 years ago) was the primary driver of biological change, the data indicates the most intense period of natural selection occurred during the Bronze Age (5,000 to 2,000 years ago).
• This intensification correlates with a "wrenching" shift to high population densities, urbanization, and close proximity to domesticated animals, creating new selective pressures for immunity and metabolism.
• Specific examples of this shift include:
  • Lactase persistence: Strong selection for the ability to digest milk coinciding with the intensification of cattle herding.
  • TIK2 variant: A risk factor for tuberculosis that increased in frequency (likely due to protection against another pathogen) before declining sharply 3,000 years ago as TB became endemic.
  • Skin pigmentation: The strongest period of depigmentation in Europeans occurred between 4,000 and 2,000 years ago.
  • ABO blood groups: A shift in frequency favoring the B allele over the A allele occurred during this period.

Evolutionary Mismatch and Trade-offs

• The data suggests an "evolutionary mismatch" where hunter-gatherer genomes were不适应 to the new agricultural and Bronze Age environments, driving rapid adaptation.
• Selection against obesity and body fat mass has been consistent over the last 10,000 years, supporting the "thrifty gene" hypothesis in the context of stable agricultural food supplies compared to the boom-bust cycles of hunting.
• Reich notes that while intelligence selection peaked in the Bronze Age, there is no evidence of selection for such traits in the last 2,000 years, contradicting hypotheses that modern societal complexity requires further cognitive evolution.

Hypotheses on Archaic Human Relationships

• Reich proposes a controversial new model for the relationship between modern humans, Neanderthals, and Denisovans to explain genetic anomalies where mitochondrial DNA and Y-chromosomes cluster modern humans with Neanderthals, while the rest of the genome clusters Neanderthals with Denisovans.
• The hypothesis suggests a "Middle Stone Age" cultural revolution originating in the Near East/East Africa ~300,000 years ago that spread via a modern human population expansion.
• As this modern population expanded into Europe, it interbred with local archaic humans; due to patrilineal or matrilineal social structures and social selection, modern human mtDNA or Y-chromosomes swept to 100% frequency despite modern humans constituting only ~5% of the final archaic genome.
• This model posits that the common ancestor of Neanderthals and most modern humans is more recent than the divergence of Denisovans, implying Neanderthals may share a more recent cultural and potentially genetic lineage with modern humans than previously thought.

Historical Context of Human Variation

• Genetic data confirms that the "cognitive revolution" and behavioral modernity (art, complex tools) 50,000 years ago were not preceded by fixed genetic differences; the genetic toolkit for these behaviors existed 100,000+ years ago.
• The development of agriculture was delayed until the Holocene (last 12,000 years) primarily due to a unique period of climate stability, not a lack of genetic capability.
• Modern human genetic diversity is contained within a small ancestral population (~10,000 individuals) that existed ~70,000 years ago, which held enough latent variation to support rapid adaptation to diverse environments without requiring new mutations.

Future Research and Implications

• The ability to detect selection now depends on large sample sizes and the correction for migration/drift; small populations were previously mutation-limited, but large modern populations (8 billion) ensure every mutation occurs frequently enough for selection to act.
• The research suggests there is significant "room at the top" for many complex traits (height, immunity, cognition) to evolve further if environmental pressures shift, as the necessary genetic variation already exists in the pool.
• The study highlights that the "standard model" of human evolutionary history may be a collection of "epicycles" (patched-together explanations) and that a simpler model involving shared cultural-genetic expansions 300,000 years ago could better explain current genetic patterns.