Subsistence transition preceded population turnover in the eastern Colombian Andes

Genome-wide analysis of 209 ancient individuals from the eastern Colombian Andes reveals that a subsistence transition to maize cultivation occurred without population replacement among long-standing local hunter-gatherers, while a distinct genetic turnover associated with the emergence of the Herrera culture approximately 2,200 years ago was driven by the migration of new groups with mixed Amazonian and Chibchan-related ancestry.

The Gist

Imagine the Eastern Colombian Andes as a grand, high-altitude stage where a play has been running for over 7,000 years. For a long time, historians and archaeologists were trying to figure out the plot: Did the actors stay the same while the scenery changed? Or did the entire cast get replaced by a new troupe?

This paper uses a powerful new tool—ancient DNA—to rewrite the script. It reveals a story of two distinct acts: a long, quiet period of continuity, followed by a dramatic, sudden cast change.

Here is the story in simple terms:

Act 1: The Long, Quiet Stew (7,000 to 2,800 years ago)

For over 5,000 years, the people living on this high plateau (the Sabana de Bogotá) were essentially the same family tree. They were hunter-gatherers who lived in small, tight-knit groups.

• The "Slow Cook" Metaphor: Think of their genetic makeup like a pot of soup simmering on a low flame. Over thousands of years, the flavor changed slightly because the pot was small and the ingredients (genes) were just drifting around, but no new ingredients were added from outside. They were isolated, but stable.
• The Surprise Twist: Even though they were still genetically the same "hunter-gatherer" family, they started eating something new. About 2,800 years ago, two individuals from this group were found to have eaten maize (corn).
  • The Analogy: Imagine a family that has been eating only wild berries and deer for centuries. Suddenly, they start eating corn. In the past, scientists thought, "Oh, they must have been replaced by farmers who brought the corn." But this study says: No. The same family that had been eating berries for 5,000 years just decided to start farming corn. They adopted the new food without being replaced by new people.

Act 2: The Great Cast Swap (Around 2,200 years ago)

Then, around 2,200 years ago, the plot took a sharp turn. The "Herrera Period" began, marked by bigger villages, pottery, and more complex farming.

• The "Overnight Replacement" Metaphor: Imagine walking into a theater and seeing the same actors from Act 1, but then, suddenly, the curtain rises on Act 2, and every single actor has been swapped out. The new actors look different, speak a different "dialect" (genetically), and bring a completely new energy.
• Who were the new actors? The new people were a mix of two groups:
  1. People related to Chibchan speakers (who came from what is now Central America/Costa Rica).
  2. People related to groups living at the edge of the Amazon rainforest.
  • The Mix: These two groups had already mixed together about 4,000 years ago, creating a new "genetic cocktail." This new cocktail then migrated up into the high Andes, replacing the old hunter-gatherer population almost entirely.

The Key Takeaways

1. Food didn't bring the new people: The biggest surprise is that the local people started eating maize before the new people arrived. The locals adopted the new diet on their own. The "farmers" didn't bring the "food"; the "food" came first, and the "new people" came later.
2. The Great Reset: The transition from the old hunter-gatherers to the complex Muisca chiefdoms wasn't a slow evolution. It was a population turnover. The old genetic line was largely wiped out and replaced by the new mixed group.
3. Small Towns, Big Changes: Even though the local groups were small (like a village of 100-130 people), they managed to survive for 5,000 years without outside interference. But when the new group arrived, they didn't just blend in; they took over the stage.

Why does this matter?

This study solves a mystery that has puzzled scientists for decades. It shows that in human history, changing your diet doesn't always mean changing your people. Sometimes, the locals just learn a new trick (farming). But sometimes, the stage gets completely cleared, and a whole new cast arrives to take over the show.

In the Eastern Colombian Andes, we saw both: a long period where the locals learned to farm corn, followed by a dramatic moment where a new, mixed population arrived and built the famous Muisca civilization we know from history.

Technical Summary

1. Problem Statement

The Altiplano Cundiboyacense in the Eastern Cordillera of the Northern Andes (Colombia) possesses one of the most continuous archaeological sequences in the Americas, spanning from late Pleistocene hunter-gatherers to the complex Muisca chiefdoms. However, the demographic history of this region remained poorly understood due to a lack of ancient DNA (aDNA) data, particularly for the critical period between 6,000 and 2,000 years before present (BP).

Previous studies identified a major genetic turnover distinguishing early Holocene groups from the later Herrera and Muisca Period populations but could not resolve:

• Timing: Whether the ancestry shift was gradual or abrupt.
• Causality: Whether the shift coincided with the initial adoption of maize (C₄ plants) or the later emergence of the sedentary Herrera culture.
• Mechanism: Whether the transition involved population replacement or in-situ cultural adaptation.

Three competing hypotheses existed: (1) a middle Holocene shift (6000–4000 BP), (2) an initial late Holocene shift (4000–2500 BP) linked to early maize consumption, or (3) a later shift (~2200 BP) associated with the Herrera migration.

2. Methodology

The study employed a comprehensive paleogenomic approach to reconstruct population history over 7,000 years.

