Genomic sampling and population structure of farmer-maintained varieties reveal previously uncharacterized diversity of Theobroma cacao L. in Costa Rica

This study utilizes whole-genome resequencing of 94 farmer-maintained cacao trees across Costa Rica to reveal a complex genetic landscape containing both Criollo-derived lineages and unique local diversity, highlighting the critical role of smallholder farming systems as reservoirs for the crop's conservation and future breeding.

The Gist

Imagine the world of chocolate as a massive, ancient library. For decades, scientists have been cataloging the books in this library, but they've mostly focused on the famous bestsellers from South America. They knew very little about the unique, local stories written in the "backrooms" of Costa Rica, where farmers have been growing cacao trees for generations, often keeping secrets in their backyards.

This paper is like a team of detectives (the researchers) going into those Costa Rican backyards with a high-tech magnifying glass (whole genome sequencing) to read the genetic "stories" of 94 cacao trees. Here is what they found, translated into everyday language:

1. The "Lost" Royal Family (Criollo)

For a long time, people thought the "Royal Family" of cacao, known as Criollo, was almost extinct. They were the original, high-quality chocolate trees prized for their amazing flavor but were weak against disease. Farmers replaced them with hardier, but less tasty, "common" trees.

The Discovery: The researchers found that the Royal Family isn't dead; they are just hiding! They found three pure "Royal" trees still growing on small farms in Costa Rica. It's like finding a few original, handwritten first editions of a classic novel hidden in a attic, surrounded by thousands of mass-produced paperbacks. These rare trees are a treasure trove for making the world's best chocolate.

2. The Great Genetic Melting Pot

Most of the trees the scientists looked at weren't pure "Royal" or pure "Common." They were mixes.

The Analogy: Imagine a potluck dinner where everyone brings a dish. Instead of just one type of food, you have a giant stew where ingredients from different countries have been tossed together over centuries.

• The Result: The Costa Rican cacao is a "genetic stew." It's a complex mix of known types (like Amelonado) and some brand-new, previously unknown types that the scientists have never seen before. They named these new groups "CR1" through "CR6." It's like discovering that your family recipe for soup actually contains secret ingredients from three different continents that no one knew about.

3. The "Seed-Saving" Social Network

You might expect that trees in the north of Costa Rica would look different from trees in the south, just like people in different towns might have different accents. But the scientists found no geographic pattern.

The Metaphor: Think of the cacao trees not as neighbors who stay in their own houses, but as people at a massive, chaotic party where everyone is swapping recipes and DNA.

• Farmers have been saving seeds, cloning their favorite trees, and trading cuttings with friends across the country for decades.
• A tree in the far north might be a "cousin" to a tree in the far south because a farmer traded a branch with a friend 20 years ago. The genetic map looks like a tangled web of connections rather than neat, separate neighborhoods.

4. The "Identity Crisis" of Global Cacao

The scientists compared their Costa Rican trees to the "Global Cacao Database" (the reference books scientists use to identify trees).

The Problem: Many of the Costa Rican trees didn't fit into the existing categories. It's like trying to sort a pile of unique, hand-painted rocks into boxes labeled "Red," "Blue," and "Green," only to find that half the rocks are purple, orange, or have patterns that don't match any box.

• This means our current map of cacao diversity is incomplete. We need to update our "library catalog" to include these new, unique Costa Rican varieties.

5. Why This Matters (The "Insurance Policy")

Why should we care about these messy, mixed-up trees?

The Analogy: Imagine you are building a house. If you only use one type of brick (modern, uniform trees), your house might be strong today, but if a new storm (climate change or a new disease) hits, the whole house could collapse because every brick reacts the same way.

• The Costa Rican farmers' trees are like a giant insurance policy. Because they are so diverse and mixed, some of them might have hidden superpowers—like resistance to heat, drought, or a new bug—that the uniform "modern" trees lack.
• By studying these trees, scientists can find the "secret ingredients" needed to breed future chocolate trees that can survive a changing climate.

The Bottom Line

This paper tells us that Costa Rican farmers are the unsung heroes of cacao conservation. They have been unknowingly preserving a massive, complex, and unique genetic library in their backyards. By listening to these farmers and studying their trees, we can save the "Royal Family" (Criollo), discover new super-trees, and ensure that chocolate remains a delicious and resilient treat for the future.

Technical Summary

1. Problem Statement

Despite the global economic and cultural importance of cacao (Theobroma cacao L.), genomic resources remain heavily biased toward South American accessions (the center of origin) and specific reference collections. Central American cacao, particularly in Costa Rica, is historically significant for fine-flavor production and Indigenous heritage but is severely underrepresented in genomic studies.

• Knowledge Gap: There is a lack of high-resolution data on farmer-maintained germplasm in Costa Rica. It is unclear whether "Ancient" (unadmixed) Criollo lineages persist or if the region harbors unique, locally adapted genetic groups distinct from the 10 major global genetic clusters previously defined.
• Conservation Concern: Traditional landraces are at risk of displacement by globally sourced "improved" clones, potentially leading to the loss of unique alleles crucial for disease resistance and climate resilience.

2. Methodology

The study employed a comprehensive whole-genome resequencing approach to characterize genetic diversity across 94 trees sampled from 17 farms in four distinct regions of Costa Rica (Guatuso-Upala, Guácimo, Talamanca, and Puntarenas).

