Rethinking human AMY1 copy number evolution in light of demographic history

By integrating new data from 390 individuals across 30 Sub-Saharan African populations with global datasets and employing ancestry-aware models, this study challenges the long-held "agriculture hypothesis" by demonstrating that demographic history, rather than dietary shifts to starch-rich farming, is the primary driver of human AMY1 copy number evolution.

The Gist

The Big Idea: It's Not About What You Eat, It's About Your Family Tree

For a long time, scientists believed a specific story about human evolution: When humans started farming and eating lots of starchy foods (like wheat, rice, and corn), our bodies adapted by growing more copies of a gene called AMY1.

Think of AMY1 as a factory that produces "starch-digesting workers" (an enzyme called amylase). The old theory was: More starch in the diet = The body builds more factories to handle the load.

This new paper says: "Actually, that story is mostly wrong."

The researchers found that the number of these "factories" in our bodies isn't really determined by whether your ancestors were farmers or hunter-gatherers. Instead, it's determined by who your ancestors were and where they came from.

---

The Investigation: A Global Detective Story

The authors (a huge team of scientists from Africa, Europe, and the US) decided to re-investigate this mystery.

1. The Old Clues: Previous studies mostly looked at a few groups of people, mostly from Europe and Asia. It was like trying to understand the weather in the whole world by only looking at a few cities in France.
2. The New Evidence: They went out and collected brand-new data from 390 people representing 30 different groups across Sub-Saharan Africa. They combined this with existing data from over 1,300 people worldwide.
3. The Method: They used a high-tech tool called "droplet digital PCR" (imagine a super-precise DNA counter) to count exactly how many copies of the AMY1 gene each person had.

The Findings: The "Family Tree" vs. The "Menu"

When they crunched the numbers, they found two major things:

1. The "Menu" Doesn't Matter (As Much as We Thought)

If you look at people in Africa, whether they are farmers, hunters, gatherers, or herders, there is no consistent link between their diet and their gene count.

• The Analogy: Imagine a school cafeteria. The old theory said, "If the cafeteria serves pizza every day, the students will grow extra stomachs." This study says, "Nope. The students who have extra stomachs are just the ones whose parents had extra stomachs, regardless of what they ate today."

2. The "Family Tree" is the Real Boss

The strongest predictor of how many AMY1 copies you have is your genetic ancestry.

• The Pattern: People with deep roots in Sub-Saharan Africa tend to have fewer copies on average. People with ancestry from Europe, East Asia, or the Americas tend to have more copies.
• The Analogy: Think of AMY1 copies like a family heirloom, such as a specific type of watch. If your great-grandfather had three watches, you are likely to have three, regardless of whether you are a banker (farmer) or a fisherman (hunter-gatherer). The number of watches is about inheritance, not your job.

The "Out-of-Africa" Bottleneck

The paper suggests that the differences we see today are actually the result of ancient history, not recent farming.

• The Bottleneck: When a small group of humans left Africa to populate the rest of the world thousands of years ago, they carried only a specific set of genes with them.
• The Analogy: Imagine a giant jar of marbles (representing all human genes). A small cup was scooped out to take to the rest of the world. By pure chance, that cup had more "high-copy" marbles than the jar left behind in Africa. As those people spread out, they kept those extra copies. It wasn't because they started eating more starch; it was just genetic luck during a long journey.

Why Did We Get It Wrong Before?

The paper explains that earlier studies were like looking at a blurry photo. Because they didn't have enough data from Africa, they accidentally mixed up "ancestry" with "diet."

• They saw that Europeans (who are farmers) had high gene counts.
• They saw that some African groups (who were hunter-gatherers) had low gene counts.
• They assumed the diet caused the difference.
• Reality: The difference was caused by the ancestry. If you compare a European farmer to an African farmer, the gene counts are actually much more similar than previously thought.

The Takeaway

This study is a reminder that evolution is complicated.

While our bodies do adapt to our environment, this specific gene (AMY1) isn't a simple "switch" that flips on when we start farming. Instead, the variation we see today is mostly a fossil record of our ancient migrations.

In short: Don't blame your diet for your gene count; blame your great-great-great-grandparents' travel itinerary!

Technical Summary

1. Problem Statement

The prevailing hypothesis in human evolutionary genomics posits that the increased copy number (CN) of the salivary amylase gene (AMY1) is a direct adaptation to high-starch agricultural diets. This "agriculture hypothesis" suggests that as humans transitioned to farming (~12,000 years ago), natural selection favored individuals with higher AMY1 CN to improve starch digestion.

However, previous evidence supporting this model suffers from two critical limitations:

1. Limited Sampling: Most studies relied on small datasets with poor representation of Sub-Saharan African populations, the cradle of human diversity.
2. Confounding Factors: Analyses often failed to adequately control for shared ancestry and population structure. Consequently, observed correlations between diet and AMY1 CN may be artifacts of demographic history (e.g., gene flow, drift) rather than direct selection.

