{ "id": "polyploidy-and-the-arabica-genome", "url": "https://coffearium.org/entries/polyploidy-and-the-arabica-genome/", "license": "CC-BY-4.0", "title": "Polyploidy and the arabica genome", "summary": "Arabica coffee carries the full chromosome sets of two species, Coffea canephora and Coffea eugenioides. This double genome explains why arabica pollinates itself, how it is inherited and why its age is hard to date.", "domain": "botany", "kind": "Concept", "stage": "plant", "topics": [ "science" ], "topicSections": {}, "lang": "en", "status": "verified", "reviewer": "Coffearium verification pass", "lastVerified": "2026-10-22T00:00:00.000Z", "updated": "2026-10-22T00:00:00.000Z", "added": "2026-10-22T00:00:00.000Z", "brief": [ "*C. arabica* is an allotetraploid: 2n = 4x = 44 chromosomes, from a hybridisation of *C. canephora* and *C. eugenioides*, both diploid with 2n = 22.1,4", "At meiosis, arabica generally shows bivalent pairing of homologous chromosomes and disomic inheritance.1", "The projected arabica genome is 1,281 Mb, against 705 Mb for *C. canephora* and 682 Mb for *C. eugenioides*.1", "A 2011 study links arabica’s self-compatibility to its polyploid origin; most diploid coffee species are self-incompatible.8", "The founding hybridisation is dated to 350,000–610,000 years ago by a 2024 genome study, and to 10,000–50,000 years by a 2012 study.1,5" ], "quickFacts": [ { "label": "Chromosomes", "value": "2n = 4x = 44 (allotetraploid)" }, { "label": "Parent species", "value": "C. canephora × C. eugenioides (both 2n = 22)" }, { "label": "Subgenomes", "value": "subCC (canephora-derived), subEE (eugenioides-derived)" }, { "label": "Inheritance", "value": "Disomic, bivalent pairing" }, { "label": "Genome size", "value": "≈1.28–1.33 Gb (haploid)" }, { "label": "2C nuclear DNA", "value": "2.71 pg (flow cytometry)" }, { "label": "Reference genomes", "value": "ET-39 di-haploid (2024); Bourbon (2024)" } ], "aka": [ "Arabica genome", "Coffee polyploidy" ], "linkTerms": [ "arabica genome", "polyploidy", "polyploid", "subgenome", "disomic inheritance", "homoeologous exchange" ], "related": [ "coffea-arabica", "coffea-canephora", "coffea-genus", "coffea-eugenioides", "the-arabica-bottleneck", "the-timor-hybrid", "arabica-varieties", "ethiopian-landraces", "flowering-and-pollination", "f1-hybrids" ], "contested": [ "Age of the founding hybridisation: a 2024 genome study gives 350,000–610,000 years and suggests that more recent datings could be underestimates1; a 2012 study dated a homoeologous exchange to 10,000–50,000 years before present and took it as the most recent estimate of the origin5; a 2020 study summarised the earlier literature as 10,000 or 665,000 years before present.3", "Regularity of meiosis: the 2024 Nature Genetics study describes generally bivalent pairing and disomic inheritance1; the 2024 Nature Communications study cites cytogenetic evidence that irregularities in disomic meiotic behaviour are ordinary in arabica and found whole-chromosome aneuploidies in four accessions.2" ], "corrections": [ { "date": "2026-10-22T00:00:00.000Z", "change": "The figure legend no longer counts three shared homoeologous exchanges; it names the chromosomes (2, 7 and 10) on which the 2024 study found exchanges fixed in all accessions.", "reason": "The 2024 Nature Communications study reports exchange events on three chromosomes, not a count of three exchanges." } ], "sources": [ { "id": "s1", "title": "The genome and population genomics of allopolyploid Coffea arabica reveal the diversification history of modern coffee cultivars", "authors": "Salojärvi, J.; Rambani, A.; Yu, Z.; Guyot, R.; et al.", "year": 2024, "publisher": "Nature Genetics 56(4), 721–731", "url": "https://repository.up.ac.za/items/a8e96560-6731-4d2a-bc81-ca77e24236ce/full", "doi": "10.1038/s41588-024-01695-w", "tier": "A", "quote": "We therefore consider 610–350 ka a likely time range for the polyploidization event (Fig. 2e).", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s2", "title": "A chromosome-scale assembly reveals chromosomal aberrations and exchanges generating genetic diversity in Coffea arabica germplasm", "authors": "Scalabrin, S.; Magris, G.; Liva, M.; Vitulo, N.; et al.; Morgante, M.", "year": 2024, "publisher": "Nature Communications 15, 463", "url": "https://doi.org/10.1038/s41467-023-44449-8", "doi": "10.1038/s41467-023-44449-8", "tier": "A", "quote": "It has an estimated genome size of 1.33 Gbp based on flow cytometry assays.", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s3", "title": "A single polyploidization event at the origin of the tetraploid genome of Coffea arabica is responsible for