• Sample Collection & Dating: The authors generated genome-wide data from 209 ancient individuals (enriching the regional dataset 11-fold) from the Altiplano Cundiboyacense. This included 37 new direct radiocarbon ($^{14}$C) dates on bone samples.
• Sequencing & Bioinformatics: DNA was extracted from skeletal elements (cochlea, teeth, bone) and enriched for ~1.4 million SNPs (using the Twist Ancient DNA assay or 1240k capture). Data were processed using standard aDNA pipelines (adapter trimming, mapping to hg19/RSRS, duplicate removal, and pseudohaploid calling).
• Dietary Analysis: Stable isotope analysis ($\delta^{13}$C and $\delta^{15}$N) was performed on collagen from 33 individuals to track subsistence shifts, specifically the incorporation of C₄ plants (maize).
• Statistical Analysis:
  • PCA (Principal Component Analysis): To visualize broad genetic clusters relative to modern and ancient Native Americans.
  • qpWave/qpAdm: To test genetic homogeneity, cladality, and model ancestry proportions (admixture modeling).
  • f-statistics (f3, f4): To measure shared genetic drift and allele sharing asymmetries.
  • DATES: To estimate the timing of admixture events.
  • hapROH: To estimate effective population sizes ($N_e$) and detect runs of homozygosity (inbreeding).
  • mtDNA & Y-chromosome: Haplogroup calling to track maternal and paternal lineage shifts.

3. Key Contributions

• Temporal Resolution: Filled a critical 4,000-year sampling gap (6000–2000 BP) in the region's genetic record.
• Decoupling Subsistence and Ancestry: Provided definitive genomic evidence that the adoption of maize agriculture occurred before the major population turnover, challenging the assumption that agricultural shifts in the Andes were driven solely by migration.
• Ancestry Reconstruction: Identified the specific dual-ancestry composition of the Herrera-Muisca populations, revealing a mixture of Central American Chibchan-related groups and a distinct "Amazonian North" component.

4. Key Results

A. Five Millennia of Genetic Continuity (Pre-Herrera)

• Hunter-Gatherer Lineage: Individuals from the early-middle Holocene (10,000 BP) through the initial late Holocene (2500 BP) form a distinct genetic clade ("H-G Cluster").
• Drift vs. Migration: Despite spanning 5,000+ years, these groups show no evidence of external gene flow. Sub-structure observed is attributed to genetic drift in small, isolated communities (estimated effective population size $N_e \approx 110-133$).
• Subsistence Shift Without Replacement: Two individuals from the Aguazuque site (~2800 BP), genetically indistinguishable from the local hunter-gatherer lineage, exhibit isotopic signatures ($\delta^{13}$C) indicating significant maize (C₄) consumption. This proves that maize was incorporated into the diet by resident populations without population replacement.

B. Abrupt Population Turnover (~2800–2000 BP)

• Genetic Break: By ~2000 BP (Herrera Period), the local gene pool was almost entirely replaced. The "H-G Cluster" disappears, replaced by a distinct "Herrera-Muisca Cluster."
• No Admixture Signal: Statistical tests (f4-statistics) show no significant excess allele sharing between pre-Herrera and Herrera-Muisca groups, suggesting a near-total replacement rather than gradual admixture.
• mtDNA Shift: A sharp decline in haplogroup C and the appearance of haplogroup D and the derived A2ac1 lineage (dominant in Herrera-Muisca) support a demographic shift.

C. Origins of the Herrera-Muisca Ancestry

• Dual Ancestry Model: The Herrera-Muisca populations are best modeled as a mixture (~4000 BP) of two sources:
  1. Chibchan-related ancestry: Proxied by ancient Panamanians and modern Cabécar (Central America).
  2. Amazonian/Andean-related ancestry: Proxied by modern "Amazonian North" groups (e.g., Kichwa Orellana) living on the eastern slopes of the Andes.
• Admixture Timing: DATES analysis estimates the admixture event occurred approximately 3964 ± 446 years ago, predating the arrival of these groups on the Altiplano Cundiboyacense.
• Regional Substructure: Within the Herrera-Muisca period, there is subtle genetic differentiation between the southern (Cundinamarca) and northern (Boyacá) regions, with the north showing stronger Chibchan affinity and the south showing stronger "Amazonian North" affinity.

D. Demographic Expansion

• Effective population sizes increased significantly during the Herrera (310–438) and Muisca (513–601) periods, aligning with archaeological evidence of sedentism, population growth, and social stratification.
• While close-kin unions were rare, specific instances of consanguinity were detected in the Muisca period, consistent with complex social structures.

5. Significance

• Redefining Andean Transitions: The study overturns the "demic diffusion" model for the initial adoption of maize in the Eastern Colombian Andes. It demonstrates that subsistence change (agriculture) preceded population turnover, indicating that local hunter-gatherers actively adopted new crops before being replaced by a new migrating population.
• Migration Routes: The genetic profile of the incoming Herrera-Muisca people suggests a complex migration history involving a pre-existing admixture between Chibchan speakers (from Central America) and Amazonian/Andean groups. This likely occurred outside the Altiplano Cundiboyacense, possibly along the Magdalena River Valley or the eastern Andean foothills, before expanding onto the plateau.
• Cultural vs. Biological Change: The paper highlights that major cultural shifts (sedentism, chiefdoms) can occur via population replacement, while earlier economic shifts (dietary diversification) can occur via cultural transmission within continuous populations.
• Methodological Benchmark: By increasing the sample size 11-fold and integrating isotopic data, this study sets a new standard for resolving complex demographic histories in the Americas where archaeological and genetic records have previously been in conflict.