The Neanderthal population history and the introgression landscape inferred from the UK Biobank

By analyzing 45,000 UK Biobank genomes, this study provides high-resolution insights into Neanderthal population history and introgression patterns, revealing ongoing selection on Neanderthal segments and identifying Human Accelerated Regions within Neanderthal deserts as potential targets of modern human adaptation.

The Gist

The Big Picture: A Genetic Time Capsule

Imagine that every non-African person alive today carries a tiny, invisible backpack. Inside this backpack are fragments of DNA inherited from our ancient cousins, the Neanderthals. These cousins interbred with our ancestors roughly 50,000 years ago.

For a long time, scientists could only peek at a few of these backpacks because they didn't have enough data. It was like trying to understand a library by reading just a few pages from three different books.

This new study is a game-changer. The researchers opened up 45,000 backpacks (using data from the UK Biobank) to see exactly what's inside. They didn't just count the Neanderthal DNA; they mapped it, measured it, and asked: What happened to this DNA after it entered the human family? Did it help us? Did it hurt us? And what does it tell us about our ancient cousins?

---

1. The "Neanderthal Desert" vs. The "Oasis"

Think of the human genome as a vast landscape.

• The Oases (Introgression): These are patches of land where Neanderthal DNA is thriving. In some spots, the Neanderthal DNA is so common it feels like it belongs there.
• The Deserts (Introgression Deserts): These are huge stretches of land where Neanderthal DNA is completely missing.

What the study found:
Even with 45,000 people, these "Deserts" are still there. They aren't just empty spots because we didn't look hard enough; they are empty because nature swept them clean.

• Why? Imagine Neanderthals had a lot of "bad software" (harmful genetic mutations) in their code because their population was small and isolated. When they mixed with humans, that bad code got dragged in. Natural selection acted like a janitor, quickly deleting those harmful chunks. This created the "Deserts."

2. The "Ghost" Neanderthals

The researchers found something surprising. Even with such a huge dataset, they kept finding new rare Neanderthal DNA fragments that no one had seen before.

• The Analogy: Imagine you are trying to find all the unique seashells on a beach. You've looked at 100 people's buckets, but every time you look at a new person, you find a shell you've never seen.
• The Lesson: The Neanderthal population was more diverse than we thought, and there are still "ghost" fragments of their DNA hiding in us, waiting to be discovered.

3. Who Were These Neanderthals?

By analyzing the DNA, the team built a family tree.

• The Match: The Neanderthals who mixed with humans were most closely related to the Vindija Neanderthals (found in Croatia).
• The Meeting Spot: They likely met in the Persian Plateau (modern-day Iran/Iraq area). It was like a busy train station where two different groups of travelers met, swapped stories (and genes), and then went their separate ways.
• The Population Size: The study estimated that the Neanderthal group meeting humans wasn't tiny and dying out; they were actually a decent-sized group (about 6,500 people). They weren't necessarily "weak," but perhaps their social groups were fragmented, making it hard for them to survive long-term.

4. The "Good" vs. The "Bad" Neanderthal DNA

The researchers looked at the DNA to see if it was helping or hurting us.

The "Bad" (Purifying Selection):
As mentioned, the "Deserts" are where the bad stuff was kicked out. These areas are often related to brain development and reproduction. Nature said, "No, we don't want that version of the gene."

The "Good" (Adaptive Introgression):
But some Neanderthal DNA stuck around because it was super useful. The study found 545 spots where Neanderthal DNA is being actively favored.

• The Analogy: It's like inheriting a winter coat from a great-grandparent. Even though it's old, you keep wearing it because it keeps you warm in the cold.
• What did it help with?
  • Immunity: Fighting off new viruses.
  • Stress: Handling environmental stress.
  • Appearance: Skin and hair traits.
  • Brain: Surprisingly, some Neanderthal DNA is linked to how our brains work, specifically in areas dealing with social behavior and glial cells (the support crew of the brain).

5. The "Human-Only" Brain Upgrade

One of the coolest findings is about the Cerebellum (the part of the brain at the back that handles balance, coordination, and complex thinking).

The study found that there are specific "Deserts" in our DNA where Neanderthal DNA is missing, but these spots are filled with Human Accelerated Regions (HARs).