the extremely low genetic variation in wild and cultivated germplasm", "authors": "Scalabrin, S.; Toniutti, L.; Di Gaspero, G.; et al.; Bertrand, B.", "year": 2020, "publisher": "Scientific Reports 10, 4642", "url": "https://doi.org/10.1038/s41598-020-61216-7", "doi": "10.1038/s41598-020-61216-7", "tier": "A", "quote": "The allopolyploid speciation of C. arabica has a broad time interval estimate spanning 10,000 or 665,000 years BP", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s4", "title": "Molecular characterisation and origin of the Coffea arabica L. genome", "authors": "Lashermes, P.; Combes, M. C.; Robert, J.; Trouslot, P.; D’Hont, A.; Anthony, F.; Charrier, A.", "year": 1999, "publisher": "Molecular and General Genetics 261(2), 259–266", "url": "https://agritrop.cirad.fr/519522", "tier": "A", "quote": "These results clearly suggest that C. arabica is an amphidiploid formed by hybridisation between C. eugenioides and C. canephora", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s5", "title": "Genome evolution in diploid and tetraploid Coffea species as revealed by comparative analysis of orthologous genome segments", "authors": "Cenci, A.; Combes, M. C.; Lashermes, P.", "year": 2012, "publisher": "Plant Molecular Biology 78(1–2), 135–145", "url": "https://www.documentation.ird.fr/hor/fdi:010054234", "doi": "10.1007/s11103-011-9852-3", "tier": "A", "quote": "Based on nucleotide substitution frequency, the HNRT was dated at 10,000-50,000 years BP, which is also the most recent estimation of the origin of C. arabica.", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s6", "title": "Genome Size Variations in Diploid African Coffea Species", "authors": "Noirot, M.; Poncet, V.; Barre, P.; Hamon, P.; Hamon, S.; de Kochko, A.", "year": 2003, "publisher": "Annals of Botany 92(5), 709–714", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4244848", "doi": "10.1093/aob/mcg183", "tier": "A", "quote": "All taxa are diploid with the same chromosome number (2n = 2x = 22), except C. arabica (2n = 4x = 44).", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s7", "title": "The coffee genome provides insight into the convergent evolution of caffeine biosynthesis", "authors": "Denoeud, F.; Carretero-Paulet, L.; Dereeper, A.; et al.", "year": 2014, "publisher": "Science 345(6201), 1181–1184", "url": "https://pages.uoregon.edu/chendon/coffee_literature/old_literature/2014%20Science,%20Robusta%20genome%20and%20caffeine%20production.pdf", "doi": "10.1126/science.1255274", "tier": "A", "quote": "We annotated 25,574 protein-coding genes", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s8", "title": "Expression and Trans-Specific Polymorphism of Self-Incompatibility RNases in Coffea (Rubiaceae)", "authors": "Nowak, M. D.; Davis, A. P.; Anthony, F.; Yoder, A. D.", "year": 2011, "publisher": "PLoS ONE 6(6), e21019", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC3120821/", "doi": "10.1371/journal.pone.0021019", "tier": "A", "quote": "Self-compatibility in C. arabica is thus not surprising given the strong correlation between polyploidy and the breakdown of SI", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s9", "title": "The Coffee Tree – Genetic Diversity and Origin", "authors": "Herrera, J. C.; Lambot, C.", "year": 2017, "publisher": "In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 1, pp. 1–16", "url": "https://doi.org/10.1016/B978-0-12-803520-7.00001-3", "doi": "10.1016/B978-0-12-803520-7.00001-3", "tier": "B", "quote": "In 1999, molecular and cytogenetic analyses finally allowed the elucidation of the origin of this species, which resulted from the natural hybridization between two ecotypes related to C. eugenioides and C. canephora species (Lashermes et al., 1999).", "accessed": "2026-10-22T00:00:00.000Z" } ], "markdown": "## Four sets of chromosomes\n\nMost coffee species carry two sets of chromosomes, one from each parent. Biologists call such plants diploid. In coffee, one set holds 11 chromosomes, so a diploid plant has 22.6,1\n\n*Coffea arabica* is different. It carries four sets, 44 chromosomes in all. A plant with more than two sets is called polyploid; with four sets, tetraploid.1,3 In a 2003 study of 15 African coffee species and six further taxa, all were diploid with 22 chromosomes; arabica was named as the one exception.6\n\n## Two parents in one genome\n\nArabica’s four sets come from two different species. Two sets descend from *C. canephora*, the species grown as robusta. The other two descend from *C. eugenioides*, a wild coffee species.4,1 A polyploid built from two species is called an allotetraploid.1 The second parent is described in [Coffea eugenioides](/entries/coffea-eugenioides/).\n\nA 1999 study first showed this parentage. It compared genetic markers and labelled arabica chromosomes with DNA of the candidate parents.4,9 Genome assemblies of arabica and of both parents confirmed it in 2024.1\n\nEach parental half of the genome is called a subgenome. The 2024 study names them subCC, from the canephora side, and subEE, from the eugenioides side.1 The plastid genome of arabica, the small genome inside its chloroplasts, comes from *C. eugenioides*.1\n\n
