New tool gives arabica breeders a faster, cheaper way to select better coffee trees
A genotyping panel with 5,000 markers developed by World Coffee Research and partners in five countries helps researchers use genomic selection for more effective breeding.
Key findings
- The first openly accessible genotyping panel for arabica. Breeders now have a shared, cost-effective way to genotype young plants for breeding: 5,053 single nucleotide polymorphism (SNP) markers that could be used for molecular breeding applications to predict which young seedlings are most promising vs. waiting years for field evaluation.
- Reliable, genome-wide coverage. Markers passed quality control, reliably covering 98.7% of the genome across both of arabica’s subgenomes.
- More detail on diversity and inheritance. The panel separates wild from improved material and introgressed material (containing genes borrowed from other coffee species, which is especially relevant for disease resistance), even telling apart recently selected Bourbon derivatives. It also produced the most marker-dense arabica linkage map to date: 1,854 markers, roughly double the previous best.
Breeding a new coffee variety can take decades, and breeders must evaluate the growth of thousands of plants in the field, a process called phenotyping, before they know which to advance. Molecular markers can be used to shorten this cycle by letting breeders study a seedling’s DNA and predict its performance early, an approach known as marker-assisted selection. Until now, arabica coffee has lacked a cost-effective platform that would allow breeders to do this routinely. A new study just published in Frontiers in Plant Science describes such a tool—a medium-density genotyping panel for Coffea arabica (arabica).
The panel was developed by World Coffee Research (WCR), led by Drs. Jorge Berny and Santos Barrera Lemus, alongside partners from DaRT in Australia, Emprapa in Brazil, Universidad de Frontera in Chile, ICAFE and CATIE in Costa Rica, and the Rwanda Agriculture Board.
Better, faster, cheaper
WCR is using the panel to make its own breeding program better, faster, and cheaper. Marker-assisted selection lets breeders identify promising seedlings years before they fruit. Markers “allow early selection, shorter breeding cycles and reduced phenotyping costs,” the authors note. The panel will also support better-informed choice of parents and trait mapping, linked to our variety-verification work.
“Arabica breeding has lacked a shared, affordable genotyping tool,” says lead author Dr. Jorge C. Berny Mier y Teran, Research Scientist , Plant Breeding & Genomics at WCR. “This panel exists because breeding programs, research institutes and genotyping specialists pooled germplasm, expertise and data. Very high-density marker sets are valuable for some academic research, but most breeding applications, such as genomic prediction, diversity analysis and linkage mapping, don’t necessarily need them. They add cost and data-handling burden without much practical gain. A medium-density panel gives breeding programs the information they need at a price they can use routinely.”
Work is underway to develop a matching panel for robusta, led by World Coffee Research. All six countries participating in the Innovea Global Robusta Breeding Network (Ghana, India, Indonesia, Rwanda, Uganda, and Vietnam) have joined this project, along with CATIE.
In order to utilize the mid-density SNP markers to predict plant performance, researchers must first obtain leaf samples from coffee seedlings in the breeding pipeline.
Open to every breeding program
Any public or private breeding program can order genotyping with the panel through Diversity Arrays Technology (product code Coffee_DArTag_WCR 1.0, custom code CoffeeDArTagWCRv1). Because programs using it share the same markers, their results can be compared and pooled across institutions.
Why breeders need a new tool
Genomic prediction has shown promise in arabica, but genotyping-by-sequencing, the method used in those studies, is not yet economically viable for routine use. An earlier 8,500-marker array was more cost-effective but could not distinguish between arabica’s two subgenomes. This matters because arabica carries two subgenomes, one from each of its two ancestor-species, and they are nearly identical in places. A marker that can’t be placed on the correct subgenome can give misleading results, especially relevant for disease resistance, a critical trait farmers. Subgenome-specific markers allow breeders to follow disease resistance introduced to arabica from other coffee species, including C. canephora and C. liberica, which sits in one subgenome, and tell introgressed plants from non-introgressed ones.
“Coffee breeding programs have not yet benefited from routine use of molecular markers for genomic prediction and mapping due to the lack of a cost-effective genotyping platform,” the authors write.
What the team built
The new panel uses 5,053 SNPs, which are single-nucleotide variations in the DNA sequence. That medium density is enough for genomic prediction and trait mapping without the cost and data-processing demands of high-density approaches.
Arabica’s genome made the design harder. The species is allotetraploid, carrying two subgenomes inherited from its parent species Coffea canephora and Coffea eugenioides, and the markers had to cover both. To find them, the team sequenced 94 diverse plants from the collections at five institutions, spanning wild Ethiopian accessions, Bourbon and Typica founders, introgressed lines carrying disease resistance from other coffee species, and F1hybrids. The panel also includes 53 markers used in our low-density variety identification panel, so new data connects with existing datasets. The institutions providing material included: Brazilian Agricultural Research Corporation (Embrapa), Instituto del Café de Costa Rica, Rwanda Agriculture and Animal Resources Development Board, CATIE, and World Coffee Research.
To develop the new panel, the team sequenced 94 plants from five breeding institutions to select 5,053 SNP markers. This is a tanglegram comparing hierarchical clusters of the two subgenomes in the discovery set accessions. The accessions are colored according to their origin classification. Highlighted in dotted lines are the unique nodes, which contain a combination of samples that are not present in the other tree. The connecting lines are colored based on the common subtrees on both dendrograms. Genetic clustering of the C. canephora subgenome primarily separates introgressed hybrids from traditional, non-introgressed lineages. Meanwhile, the C. eugenioides subgenome differentiates core Ethiopian varieties from domesticated groups like Typica and Bourbon.
Key findings
Telling near-identical subgenomes apart has long challenged sequencing-based genotyping in crops such as peanut and cotton. The new panel handles it well, with consistent results across nearly the whole genome.
It also grouped plants accurately by ancestry and breeding history, separating even recently selected Bourbon derivatives, which an earlier study using nine microsatellite markers could not. This helps breeders choose parents and trace where introgressed resistance genes sit.
The team built a genetic linkage map, which orders markers along each chromosome based on how they are inherited together. Linkage maps are the foundation for locating the genome regions behind agronomic traits, including disease resistance. This one was built from 215 plants of a cross between Marsellesa, an improved rust-resistant variety, and Geisha, a traditional variety.
The panel, the authors conclude, “makes routine, high-throughput genotyping for C. arabica more accessible and affordable, addressing a critical bottleneck in modern coffee breeding.”