2. Methodology

The authors employed a comprehensive, ancestry-aware approach to re-evaluate AMY1 evolution using the largest dataset assembled to date.

• Data Generation & Integration:

  • New Data: Generated droplet digital PCR (ddPCR) estimates for AMY1 CN in 390 individuals from 30 diverse Sub-Saharan African populations.
  • Global Integration: Combined new data with published read-depth estimates from the Human Genome Diversity Project (HGDP), Simons Genome Diversity Project (SGDP), and 1000 Genomes Project (KGP), totaling up to 1,307 individuals from 86 global populations.
  • Ancient DNA: Included 533 ancient European genomes to analyze temporal trends over the last 12,000 years.
  • Subsistence Annotation: Populations were annotated with detailed subsistence strategies (agriculture, pastoralism, hunting, gathering, fishing) using the D-PLACE database.

• Statistical Modeling:

  • Ancestry-Aware Frameworks: The study utilized two complementary regression frameworks to disentangle diet from demography:
    1. Bayesian Models (brms): Incorporated individual-level phylogenetic autocorrelation via a variance-covariance matrix derived from a phylogenetic tree.
    2. Frequentist Models (glmmTMB): Controlled for population structure using genome-wide Principal Components (PCs) as fixed effects.
  • Model Comparisons: Tested null models (ancestry only) against models including subsistence predictors (binary categories and continuous dependence measures).
  • Phylogenetic Signal Analysis: Used global statistics, correlograms, and Local Indicators of Phylogenetic Association (LIPA) to detect clustering of traits across the human phylogeny.

3. Key Contributions

• Largest African Dataset: Provided the first high-resolution AMY1 CN dataset specifically for diverse Sub-Saharan African populations, addressing a major geographic gap in previous literature.
• Methodological Rigor: Demonstrated the necessity of using phylogenetically informed and ancestry-aware models to avoid spurious correlations between diet and genetic traits.
• Paradigm Shift: Challenged the universality of the "agriculture hypothesis" by showing that demographic history is a stronger predictor of AMY1 CN than subsistence strategy.

4. Key Results

• No Association in Africa:

  • Across 390 individuals from 30 African populations, no consistent association was found between AMY1 CN and agriculture once population structure was accounted for.
  • Genetic ancestry (specifically principal components) was the primary predictor of variation. For instance, African populations with higher Eurasian ancestry (e.g., Ethiopians) showed higher AMY1 CN, regardless of their subsistence strategy.

• Global Scale Findings:

  • When analyzing the global dataset (1,307 individuals), a statistically significant but modest difference (0.53 diploid copies) was observed between agriculturalists and non-agriculturalists.
  • However, Bayesian effect-size assessments indicated that the contribution of agriculture to global variation was negligible once ancestry was controlled. The effect size was substantially smaller than previously reported (dropping from ~1.28 copies in early studies to 0.53 in this study).
  • Subsistence Specifics: Among agriculturalists globally, dependence on gathering showed a weak positive association, while fishing showed a negative association, but these effects were small and context-dependent.

• Phylogenetic Structure & Baseline Differences:

  • Deep Demographic Split: Phylogenetic analyses revealed distinct baseline distributions: Sub-Saharan African populations generally exhibit lower AMY1 CN, while non-Sub-Saharan (Eurasian, American, East Asian) populations exhibit higher baselines.
  • Phylogenetic Signal: Strong phylogenetic autocorrelation was detected in non-African clades but was absent in Sub-Saharan African populations, suggesting that the variation is driven by deep demographic processes (e.g., serial bottlenecks during the Out-of-Africa expansion) rather than recent local adaptation.
  • Ancient Europe: In ancient European genomes, AMY1 CN increased over time, but this correlated more strongly with ancestry turnover (e.g., the arrival of Anatolian farmers) than with sample age alone, suggesting demographic replacement rather than direct selection on diet.

5. Significance

• Refutation of a Canonical Example: This study fundamentally revises the understanding of AMY1 evolution, demonstrating that it is not a straightforward signature of agricultural adaptation. The previously cited correlation was largely an artifact of limited sampling and uncorrected population structure.
• Demography as the Driver: The primary driver of global AMY1 CN variation is demographic history, including the Out-of-Africa dispersal, serial bottlenecks, and subsequent gene flow. The high mutation rate of the AMY1 locus (via non-homologous recombination) combined with these demographic events creates stable, population-level differences that mimic selection.
• Broader Implications for Evolutionary Genomics: The paper underscores the critical need for:
  1. Comprehensive Geographic Sampling: Particularly from Africa, to avoid bias.
  2. Ancestry-Aware Analyses: To distinguish between true selective signals and neutral demographic effects, especially for loci with extreme structural variability.
• Functional Complexity: The results suggest that AMY1 CN variation may be influenced by factors beyond starch digestion, such as immunity, stress response, or oral health, given the gene's pleiotropic roles.

In conclusion, the authors argue that AMY1 CN diversity reflects a complex interplay of mutation dynamics and deep demographic history, rather than a simple, universal adaptation to the Neolithic agricultural transition.