\n
\n
    \n
  1. C. canephora, diploid. Two sets of 11 chromosomes, 2n = 2x = 22.7,1
    2. \n
  2. C. eugenioides, diploid. Two sets of 11 chromosomes; the source of arabica’s plastid genome.1
    3. \n
  3. Founding polyploidy event. A single hybridisation, 350,000–610,000 years ago by the 2024 dating; other datings differ.1,3
    4. \n
  4. C. arabica, 2n = 4x = 44. Both parental genomes are kept.4,1
    5. \n
  5. Canephora-derived subgenome (subCC). Its two copies pair with each other at meiosis.1
    6. \n
  6. Eugenioides-derived subgenome (subEE). Its two copies pair with each other as well.1
    7. \n
  7. Homoeologous exchange. A segment of one subgenome has replaced its counterpart in the other; exchanges on chromosomes 2, 7 and 10 are shared by all accessions studied.5,2
    8. \n
\n
Parentage and subgenomes of C. arabica. Filled: chromosomes from C. canephora; hollow: from C. eugenioides. Schematic, not to scale.1,2,4
\n
\n\n## Why arabica pollinates itself\n\nArabica mainly fertilises itself; pollinators may occasionally carry pollen between plants.3,1 Most diploid coffee species cannot do this. They reject their own pollen through a mechanism called gametophytic self-incompatibility.8\n\nA 2011 study of the genes behind this mechanism links the difference to polyploidy. It states that arabica’s self-compatibility is likely the result of its allotetraploid origin, given the strong link between polyploidy and the breakdown of self-incompatibility in plants.8 Two diploid species, *C. heterocalyx* and *C. anthonyi*, are also self-compatible; the cause in those is unknown.8 How coffee flowers are pollinated is covered in [Flowering and pollination](/entries/flowering-and-pollination/).\n\n## How large the genome is\n\nThe arabica genome is close to the size of its two parents’ genomes together. A 2024 genome study projects 1,281 Mb (million base pairs) for arabica, 705 Mb for *C. canephora* and 682 Mb for *C. eugenioides*; the two parents add up to 1,387 Mb by our calculation.1 A 2014 study gave about 710 Mb for *C. canephora*.7 A second 2024 study estimated 1.33 Gb for arabica by flow cytometry, which measures the DNA in cell nuclei.2\n\n
\n
\n
Genome size of C. arabica and its two parent species, in megabases, from three genome studies. Projected sizes are estimates of the whole genome, not the assembled length.1,2,7
\n
\n\nArabica’s narrow genetic diversity also goes back to its single origin and later bottlenecks; that story is told in [The arabica bottleneck](/entries/the-arabica-bottleneck/).3,1\n\n\n\n## Subgenomes and meiosis\n\n*C. arabica* (2n = 4x = 44) is an allopolyploid whose subgenomes derive from *C. canephora* and *C. eugenioides*.3,4 The 1999 study used restriction fragment length polymorphism (RFLP) markers and genomic in situ hybridisation (GISH); it concluded that arabica is an amphidiploid formed from these two species or closely related ecotypes.4 It also found low divergence between each subgenome and its parent species, which pointed to a relatively recent origin.4 According to a 2017 textbook chapter, this low divergence supports the hypothesis of a very recent speciation.9\n\nArabica generally shows bivalent pairing of homologous chromosomes and disomic inheritance.1 Each chromosome pairs with its partner from the same subgenome, not with the matching (homoeologous) chromosome of the other. A 2024 study cites cytogenetic evidence that irregularities in this disomic behaviour are ordinary in arabica.2\n\nThe genome structure has changed little since the hybridisation. The 2024 Nature Genetics study found largely conserved structures between the diploid parents and the descendant subgenomes, with no obvious global dominance of either subgenome.1 Genes were lost in similar proportions from both subgenomes, and the polyploidy event seemingly did not affect the rate of this gene loss (fractionation).1 *C. canephora* itself shows no sign of a whole-genome polyploidisation in its lineage since the ancient gamma triplication.7\n\n## Exchanges and chromosomal changes\n\nSegments can move between the subgenomes. In a homoeologous exchange, a