• Sampling Strategy: Trees were selected based on farmer identification of local adaptation, historic ties, or Criollo heritage.
• Sequencing:
  • Library Prep: Nextera tagmentation protocol using a homebrewed Tn5 transposome.
  • Sequencing: Illumina NovaSeqX (2x150 bp), generating an average of 12.9 Gbp per sample.
  • Reference Panel: 94 new Costa Rican samples were analyzed alongside 166 previously published reference accessions representing 10+ major genetic groups (e.g., Criollo, Amelonado, Iquitos, Contamana).
• Bioinformatics Pipeline:
  • Alignment: Reads mapped to the B97-61/B2 Criollo reference genome (v2) using bwa mem2.
  • Variant Calling: SNPs called with BCFtools and filtered for quality, missing data (<30%), and minor allele frequency.
  • Pruning: Linkage disequilibrium (LD) pruning performed using PLINK and PopLDdecay.
• Analytical Approaches:
  • Population Structure: sNMF (in R package LEA) to estimate ancestry coefficients ($K=1$ to $25$) and identify pure lines (MaxQ $\ge$ 0.8).
  • Phylogenetics: Maximum Likelihood trees (RAxML-NG) and phylogenetic networks (NeighborNet in SplitsTree) to visualize reticulate evolution and admixture.
  • Diversity Metrics: Calculation of observed heterozygosity ($H_o$), inbreeding coefficients ($F$), and pairwise kinship (KING-robust).
  • Spatial Analysis: Isolation by Distance (Mantel tests), AMOVA, and Spatial AMOVA (SAMOVA) to assess geographic structuring.

3. Key Contributions

• First High-Resolution Genomic Survey of Costa Rican Cacao: This study provides the first whole-genome resequencing dataset specifically targeting farmer-maintained, heritage cacao in Costa Rica.
• Discovery of Novel Genetic Groups: The analysis identified six previously uncharacterized ancestral lineages (designated CR1–CR6) that are distinct from the standard global reference groups.
• Validation of "Ancient" Criollo: The study confirmed the presence of genetically pure (unadmixed) Ancient Criollo trees within farmer plots, challenging the narrative that this lineage is extinct in cultivation.
• Refinement of Global Population Structure: The findings suggest that global cacao population structure is more reticulate and fluid than previously thought, with many reference accessions showing unexpected admixture when analyzed with broader, localized datasets.

4. Key Results

A. Genetic Diversity and Heterozygosity

• Heterozygosity Range: Observed heterozygosity ($H_o$) varied drastically, from near-zero (0.00068) in specific samples to 0.01350.
• Inbreeding: Samples with extremely low heterozygosity corresponded to high inbreeding coefficients ($F \approx 0.8$), consistent with pure Criollo lineages. Conversely, the majority of samples showed negative $F$ values, indicating high heterozygosity and outcrossing.
• Duplication: Kinship analysis revealed several duplicate accessions (clones) across different farms (e.g., CCN-10, Matina1-6, TSH-565), confirming that farmers propagate and share specific elite genotypes across the country.

B. Population Structure and Ancestry

• Global Context ($K=17$): When compared to global references, Costa Rican samples clustered into known groups (Criollo, Amelonado) but also formed distinct clusters.
• Novel Lineages (CR1–CR6):
  • CR1: Associated with United Fruit (UF) and Imperial College Selection (ICS) accessions.
  • CR2: Distinct from known references, containing accessions like UF-613.
  • CR3: A distinct group formed by Trinidad Select Hybrid (TSH) accessions (TSH-565, TSH-615).
  • CR4: A pervasive, uncharacterized lineage split into three sub-groups (CR4-A, B, C) upon finer analysis ($K=9$). It was present in 90% of Costa Rican samples as an admixed component.
  • CR5 & CR6: Unique to Costa Rican samples, with CR5 clustering around CCN-10 and CR6 representing a distinct pure lineage (UPA1-3, TAL9-1).
• Criollo Persistence: Three samples were identified as pure Ancient Criollo (GCM11-4, PUNT17-1, and PUNT17-1F2). While rare, their presence confirms the survival of this lineage.

C. Spatial and Landscape Patterns

• No Isolation by Distance (IBD): Mantel tests revealed no significant correlation between genetic and geographic distance ($r = 0.05, p = 0.11$).
• Human-Mediated Gene Flow: The lack of geographic structuring, combined with the presence of identical genotypes in geographically distant regions (e.g., Upala and Talamanca), indicates that gene flow is driven primarily by human activities (seed saving, clonal propagation, and farmer-to-farmer exchange) rather than natural dispersal or landscape barriers.
• AMOVA: The majority of genetic variation (82–88%) was found within farms, with only small proportions attributed to differences between farms or regions.

D. Phylogenetic Insights

• Phylogenetic networks revealed a complex, reticulate structure rather than a strictly bifurcating tree.
• Admixed individuals occupied intermediate positions between major clusters, confirming extensive historical and ongoing hybridization.
• Some reference accessions previously considered "pure" (e.g., certain Napo or Apaporis accessions) showed significant admixture in this broader context.

5. Significance and Implications

• Conservation Value: Farmer-managed systems in Costa Rica act as critical reservoirs for genetic diversity, including rare Ancient Criollo and unique local lineages (CR1–CR6) not captured in global genebanks.
• Breeding Resources: The identified unique genetic groups and the high heterozygosity in many local varieties offer a rich source of alleles for breeding programs focused on disease resistance, climate resilience, and fine-flavor traits.
• Re-evaluation of Cacao Taxonomy: The study suggests that the rigid classification of cacao into 10 major groups may be insufficient. The fluid boundaries and pervasive admixture indicate a need for a more dynamic, network-based understanding of cacao population structure.
• Policy and Management: The findings underscore the importance of supporting in-situ conservation of farmer varieties. Displacing these landraces with uniform commercial clones could result in the irreversible loss of locally adapted genetic resources essential for future food security.

In conclusion, this paper demonstrates that Costa Rican cacao is a complex mosaic of known global lineages and unique, locally evolved genetic groups, maintained through active farmer selection and exchange, serving as a vital resource for the future resilience of the species.