• What are HARs? These are parts of the DNA that changed very fast in humans but stayed the same in almost every other animal (including Neanderthals).
• The Metaphor: Imagine Neanderthals had a standard model of a car engine. Humans took that engine and added a turbocharger and a new computer system. The study found that the "turbocharger" parts are located in the cerebellum.
• The Result: We purged the Neanderthal version of these genes because our human version was better. This suggests that the unique "human-ness" of our brains—our ability to think, plan, and balance—was a specific upgrade that happened after we split from Neanderthals.

Summary: What Does This All Mean?

This paper is like upgrading from a blurry, low-resolution photo of our past to a 4K high-definition movie.

1. We have more Neanderthal DNA than we thought: Even rare, hidden fragments are still being found.
2. Nature is a strict editor: It ruthlessly deleted the harmful Neanderthal DNA (creating deserts) but kept the helpful parts (creating oases).
3. We are unique: The study highlights that our brains, specifically the parts that make us "human," underwent rapid, specific changes that Neanderthals didn't have. We didn't just inherit their DNA; we actively edited it to make ourselves who we are today.

In short: We are a mosaic. We carry the scars of ancient battles (the deserts), the gifts of ancient allies (the oases), and the unique spark that makes us modern humans.

Technical Summary

1. Problem Statement

While Neanderthal introgression into non-African modern humans is well-established (~1-2% ancestry), current understanding of the introgression landscape and the demographic history of the introgressed lineage is limited by:

• Sample Size: Previous studies relied on modest sample sizes (typically ~500 genomes), limiting the detection of rare variants and the resolution of the Site Frequency Spectrum (SFS).
• Genomic Coverage: Analyses were often restricted to subsets of the genome or relied heavily on reference-based approaches that miss archaic variation absent from sequenced Neanderthal genomes.
• Demographic Inference: There is a lack of precise estimates regarding the effective population size ($N_e$) and divergence times of the specific Neanderthal lineage that admixed with modern humans, distinct from the sequenced high-coverage Neanderthals (Altai, Vindija, Chagyrskaya).
• Selection Dynamics: The extent of ongoing positive selection on introgressed alleles and the precise nature of "Neanderthal deserts" (regions depleted of archaic ancestry) remain to be fully characterized at a genome-wide scale.

2. Methodology

The authors leveraged the UK Biobank (UKBB), a massive dataset of ~45,000 imputed and phased genomes of European descent, to overcome previous limitations.

• Data Processing:

  • Selected 45,532 individuals with brain MRI phenotypes; pruned to 38,406 unrelated European samples.
  • Used SHAPEIT5 for phasing and GLIMPSE2 for imputation.
  • Merged data with 1000 Genomes Project data and three high-coverage Neanderthal genomes (Altai, Vindija, Chagyrskaya) plus a Denisova genome.

• Introgression Detection:

  • Employed SSTAR, a reference-free method using the $S^*$ statistic, to identify haplotypes with an enrichment of derived alleles relative to a Yoruba (YRI) reference. This approach captures archaic variation not present in currently sequenced Neanderthal genomes.
  • Applied quality control using f4-ratio statistics to estimate the proportion of modern human contamination in inferred Neanderthal segments.

• Demographic Modeling:

  • Constructed the Derived Allele Frequency (DAF) spectrum of variants within the introgressed Neanderthal segments.
  • Used fastsimcoal2 to infer demographic parameters. Crucially, they developed an innovative approach to explicitly incorporate the post-introgression demographic history of the modern European population (growth rates, timing) into the simulation of the introgressed Neanderthal lineage to avoid confounding effects.
  • Validated relationships using TREEMIX and qpGraph.

• Selection Scans:

  • Implemented a Hardy-Weinberg Equilibrium (HWE) test on introgressed haplotypes to detect regions with an excess of homozygotes, a signature of ongoing positive selection.
  • Filtered candidates to ensure deviations were driven by introgressed alleles, not linked non-introgressed variants.

• Desert Analysis:

  • Defined "Neanderthal deserts" as 10-kb windows with near-zero introgression rates.
  • Classified gaps into "strict" (containing Neanderthal-specific SNPs) and "ILS" (Incomplete Lineage Sorting, lacking specific SNPs).
  • Analyzed Time to Most Recent Common Ancestor (TMRCA) within modern humans and between modern humans and Neanderthals to distinguish ancient purifying selection from recent human-specific sweeps.
  • Overlapped results with Human Accelerated Regions (HARs) and Gene Ontology (GO) terms.