stretch of one subgenome replaces the corresponding stretch of the other. A 2012 study characterised a 50 kb region of the eugenioides-derived subgenome that had replaced its canephora-derived counterpart.5 The 2024 Nature Genetics study detected a bias in such exchanges in favour of subCC.1\n\nThe 2024 Nature Communications study, based on a chromosome-scale assembly of the variety Bourbon, found exchanges on chromosomes 2, 7 and 10 that occurred very early in the arabica lineage and are fixed in all accessions analysed.2 It also found many recent chromosomal aberrations, including aneuploidies, deletions, duplications and exchanges, that remain polymorphic in the germplasm.2 Four accessions showed whole-chromosome aneuploidies.2 The authors propose such variants as a source of variation in a species with otherwise low diversity.2\n\nThe same study found evidence of cryptic robusta introgression, via derivatives of the Timor Hybrid, in arabica-like germplasm.2 The Timor Hybrid, a natural cross of arabica and *C. canephora*, is covered in [The Timor Hybrid](/entries/the-timor-hybrid/).\n\n## Dating the hybridisation\n\nThe subgenomes have diverged little from their parent species.4 Published datings span two orders of magnitude.1\n\nA 2012 study dated its homoeologous exchange to 10,000–50,000 years before present, from nucleotide substitution frequency, and called this the most recent estimate of arabica’s origin.5 A 2017 textbook chapter, citing that study, gives the speciation as 10,000–50,000 years BC.9 A 2020 study summarised the literature as a broad interval of 10,000 or 665,000 years before present.3\n\nThe 2024 Nature Genetics study considers 350,000–610,000 years a likely range for the founding polyploidy event.1 It notes that earlier datings ran from 10,000 to 1 million years ago and that more recent timings, such as 20,000 years, could be underestimates.1 The same study times the bottlenecks that followed with a generation time of 21 years.1 The 2024 Nature Communications study gives no date of its own and speaks of a very recent origin.2\n\n
\n
\n
Published datings of the hybridisation that founded C. arabica, in thousand years before present. The 2012 interval dates an exchange between the subgenomes, which that study took as the most recent estimate of the origin.1,5
\n
\n\nWhere the hybridisation took place is also open. The 1999 study noted that it is difficult to locate, because the extent of tropical forest in Central Africa changed considerably during the late Quaternary.4\n\n## Reference values\n\n| Parameter | *C. arabica* | *C. canephora* | *C. eugenioides* |\n|---|---|---|---|\n| Chromosomes | 2n = 4x = 441,6 | 2n = 2x = 227,6 | 2n = 2x = 221,6 |\n| Projected genome size | 1,281 Mb1 | 705 Mb1; ≈710 Mb7 | 682 Mb1 |\n| Genome size, flow cytometry | 1.33 Gb2 | – | – |\n| 2C nuclear DNA, flow cytometry | 2.71 ± 0.04 pg2 | 1.440 pg (Central African Republic population)6 | 1.364 pg6 |\n| Assembly length | 1,198 Mb (HiFi, ET-39)1; 1.32 Gb (Bourbon)2 | 672 Mb1; 568.6 Mb (2014)7 | 661 Mb1 |\n| Gene models | 69,314 (HiFi assembly)1 | 28,857 (text; 28,880 in Table 1)1; 25,574 (2014)7 | 32,1921 |\n| Transposable elements | 63.1% and 63.8% (two subgenomes)1 | 67.5%1 | 59.7%1 |\n| Breeding system | Self-compatible, mainly selfing8,3 | Self-incompatible (gametophytic)8 | Self-incompatible (listed in the 2011 study’s table)8 |\n\nAmong the diploid African species measured in 2003, 2C values ranged from 1.04 pg in *C. racemosa* to 1.76 pg in *C. humilis*; that study used flow cytometry with *Petunia hybrida* as the internal standard.6 In the subgenus *Coffea*, genome size generally increases from East to West Africa, a gradient the authors relate to rainfall while cautioning that a correlation does not show cause.6 The 2024 arabica reference was a di-haploid accession, ET-39; the Bourbon assembly was made separately.1,2\n\n## Open questions\n\n- **Age of the species.** Datings range from tens of thousands to several hundred thousand years, depending on the method and on how bottlenecks are treated (see Contested).1,5\n- **Place of origin.** Shifts in Central African forest during the late Quaternary make the site of the hybridisation hard to locate.4\n- **Cause of self-compatibility.** The 2011 study links it to polyploidy and found three putative self-incompatibility RNase alleles in arabica, one possibly a pseudogene; the mechanism of the breakdown was not shown.8\n- **Meiotic stability.** How often irregular pairing produces aneuploid or rearranged plants, and what that means for breeding, is not established in the studies opened for this entry.2\n\n## Key studies\n\n
\n

Lashermes et al., 1999 – The two parents of arabica. Compares arabica with candidate diploid species using RFLP markers and labels chromosomes with the parents’ DNA. Identifies C. eugenioides and C. canephora as the sources of arabica’s two genomes and finds them little diverged from their parents.

\n

Molecular and General Genetics 261(2), 259–266 · agritrop.cirad.fr/519522

\n
\n\n
\n

Noirot et al., 2003 – Genome size across diploid coffees. Measures nuclear DNA by flow cytometry in 15 African species and six further taxa. Finds 2C values from 1.04 to 1.76 pg among the diploids, generally increasing from East to West Africa.

\n

Annals of Botany 92(5), 709–714 · doi:10.1093/aob/mcg183

\n
\n\n
\n

Cenci, Combes & Lashermes, 2012 – Orthologous segments compared. Sequences matching genome regions of diploid coffees and of both arabica subgenomes. Describes a homoeologous exchange in arabica and dates it to 10,000–50,000 years before present.

\n

Plant Molecular Biology 78(1–2), 135–145 · doi:10.1007/s11103-011-9852-3

\n
\n\n
\n

Salojärvi et al., 2024 – Arabica and its parents assembled. Presents chromosome-level genomes of a di-haploid arabica and of modern representatives of both parents. Finds conserved subgenome structure, disomic inheritance and a founding event 350,000–610,000 years ago.

\n

Nature Genetics 56(4), 721–731 · doi:10.1038/s41588-024-01695-w

\n
\n\n
\n

Scalabrin et al., 2024 – Chromosomal changes in arabica germplasm. Assembles the genome of Bourbon and resequences a large panel of accessions. Finds early exchanges between the subgenomes and many recent aneuploidies, deletions and duplications that add variation to a narrow species.

\n

Nature Communications 15, 463 · doi:10.1038/s41467-023-44449-8

\n
" }