3. Key Contributions

• Scale: The study analyzes ~45,000 genomes, a ~90-fold increase over typical previous studies, allowing for the detection of rare Neanderthal variants and a near-saturation of the introgression landscape.
• Reference-Free Discovery: By using SSTAR, the authors identified ~50,680 additional Neanderthal-derived variants that are absent from the three high-coverage Neanderthal genomes but strongly associated with introgressed haplotypes.
• Demographic Refinement: Provided the first precise estimates of the effective population size and divergence time of the specific Neanderthal lineage that contributed to modern Europeans, distinct from the sequenced reference individuals.
• Selection Detection: Identified 545 independent loci showing signatures of ongoing positive selection via HWE distortion, expanding the known landscape of adaptive introgression.
• Desert Characterization: Refined the map of Neanderthal deserts, demonstrating they are not artifacts of small sample sizes, and identified a subset of deserts driven by recent, human-specific selective sweeps.

4. Key Results

A. Introgression Landscape and Rare Variants

• The study recovered 5,258,470 polymorphic positions covered by Neanderthal segments.
• The discovery rate of new variants has not plateaued, indicating that very rare introgressed variants remain undetected even at this scale.
• The average Neanderthal ancestry proportion is 1.11% ± 0.09%.
• Validation via f4-ratio showed minimal false positives (max ~3.7% modern human contamination in specific batches).

B. Demographic History of the Introgressed Lineage

• Divergence Time: The introgressed Neanderthal lineage diverged from the Vindija Neanderthal approximately 2,061 generations ago (61.8 kya, assuming 30-year generations). This predates the introgression event (55.6 kya), confirming Vindija as a near-direct representative of the admixing population.
• Effective Population Size ($N_e$): The $N_e$ of the introgressed lineage prior to introgression was estimated at 6,564. This is comparable to the $N_e$ of modern humans during the Out-of-Africa expansion (~3,124), suggesting Neanderthals were not demographically negligible compared to early modern humans.
• Bottleneck: A prolonged bottleneck occurred during introgression and subsequent purging, involving as few as 328 individuals.

C. Ongoing Selection (HWE Analysis)

• Identified 545 independent loci with significant excess of Neanderthal homozygotes ($p < 10^{-8}$), indicative of ongoing positive selection.
• Functional Enrichment: These loci are enriched for:
  • Immune regulation (B cell activation).
  • Cellular stress response.
  • Craniofacial and musculoskeletal morphology.
  • Neural/behavioral phenotypes (glial cell apoptosis, social withdrawal).
• Notable genes include SLCO1B1 (liver), OR9G1, PPDPF, STARD9, and NECAB1.

D. Neanderthal Deserts and Human-Specific Adaptation

• Desert Extent: Even with 45k samples, 35.1% of the genome remains devoid of Neanderthal introgression (expanded gaps), confirming these are biological realities, not sampling artifacts.
• HARs in Deserts: Seven Human Accelerated Regions (HARs) were found within deserts that have a recent TMRCA within modern humans (<650 kya) but an ancient TMRCA relative to Neanderthals. This suggests these regions underwent strong, recent selective sweeps in modern humans that purged Neanderthal variation.
• Cerebellum Connection: Four of these HARs are located in introns of genes highly expressed in the cerebellum (RBFOX2, CADM1, BACH2, TCF4). This supports the hypothesis that modern human-specific adaptations in brain morphology (specifically cerebellar expansion) were driven by selective sweeps that removed archaic alleles.
• GWAS Traits: Gaps with recent human-specific sweeps are enriched for traits related to cognition, body size, and lipid metabolism.

5. Significance

• Demographic Insight: The study resolves the demographic history of the specific Neanderthal population that admixed with modern humans, showing it was a relatively large, structured population closely related to Vindija, rather than a small, isolated group.
• Evolutionary Dynamics: It demonstrates that the legacy of Neanderthal introgression is dynamic. While purifying selection removed deleterious archaic alleles (creating deserts), positive selection has actively maintained and driven to high frequency specific adaptive alleles (e.g., in immunity and skin/hair).
• Human Specificity: The identification of recent selective sweeps in the cerebellum-associated genes within Neanderthal deserts provides strong genetic evidence for the evolution of modern human brain structure and cognitive functions post-admixture.
• Methodological Benchmark: The study establishes a framework for using massive biobank datasets to reconstruct archaic population history and detect subtle selection signals that were previously invisible to smaller-scale studies.